Cryosphere Posters

Himalayan glaciers are quite reliable indicators of changes in the climate system. The study of a debris covered glacier – Lirung glacier (28.241 N, 85.556 E), Lantan sub-basin of Gandaki river, Nepal through remote sensing and field survey reveals the glacier dynamics. The glacier recession is active and glacier thinning is seriously observed in Lirung glacier. A proglacial lake formed at the evacuated historical terminus of the glacier bounded by the lateral moraines. The glacier broken as two separate bodies; the accumulation zone and ablation zone are separated. This activity is mentioned here as “One-By-Two” phenomenon. As the glaciers not receiving any ice mass from accumulation zone, the mass balance will be always negative. Many supraglacial activities such as formation of large supraglacial lakes, exposing ice cliffs initiated in the broken ablation zone of the glacier. The supraglacial lakes and the exposed ice cliffs of the lakes receive more solar radiation and fasten the melting of glacier ice. Finally the ablation part melts faster and disappears. The accumulation area turns to become a “glacieret” with no dynamics. This type of activity is getting more prevalent in the Himalayan cryosphere due to change in the precipitation and temperature increase. Such breaking of glacier will have pronounced impact on the hydrology of rivers originating from the Himalayan cryosphere.

The aim of this work is to assess the current state of the highlands’ glaciation by the high resolution satellite images, to compare with the Glacier inventory data at the background of climatic parameters, such as temperature and precipitation trends over the last 50 years, as well as assessment of the evolution of glacier systems in the near future.

A recent study, done by a group of authors presents a generalization and systematization of the information on glaciers of the Chukotka and Kolyma highlands, which was given by the Magadan researcher R.V. Sedov.

Glaciers of the Kolyma highlands consist of two groups: five are located on the eastern slope of the Kolyma highlands near the western shore of the Sea of Okhotsk, the equilibrium line altitude (H_ELA) is from 700 to 1500 m a. s. l., and 14 cirque glaciers are located in the northern part of the Taigonos Peninsula, the equilibrium line altitude (H_ELA) from 700 to 1000 m.

Glaciers of the Chukchi Highlands, according to R.V. Sedov, are represented by five groups: in the Teniany Range in the Lavrentiya Bay, the mean H_ELA is 500 m a.s.l. in the Providenskiy Mountain Massif, the H_ELA is from 400 to 550 m, in Iskaten Ridge, H_ELA  from 500 to 1000 m, on the Pekulney Ridge, the mean H_ELA was 740 m, in the Amguema River basin with an average H_ELA1400 m. (Kotlyakov et al, 2011).

 To determine the parameters of glaciers the high-resolution satellite images have been used dated by August 2012, which cover the glaciers of these regions, courtesy of program AMAP (Arctic Monitoring Assessment Program), as well as LandSat-7 for the same period, which are in the public access.

We were able to detect and decipher 27 glaciers in the area of the Iskaten Range, Cross Bay , 16 glaciers in the Providenskiy Massif, 6 in the Lawrence Bay, 5 - on the Pekulnei Range and 19 - in the Kolyma Highlands (the same number as was the specified of R.V. Sedov). In general, the trend is clear: the glaciers have decreased in size as compared to the estimates of late 1980s. Particularly small glaciers of Pekulnei Range “suffered” most, there are not more than 7% of the area remained from that of indicated by R.V. Sedov. The rest of the glacial systems reduced much less, the proportion of the remaining area is of 66% (Kolyma Highlands) to ~ 40% (Iskaten Range).

Method for estimating glacier systems evolution was described in our papers (Ananicheva and Krenke, 2007, Ananicheva et al, 2010); in this work we had to make some changes, caused by the climate scenario features. In order to project future changes in the morphology and regime of the Chukchi Highlands glacier systems the output data (temperature and precipitation), calculated by the ensemble models A-31 (Kattsov, Govorkova 2013) were used.

The glaciers of the Chukchi Highlands in 2030, based on chosen climatic scenario (A-31) will reduce in size in different way. Small glaciers will completely disappear from Pekulnei Range, so far Н_ELA has shifted upward maximum (410 m) among other studied systems, and a glacier area remained is just 0.1 km². Glaciers of Iskaten Ridge (the Cross Bay) and Providence Bay Massif by 2012 have saved more than 40% of their area, and by 2030 they will lose a large part of it which will remain only 6.1%, and 13, 6%, respectively, offset by the time Н_ELAreach 350 m. The best preservation of glaciers belongs to Amguema River basin, located in northeastern Chukotka (~ 60%), but there are small glaciers there, and up to 2030 only ~ 0.7 km² covered by ice will "survive". In total, on the basis of our constructions and the scenario ~ 11.5% of glaciation will remain to 2030. Ablation-accumulation at the Н_ELAto that time will vary from 230 to 680 mm from system to system.

Objectives of presentation are demonstrate results of D-InSAR technology testing for ongoing changes estimathions in the permafrost research area for terrestrial North Eurasia Arctic ecosystems. As was early illustrated for North Canada (Short et all., 2011) and Central Siberia (Chimitdorziev et.all., 2010; 2011; 2013) the D-InSAR technology can be selected as a one of the most informative monitoring systems, which demonstrate seasonal and interannual changes of Earth surface related with dynamic of active layer depth in present. Few different adjacent cryolitozone regions was taken to account as models. Among them: European North-East (EN), Western (WS) and East (ES) Siberia. Contiguous model areas were selected, since we assume that the characteristics of permafrost East European tundra areas with more ocean-oriented climate and "weak, warm" permafrost (Astakhov, Svensen, 2011) are more vulnerable than Urals ecosystems of Siberia. More dry and harsh climatic conditions of the Trans-Urals located in the continental regions are less affected by climate change. We focused on the development of D-InSAR methodology for assessing the thawing permafrost dynamic, since we have compiled a database of land instrumental observations for the period 2000-2015 years in several monitoring sites within the Circumpolar Active Layer Monitoring (CALM) program. The combination of satellite and field observations extend and improve the D-InSAR technique for estimating the active layer. As a result We develop methodology for large-scale mapping of features of cryolitozone on local (effects of landscapes, geomorphology, glacial sediments and late Pleistocene/Holocene history) and regional (zonal climate specific, permafrost types) levels with satellite data. Possibility of estimates for the relatively warm (2007) and cold (2010) year is the basis for developing a model permafrost changes with effects on the native climatic changes in gradients of latitude-longitude scale and anthropogenic (reindeer grazing system, coal, oil and gas system) effects.

Recently the feasibility of commercial shipping in the ice-prone Northwest Passage has attracted a lot of attention. However, very little ice thickness information actually exists. We present results of the first-ever airborne electromagnetic ice thickness surveys over the Northwest Passage carried out in April and May 2011 and 2015 over first-year and multi-year ice. Results show modal thicknesses between 1.8 and 2.0 m in all regions. Mean thicknesses over 3 m and thick, deformed ice were observed over some multiyear ice regimes shown to originate from the Arctic Ocean. Mean ice thickness and deformation decrease with decreasing northern latitude. Thick ice features more than 100 m wide and thicker than 4 m occurred frequently. Although likely thinner than some 20 or more years ago, ice conditions must still be considered severe. These results have important broad implications for the prediction of ice break-up and summer ice conditions, and the assessment of sea ice hazards during the summer shipping season.

Bistatic TanDEM-X interferometric synthetic aperture radar (InSAR) data were used to produce digital elevation models (DEMs) over winter, dry snow conditions in a sub-Arctic site in Norway. Interference patterns were assessed both qualitatively and quantitatively. Open fields and forest stands (slope < 20%), were analysed. Interferometric phase uncertainty and spatio-temporal consistency were analysed by means of scattering phase centre elevation accuracy and coherence degree, γ with respect to hydro-meteorological conditions. Open fields displayed absolute height errors of < 1m. The sole approximation was to consider volume decorrelation effects as main source of elevation ambiguities and coherence decay. Principal Component Analysis was conducted to identify the combined influence of variables describing the radar response and scattering centre. The major findings were that coherence was strongly related to snowpack parameters, and local incidence angle for all land cover classes, while backscatter intensity was related to the coherence degree for the open fields and to forest parameters such as diameter at breast height and stand density for the forest stands.  However the derived DEMs attained high vertical accuracy, the sub-meter height errors might relate to snowcover effects. Coherence was strongly influenced by snowpack parameters suggesting TDM coherence inversion may enable snow cover monitoring

There have been dramatic changes in the Arctic in the past three decades. One major change is the accelerated declining in the sea ice especially in the summer. Ice-free Arctic summers are predicted by climate models for the coming decades. The cause of the accelerated sea ice decline and its impact on other components of the Arctic climate system is intriguing and under investigation. The cloud feedback is the primary source of uncertainty in model simulations, especially in the Polar Regions. A better understanding of the interaction between sea ice and clouds will provide valuable insight into the Arctic climate system and may ultimately help in improving climate model parameterizations.

 

Space-based observations play a critical role in polar climate research. Here we use a variety of satellites to monitor and study the Arctic climate system. Visible, infrared, passive microwave, lidar, and radar altimeter data from Suomi National Polar-Orbiting Partnership (S-NPP), Aqua and Terra, Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO), CryoSat-2, Defense Meteorological Satellite Program (DMSP), and NOAA satellites are used individually and collectively to examine the influence of sea ice on cloud cover and vice versa. Two major findings are presented. Results show that a 1% decrease in Arctic sea ice concentration leads to a 0.36 – 0.47% increase in cloud cover, suggesting that a further decline in sea ice cover will result in an even cloudier Arctic. This relationship implies a positive feedback between sea ice and clouds. Results also demonstrate that Arctic wintertime surface cloud forcing can contribute to summer sea ice variations. In particular, winter cloud forcing anomalies in the Beaufort, Chukchi, and East Siberian Seas explain approximately 30% of the variance in late summer sea ice concentration over the period 1982-2014.

 

One case study on the application of multiple satellite products in studying the cloud and sea ice interaction is presented. In September 2012, Arctic sea ice extent reached a record minimum for the satellite era. The following winter the sea ice quickly returned, carrying through to the summer when ice extent was 48% greater than the same time in 2012. Using five different satellite products, two climate reanalysis, aircraft data, and a simple ice growth model we show that another factor, below average Arctic cloud cover in January - February 2013, resulted in a more strongly negative surface radiation budget, cooling the surface and allowing for greater ice growth. Areas of ice growth estimated from the negative cloud cover anomaly and advected from winter to summer with ice drift data correspond well with the September ice concentration anomaly distribution.

Satellite radar altimetry over ice sheets provides elevation measurements which reveal important properties and behaviours, including fluctuations in mass balance and in the hydrological connections between subglacial lakes. The reliability of these measurements is affected by topographical and sub-surface properties of the snowpack, which exhibit high spatiotemporal variability across Antarctica and Greenland. Of particular note is the observation of temporally correlated fluctuations in elevation and power, which impact radar altimeter range measurements. A numerical model of microwave backscattering is presented which, when applied to satellite waveforms, determines the extent and variation of backscattering across the Antarctic ice sheet, and allows the mapping of sub-surface features. This model is applied to CryoSat-2 data to develop an improved understanding of the variation of scattering properties within the snowpack. In addition, the incorporation of a firn evolution model is explored in order to investigate the source of observed backscatter variations.

The reliability of Surface elevation change in Essential Climate Variables in CCI has been analyzed. A complete analysis of ICESAT laser altimeter data between 2003 and 2008 improved understanding of snow drifting. Elevation change of the surface on the Centre part of Antarctic ice sheet with ice velocity about 20 cm/year depends mostly on snow drifting. This process is described as scaling phenomena in order to form a basic understanding provided by the fractional calculus. This non-stationary process is characterized by spatial frequencies in the range 1.5 – 22.5 km. Our data show that short-term changes in drifting snow are common throughout the Central part of the Antarctic ice sheet. Local elevation changes may be in the range ±20 cm. The crucial feature of the process is elevation changes along the repeat ICESAT tracks. They exhibit specific kind of fractal-like behavior and self-similarity in the range of 70 km. An analytical description is difficult to do. Consequently, drifting snow has no characteristic scale length spatial heterogeneity. It became apparent that not just static structures of ice surface especially above subglacial lakes are fractal, but the dynamics are fractal as well including their statistical fluctuations of the actual surface which are Gaussian. The fractal dimension provides an indication of how rough a surface elevation change is. The Hurst exponent directly related to the fractal dimension. Note that all track processes are characterized by Gaussian one-dimensional distribution. The association between seasons and intensity of drifting snow reflects the contrast in fractal dimension and scaling range. Snow drifting is driven by fluctuating velocity wind fields which are fractal statistical process. Spatial frequency of elevation change due to snow drifting was calculated by using wavelet transform. This provided the spatial frequency about 5.5 km and frequency location. There is evidence that the Center part of the Antarctic ice sheet is in the state of dynamical equilibrium.

Albedo is defined as diffuse reflectivity calculated by the directional integration of reflectance over all sun-view geometries. It can be interpreted as the ratio of solar irradiance reflected from the surface and can be expressed on an interval that ranges from 0 (0%) to 1 (100%). The average albedo of the earth is around 0,31, whereas the albedo of snow and ice ranges from 0,6 to 0,9. During the last centuries climate change and global warming are major concerns as polar ice caps are shrinking and glaciers are retreating. While more ice is melting, more energy irradiated by the sun is absorbed at polar regions and albedo diminishes. As one consequence, ice melt accelerates. The so called positive feedback occurs.

A couple of scientific studies and monitoring programs have revealed that West Antarctica is one of the fastest warming regions globally. Analysis of ground based and satellite observations recorded during the last 50 years show a warming trend of 0,17°C per decade. Significant warming during austral summer period was reported recently, which might result in extensive surface melt with more frequency.

One possibility to estimate changes in glaciers, ice shelves, etc. is through remote sensing and satellite image processing and analysis techniques. Calculating the albedo of an exposed surface based on remote sensing has repeatedely generated reliable results, which were used in climate change research. We propose to calculate the albedo in Antarctic polar regions by the use of data gathered by optical high resolution remote sensing sensors, more specifically the Pleiades mission. In literature nothing can be found about albedo calculation based on Pleiades satellite imagery. Ground Sampling Distance of panchromatic observations is 50cm and of the 4 spectral bands (R, G, B, NIR) for multispectral imagery it is 2m.

In particular, we carried out a case study that shows how to calculate the albedo on the northern edge of the George VI ice shelf. In this study we developed an algebraic formula to calculate the albedo. This formula is a linear equation, which comprises the four spectral bands R, G, B and NIR, and 4 coefficients, one for every spectralband. These coefficients have been determined from samples of different snow and ice types. A comparison of the albedo values calculated in this study with those derived from other experiments that can be found in literature indicate that they almost coincide. Statistical significance was analysed based on Fisher's exact test and at 99% level of confidence the test did not fail. Therefore, we conclude that the proposed methodology allows to achieve a high degree of plausibility to calculate albedo from Pleiades high resolution satellite imagery.

Synthetic aperture radar (SAR), which is hardly influenced by weather conditions (such as cloud and precipitation) and solar illumination, provides strong support for glacier change monitoring. SAR is also sensitive to ground moisture and roughness, which enables it to monitor glacier surface changes. For high-resolution SAR, it is found that it presents more details on glacier surface which are invisible on low-resolution or optical images. In this paper, the phase congruency method is applied to extract glacier outlines and stripes using X-band high-resolution SAR imagery in the Qinghai-Tibet Plateau collected by the TerraSAR satellites. The phase congruency method, which is independent of the image illumination and magnification, performs excellently in glacier outline and stripe extraction. Two kinds of stripes can be observed and extracted. The first type which is basically parallel to the glacier center axis, is eroded by surface melting water. The second type which is basically perpendicular to the glacier center axis, is the crevasses originating from the tension inside the ice body. In cloudy glacial regions, the SAR-derived outlines are essential tools for yearly glacier mapping. By contrast, the extracted stripes on high-resolution SAR image may provide new long-term indicators of the glacier changes.

Polarimetric Synthetic Aperture Radar (PolSAR) sensors represent, nowadays, an established tool for the observation of glaciated areas. Being active microwave systems, SAR systems can operate in nearly all-weather conditions, during the long and dark polar winters. Moreover, exploiting the synthetic aperture concept, they are able to perform large scale and high-spatial resolution observations. At dry conditions, they combine the penetration capability of microwaves into snow/ice and the sensitivity of polarimetry to different scattering mechanisms. Therefore, PolSAR measurements are sensitive to glacier surface as well as near-surface features, especially at lower frequencies (e.g., L-band).

In the past few years, Circular SAR (CSAR) has become of increasing interest due to its enhanced (sub-wavelength) spatial resolution and 3-D reconstruction capability allowed by multi-angular measurements over 360° [1]-[3]. In this sense, CSAR has the potential to provide a better characterization of the imaged scene, compared to the conventional linear imaging modes, such as stripmap. In fact, the circular trajectory of the sensor allows to observe the scene from a number of different directions and to mitigate the dependency of the backscatter on the imaging geometry (i.e., flight direction and incidence angle).

The objective of this study is to provide a physical interpretation of L-band polarimetric CSAR (Pol-CSAR) measurements of the Findel Glacier, a temperate glacier located in the Swiss Alps. The data collection took place in October 2014 and March 2015, when two airborne campaigns were flown by the airborne F-SAR sensor of the German Aerospace Center (DLR). Data analysis and interpretation are performed applying, for the first time, a polarimetric scattering model to Pol-CSAR data. For this, the model proposed in [4] is employed to retrieve snow and firn parameters of the shallow glacier subsurface [5], of relevance for the estimation of surface mass balance. Validation of the results is performed by means of ground measurements (i.e., GPR profiles, snow depth) collected in coordination with the SAR campaigns.

 

References 

[1] Ponce, Octavio, et al. "Fully polarimetric high-resolution 3-D imaging with circular SAR at L-band." IEEE Transactions on Geoscience and Remote Sensing, vol. 52, no. 6, pp. 3074-3090, 2014.

[2] Ponce, O., Prats-Iraola, P., Scheiber, R., Reigber, A. and Moreira, A. "Polarimetric 3-D reconstruction from multicircular SAR at P-band." IEEE  Geoscience and Remote Sensing Letters, vol. 11, no. 4, pp. 803-807, 2014.

[3] Ponce, O., Prats, P., Scheiber, R., Reigber, A. and Moreira, A. "Polarimetric 3-D Imaging with Airborne Holographic SAR Tomography over Glaciers”, Proceedings of IGARSS 2015, Milan, Italy.

In the framework of a study of new-ice formation in Antarctica, SAR image acquisitions were planned over Terra Nova Bay (TNB). Thanks to the ESA Third Party Mission program, Cosmo-SkyMed and Radarsat-2 image acquisition were ordered for the period 20 February – 20 March, 2015; in addition, available Sentinel-1images over TNB for the same period were retrieved from the ESA Scientific Data Hub. The first inspection of the images revealed the presence of a prominent gyre of surface ice, probably produced by the wind blowing from the continent. Our first goal was to investigate the correlation between gyre area and wind field. Wind data were obtained from the AWS ‘Eneide’ located in proximity of the Italian Antarctic base “Mario Zucchelli Station” at TNB.

A processing scheme was developed for measuring the gyre area; it consists of the following blocks: 1. Non-linear filtering [1]; 2. Segmentation based on the Markov Random Field theory [2-3] and using a contextual approach which considers both the original and the filtered image; 3. Extraction of the gyre parameters, area and perimeter, by means of an active contour detection algorithm which works in an iterative fashion [4].

The correlation between gyre area and wind field was analysed by means of the running correlation coefficient function Ri which can reveal the consistency between the two variables. Ri attained high values in the period 20February – 15 March; after this date, a powerful katabatic wind completely disrupted the surface ice gyre by displaying a well-defined polynya.

 REFERENCES

[1] Perona, P., and Malik, J. (1990). Scale space and edge detection using anisotropic diffusion, IEEE Trans. PAMI, vol. 12, 629–639.

[2] S. Z. Li, S.Z. (2009). Markov Random Field Modeling in Image Analysis (Advances in Computer Vision and Pattern Recognition), Springer, 3rd edition.

[3] R. Kindermann and J. Snell, Markov Random Fields and Their Applications, Providence, RI, USA: American Mathematical Society, 1980.

[4] M. Kass, A. Witkin, and D. Terzopoulos, Snakes: active contour models, Int. Journal of Computer Vision, vol. 1, pp. 321-331, 1988.

Several meteorological institutes create daily ice charts, which are freely available to their users. In the European Arctic however, these ice charts only get delivered in the late afternoon, and not in the weekends, whereas ships and (mobile) oil rigs are out there around the clock all year round. The ice situation can change rapidly, so that even the 7-8 hours it takes to create a detailed ice chart from satellite images can lead to outdated information and possibly dangerous situations. Accurate, fast information about the ice situation is necessary. KSAT can deliver information about the ice edge in NRT.

As ships and oil rigs operate near the ice edge, they can rely on this ice edge to get fast, accurate and frequently updated information about its location and plausible movement.

The ice edge developed by KSAT, delivers accurate, high resolution information within 15 minutes after image acquisition. This ice edge can be created from several sources of SAR imagery, such as RADARSAT-2, COSMO-SKYMED or TerraSAR-X. Moreover, it works on fine scales, since it is automatically drawn with the resolution of the satellite image. In case of Wide Swath imagery, often used in the Arctic, the resolution is 100 meters.

The ice edge has been tested and approved so far in the Barents Sea, the Fram Strait and the Kara Sea, and has proven to be of use for fishing vessels and oil companies operating in or near the ice edge. KSAT has internally verified the accuracy of the line against the manual ice charts produced by leading meteorological institutes and is continually working to evaluate the statistical performance and to identify areas of relative strength and weakness. 
In short, KSAT’s ice edge gives everyone who is working in ice-infested waters, near the ice edge, consistent, accurate and near-real-time information about the situation near the ship or oil rig. 

Sea ice freeboard derived from satellite altimetry is the basis for estimation of sea ice thickness using assumption of hydrostatic equilibrium. Therefore a high accuracy of altimeter measurements and freeboard retrieval procedure is required. Currently two approaches for estimation of the freeboard using laser altimeter measurements from Ice, Cloud, and land Elevation Satellite (ICESat), referred to as tiepoints (Kwok et al., 2007) and lowest-level elevation methods (Zwally et al., 2008), have been developed and applied in different studies. We reproduced these retrieval algorithms in order to assess and analyze the sources of the differences between freeboard estimates for the Arctic sea ice. It was found that approach for determination of sea level references applied in the lowest-level elevation method for some ICESat observation periods underestimates freeboard estimates over thickest part of multy-year ice locally by up to 15 cm. In contrast, tiepoits method tends to underestimate freeboard over vast areas of thin first-year ice by 3-5 cm when using original retrieval algorithms, and by less than 2 cm if applying improvements proposed in this study. Also the effects of different along track resolutions and geoids used in the freeboard retrieval process are found to contribute to the difference between the estimates. In addition, in the studies where freeboard estimates are retrieved by tiepoints method they are corrected for snow depth and presence of ice within the ICESat footprint of samples used to determine sea surface references. An applying of these adjustments can explain a large part of the observed difference in the available products of the Arctic sea ice thickness.

Radar altimeter measurements from ERS, Envisat and Cryosat-2 ESA’s satellites have been used for study of the ice sheet elevation changes for more than two decades. The follow-on SARAL  ISRO/CNES mission with the radar altimeter AltiKa on board was launched in February 2013 on the same orbit as Envisat. However, in contrast to the previous Ku-band radar altimeters, AltiKa operates in Ka-band (36.8 GHz) resulting in smaller footprint, better vertical resolution and decreased penetration of the signal in the snowpack. This work presents Greenland ice sheet  surface elevation changes (SEC) derived from the first years of SARAL/AltiKa operation as part of the ESA’s Climate Change Initiative program, which addresses the GrIS as one of the Essential Climate Variables. Seasonal changes in elevation and radar altimeter waveform parameters are estimated using crossover and stacking methods and compared with those derived from ERS, Envisat and CryoSat data.

Sea ice concentration (SIC) is an essential sea ice parameter e.g. for ice navigation, offshore actions, weather forecasting models and climate studies. We have studiedthe use of SENTINEL-1a (S-1a) SAR Extra Wide Swath (EW) mode dual-polarized data with HH/HV polarization combination for estimating SIC over the Baltic Sea. For this purpose we have used the available EW mode HH/HV S-1a data over Baltic acquired during the winters 2014-2015 and 2015-2016. The data used in our study are S-1a level 1 GRDM (medium resolution) data, i.e. the pixel spacing is about 40m.

The data are first calibrated and rectified to Mercator projection, which is also used in the Baltic Sea nautical charts, to ensure compatibility of the SIC product with the on-board navigation systems. Before using the SAR data for automated SIC estimation, an incidence angle correction is performed for both the channels. For HH-channel we apply a linear correction defined experimentally for sea ice. Over open water SAR backscattering is dependent, in addition to the incidence angle, on the local 2-dimensional wave spectrum (local water surface roughness), and correction can not usually be performed as the exact wave spectrum information is missing. For HV-channel the incidence angle and wave spectrum dependence is smaller than for the HH-channel. However, for the HV channel a noise floor correction (equalization)
is necessary, because the S-1a noise floor significantly varies as a function of the range and with respect to the HV signal strength. We perform the noise floor equalization either by using a self-defined statistical equalization (for the season 2015-2015 imagery), or using the noise floor data provided by ESA (for the season 2015-2016 imagery). The reason for using the self-defined statistical correction is that ESA did not provide proper noise floor data until summer 2015.

After the calibration and noise floor equalization, georectification and necessary corrections a SAR segmentation is applied. This segmentation is applied to Principal Component (PC) image of the two channels. SIC estimation is performed for each segment produce by the segmentation, i.e. our SIC estimates are segment-wise.

Two SIC estimation approaches have been studied: a linear and a nonlinear method.
The inputs for the estimation methods are the HV backscattering coefficient complemented by texture measures (including auto-correlation, entropy, and segment-wise corner pixel count for both the SAR channels) computed for HH and HV channels separately. Also cross-correlation between the two SAR channels has been used as an input. The input features are computed for each segment. For the study the data is divided into training and testing data sets. Training data set is used for defining the estimation parameters and test data set as an independent  valuation
data set. The linear estimation is based on the least-squares fit of the training data set. Because it is in practice very difficult to model the nonlinear relationship between the multiple inputs and SIC, we have used a multilayer perceptron (MLP) neural network for the nonlinear estimation.

We also study the joint use of AMSR-2 radiometer data with S-1a data. This can also be seen as preparation for the SENTINEL-3 MWR data (23.8 GHz and 36,5 GHz).

The digitized FMI ice chart SIC fields have been used as training ground truth and reference data sets for the algorithm evaluation. The estimation results for test data sets from the two seasons will be presented and compared to earlier estimation results (winter 2012-2013) for RADARSAT-2 dual-polarized data. To evaluate the estimation results they will be compared to SIC of FMI ice charts and the 3.125 km ASI AMSR-2 radiometer SIC algorithm.

Cloudiness represents an important physical variable related to global climate change. This is particularly true over the Arctic where the signals of cloudiness alteration result stronger. In this study, the coupled spatial and temporal variability between the monthly anomalies of Cloud Fraction Cover (CFC) and Sea Ice Concentration (SIC) – one of the most important components of Arctic climate – is presented. CFC  and SIC data collected from SAF-CM [1] and NSIDC [2], respectively, for the period 1982 – 2009, are analysed using the Singular Value Decomposition (SVD) method [3].

Correlations between both variables and NAO [4], AO [5] and PNA [6] indices are estimated. The second and the third mode are significantly correlated with the NAO, the major mode of atmospheric variability affecting the Arctic [7].  A deepest seasonal analysis detects higher correlation in the first mode during winter and summer. The spatial patterns of the first mode show areas with the maximum covariance for SIC and CFC located over the Barents Sea and the Beaufort Sea, respectively.

[1]  Karlsson K.G., A. Riihelä, R. Müller, J. F. Meirink, J. Sedlar, M. Stengel, M. Lockhoff, J. Trentmann, F. Kaspar, R. Hollmann, and E. Wolters (2013). CLARA-A1: a cloud, albedo, and radiation dataset from 28 yr of global AVHRR data, Atmos. Chem. Phys., 13, 5351–5367.

[2]  Cavalieri, D. J., C. L. Parkinson, P. Gloersen, and H. Zwally (1996). Updated yearly Sea Ice Concentrations from Nimbus-7 SMMR and DMSP SSM/I-SSMIS Passive Microwave Data. Boulder, Colorado USA: NASA National Snow and Ice Data Center Distributed Active Archive Center.

[3]  Breterthon, C. S., C. Smith, and J. M. Wallace (1992). An Inter-comparison of Methods for Finding Coupled Patterns in Climate Data. J. Climate, 5, 541-560. 

[4]  Hurrel, J. W. (1995). Decadal trends in the North Atlantic Oscillation: Regional temperature and precipitation. Science 269,676-679.

[5]  Thompson, D. W. J. and J. M. Wallace (1998). The Arctic Oscillation signature in the wintertime geopotential height and temperature fields. Geophys. Res. Lett., 5, 9, 1297-1300.

[6]  Wallace, J. M. and D. S. Gutzler (1981). Teleconnection in the geopotential height field during the Northern Hemisphere winter. Mon. Weather Rev., 109, 784-812.

[7]  Serreze  M. C. and R. G. Barry (2014). The Arctic Climate System. 2^nd Ed. Cambridge University Press, New York.

Baltic Sea is well known for seasonal ice cover. Current study is focused on Gulf of Riga that is located in the eastern part of the Baltic Sea. Previous studies have shown that the ice conditions in Gulf of Riga can vary significantly from year to year depending on the weather conditions. Depending on the year the ice cover season starts between late November and middle January. The length of the ice season which can last until late April is in the range of 3-5 months.  In addition to interannual ice cover variations there are significant spatial variations between different Gulf areas.

The use of remote sensing methods enables to monitor ice extent during different winter scenarios. Although during the last years the emphasis in operational ice remote sensing has been on exploiting the capabilities of active sensors (e.g. SAR) the optical imagery can provide valuable information as well. Data from Moderate Resolution Imaging Spectroradiometer (MODIS) can be used for ice extent monitoring and for characterization of average winter conditions. We used MODIS data from visible range channels of spectrum with 250 m resolution (620 – 670nm ; 841 – 876 nm) to detect ice extent in the Gulf of Riga (Baltic Sea). In total 366images were used for ice extent detection.

After processing all the 366 images the average ice cover maps for different months and years were calculated. The ice cover probability maps were calculated which showed the percentage of time that each pixel was covered by ice. Based on the negative degree days, calculated from the data obtained in Kihnu meteorological station, the winter scenarios were defined. In case the sum of negative degree days (ºC day) is above 400 the winter was considered as severe (2003, 2006, 2010 and 2011). In case of medium (2004 and 2005) winters the corresponding value was between 200 and 400 and for mild (2007, 2008 and 2009) winters the sum of negative degree days was below 200.

Sea ice thickness is a key parameter for understanding changes taking place in polar regions and also for ship navigation and planning of off-shore operations in ice infested waters. In addition, the ocean-atmosphere heat, momentum and gas exchanges are controlled by the sea ice thickness distribution in the polar oceans. As in situ measurements of sea ice thickness are sparse and only represent points in time and space with limited spatial and temporal connections, it is important to have more frequent observations/models and to be able to estimate the uncertainties of these through intercomparison with the ground truth.

Currently Arctic sea ice thickness information is operationally available mainly as WMO ice type classification in ice charts provided by national Ice Services and as estimates by various ocean-sea ice models. The EC FP7 POLAR ICE project distributes in 2014-2016 three new sea ice thickness products for the Arctic based on satellite data for operational usage: (1) daily thickness map of thin sea ice in the entire Arctic (up to about 0.5 m) retrieved from brightness temperature measurements of the L-band microwave radiometer on the SMOS satellite, with a resolution of 30 to 40 km; (2) daily thickness maps of thin sea in the European Arctic (up to about 0.5 m) derived from VIIRS thermal imagery (subject to cloud cover) with resolution of 750 m, and (3) daily sea ice thickness maps (detected thin ice, and ice thickness 30 to 250 cm) of the Barents and Kara Seas derived from multisensor (AMSR2, SENTINEL-1) data and TOPAZ ocean-sea ice model at resolution of 1 km. All three thickness products are restricted to cold winter conditions (dry snow cover), roughly October to May in the Arctic. Within POLAR ICE, a state-of-the-art coupled ocean and sea ice model (ocean: HYCOM , sea ice: CICE, coupler: ESMF) runs twice a day in an operational mode and produces a 5 day forecast with a resolution of 10 km. Among other parameters, the model also predicts sea ice thickness. POLAR ICE integrates these ice thickness products with other sea ice information products derived from satellite and sea ice model data, so that end-users can readily visualise products, received in near real time, in synergy with one another.

Here we intercompare the three satellite-based ice thickness data sets and the output of the coupled ocean-sea-ice model, and validate them with available in-situ ice thickness data. The aim is to see how well the four sea ice data sets agree, and to find out where and under which conditions mismatches typically occur. The intercomparison of the SMOS and VIIRS thickness maps also allows to compare thin ice thickness estimations based on two different physical relationships; ice thickness vs. microwave emission for SMOS, and ice thickness vs. surface temperature for VIIRS. Comparison to validation data gives estimations of the absolute accuracy of the POLAR ICE thickness charts which is the main information required by end-users.

The intercomparison is conducted over the Barents and Kara Seas where all three products overlap. In addition, in the Barents Sea, East of Svalbard, a sea ice campaign was conducted in March 2014 (joint ESA SMOSIce campaign and test cruise of German IRO2 project) where sea ice thickness was measured with ship- and helicopter-borne electromagnetic induction (EM) instruments.

Coastal polynias are open water areas in the sea ice cover. They are often caused by strong katabatic winds that push the ice offshore.  While the ice around the polynia is deformed, new ice formation takes place in the polynia itself. The new ice consolidates at the border to the surrounding pack ice. This highly dynamic regime manifests itself by a high variability of sea ice characteristics. Since the variations of ice characteristics have a strong influence of the heat exchange with the atmosphere as well as on brine release and ice export, it is important to understand atmosphere - sea ice – ocean interactions close to the polyias in a more comprehensive way.

One major goal of this project is to study the potential of the upcoming Sentinel satellites for polynia research. While not all characteristics of the sea ice can be monitored in a direct way, the combination of satellite sensors and derived products can contribute to a better understanding and reduction of uncertainties. Temperature sensors such as Envisat AATSR can provide implicitly information on the ice thickness or at least thermal conductivity: thin ice and open water will show higher surface temperatures than thick ice. SAR images offer insights into the structure of the sea ice: ridges and deformation zones are highlighted in the ALOS PALSAR L-band image; Envisat ASAR C band images produce a better representation of the individual floes. Optical images from RapidEye, Modis or Landsat provide an impression in natural color, which might allow identifying snow covered regions and help to improve the interpretation of data from the other sensors.  By taking into account sea ice drift derived from a satellite image time series it is possible to identify e.g.  fast ice regions where the ice does not move. We present a hierarchical classification based on Envisat ASAR, ALOS PALSAR and Envisat AATSR and compare it to an unsupervised ISODATA classification. We demonstrate the effect of drift information on the classification and its advantage for the separation of sea ice regimes with different dynamic characteristics. As an example we present a polynia event in the Terra Nova Bay, Antarctica.. The classification is supplemented by sea ice drift fields derived from SAR time series, wind data from AMPS model simulations from the State University of Ohio, and AWS measurements from the Italian Antarctic research program (PNRA) and the University of Wisconsin.

In order to extend our studies on sea ice dynamics in and around polynjas we acquired extensive time series of TerraSAR-X images over two test sites in Antarctica during austral spring 2014. Those observations were supplemented by data of other SAR-Satellites and by images of optical sensors on Terra, Aqua, Landsat and RapidEye. Based on our anaylsis of the 2009 data we will present first results for the time series acquired 2014.

A coastline can be interpreted as the line where land and water bodies are in contact. Due to its dynamic nature it is not static and worldwide coastlines are rapidly changing due to natural physical processes and human activities. In addition, global climate change and glacial isostatic adjustment have huge impacts on coastline variability. Therefore, an exact delineation of the coastline is required to describe the physical form of the coastal zone where interaction processes happen.

In case of the Antarctic continent this task is not very simple due to its huge variablity the whole year round. Ice-cover, snowmelt, very low temperatures, day night cycle and other considerations have to be kept in mind when mapping has to be carried out. Navigation and exploration depend on information about position, shape and orientation of the Antarctic coastline. Furthermore, to understand the response of the Antarctic ice sheet to climate change, an accurate coastline map is essential.

Except the relevant cartographic products for navigation, map products for the baseline of the Chilean Antarctic Territory are obsolete. Therefore we carried out an extensive study aiming coastline estimation based on SAR observations and Digital Surface Models (obtained from TerraSAR-X and TanDEM-X stereo pairs). Furthermore we considered in-situ field surveying and ground control points determination by GNSS technologies and high resolution spaceborne optical imagery. The combination of all these techniques allow us to extract a complete coastline, with adequate precision for modern global geodetic datums, like WGS84, which are compatible with global navigation based on artificial satellites.

In particular, we studied a combination of the 2013/2014 SAR observations acquired during summer months, for 15 specific study sites between 53°W and 90°W of the Chilean Antarctic Territory with special attention on outermost islands. Our semi-automatic approach is based on a sequence of modern techniques for image enhancement and edge detection. In particular, image segmentation and classification are applied to differentiate between ice bodies (continental or floating) or rock exposures and open water. The segmented images are then transformed into vector based products to define the final coastline. Finally we compare the extracted coastline with the in-situ observations and the high resolution optical images. Unfortunately acquisition times of optical images and SAR differ. Nevertheless, the comparison lets us assume that the results are consistent as horizontal mean difference between DSM and optical images is less than 100m (except Charcot Island where horizontal accuracy was about 200m). This processing chain allows straightforward processing of DSM for coastline extraction, which is almost independent of the study area.

In conclusion, all of the results fully comply with technical aspects in relation to map generation. For cartographic and scientific applications, the Ground Sampling Distance (10m) of the SAR observations is adequate. Procedures and methodologies were adjusted which are in state of the art of geodetic surveying and modern image processing. This DSM based, and supported by high resolution optical imagery and GNSS observations, coastline allows to describe with high level of confidence the geometric shape of the study area and the boundary between the continental ice or rock exposures and the ocean.

In order to interpret ice information products correctly and efficiently, it is important to have the information presented in a consistent way. There are standards and recommendations for some ice parameter (like ice concentration), but not for all (e.g. ice thickness). In a system where the user can access many kinds of Near-Real-Time sea-ice related products from different providers, this issue is prominent. The challenge is how to present these products in a way that is consistent and non-confusing, thus fulfilling the requirement that the colour scheme should be harmonised per physical parameter. Also different physical parameters should preferably use different colour schemes so that the user immediately recognises which parameter is displayed on the screen.  Taking into account different production environments on the producers’ side, an adaptive mechanism has been implemented to ingest the products in a flexible way. The principle adopted is to use a layered approach where the values in the raster products are mapped to either explicit categories or to ranges of physical values. The physical values are then mapped to colour ranges thus ensuring that the same physical value is shown in the same colour independent of the origin.  The approach also enables changing of the colour mapping for all products showing the same physical parameter, in a simple and consistent way. Changing of the colour mapping is needed for optimal display of values in subareas or varying seasons, when – for example – the ice thickness range differ.

The paper concentrates on the visualisation issue when dealing with products from multiple providers but also describes the architecture of the Polar Ice system which is designed and implemented as part of the ongoing FP7 Polar Ice project . The building blocks, interfaces and the general principles of the user interface in Polar Ice are presented as well as feedback gathered from the users of the system during focused demonstrations.

Ice caps and glaciers are important present‐day indicators of the ongoing global climate change, and they are the component of the cryosphere currently making the largest contribution to sea level change (Vaughan et al., 2013). Arctic ice caps are of particular interest as air temperatures at high northern latitudes are increasing more rapidly than elsewhere (Serreze & Francis, 2006).

The Flade Isblink Ice Cap (FIIC) is situated in Kronprins Christian Land, Eastern North Greenland (Figure 1). With an area of 8 500 km², it is the largest independent ice cap in Greenland (Kelly and Lowell, 2009). The northern part of the ice cap is drained by two main outlet glaciers flowing to the Northwest, which have experienced several cycles of advance and retreat. While previous remote sensing studies of the FIIC outlet glaciers have observed temporal variations in ice velocity and anomalous changes in surface elevation, the mechanism behind the changes remains unknown. Using results from satellite InSAR, Joughin et al. (2010) reported that the two largest FIIC outlet glaciers slowed to sometime between 2000 and 2005. Rinne et al. (2011) showed that while the average surface elevation change rate of the FIIC was near zero (0.03 ± 0.03 m yr^-1) between 2002 and 2009, rates of ice thickening in the NW part of the ice cap were up to 3.4 ± 0.7 m yr^-1 during the period 2004 to 2008.

Additional observations are required to improve our understanding of the unusual dynamics exhibited by the FIIC, and to better predict the future contribution of Arctic ice caps to global sea level.

To measure past ice flow velocities, InSAR data acquired by ESA’s ERS satellites during the Tandem mission phase during winter 1995 were used (Palmer et al., 2010). The time series of ice flow observations was extended to the present by using the MEaSUREs dataset (Joughin et al., 2010) and ESA Sentinel-1 data. These results are shown in figure 1; changes in the direction as well as speed of ice flow can be seen.

While the data presented above helps to describe the changes in the dynamics of the outlet glaciers, they do not reveal the underlying mechanisms of change. The presence of subglacial water can profoundly alter the dynamics of ice flow, and changes in the configuration of the hydrological system at the bed are likely to have played a role in the change in ice flow between 2000 and 2005. Evidence that water flows periodically at the bed of the ice cap is provided by recent satellite observations (Willis et al., 2015) which have revealed the presence of a surface collapse basin near the summit of the ice cap. This feature has been interpreted as being the surface expression of a drained subglacial lake, examples of which have also been observed recently beneath the main Greenland Ice Sheet (Palmer at al., 2013). Given the relatively thin ice (max 535 m) and low annual mean air temperature, it is unlikely that temperature at the ice base reaches the pressure melting point, so the subglacial lake must be fed by surface meltwater.

Although the mechanism behind the observed change in ice dynamics is not yet known, ESA’s Sentinel satellites have the potential to provide the high resolution geophysical data required to improve our understanding of the FIIC and other Arctic ice caps.

References

Joughin, I., et al (2010). Greenland flow variability from ice-sheet-wide velocity mapping. Journal of Glaciology, 56(197), 415-430.

Kelly, M. A., & Lowell, T. V. (2009). Fluctuations of local glaciers in Greenland during latest Pleistocene and Holocene time. Quaternary Science Reviews, 28(21), 2088-2106.

Palmer, S. J., et al (2013). Greenland subglacial lakes detected by radar. Geophysical Research Letters, 40(23), 6154-6159.

Palmer, S. J., et al (2010). InSAR observations of ice elevation and velocity fluctuations at the Flade Isblink ice cap, eastern North Greenland. Journal of Geophysical Research: Earth Surface (2003–2012), 115(F4).

Rinne, E. J., et al (2011). On the recent elevation changes at the Flade Isblink Ice Cap, northern Greenland. Journal of Geophysical Research: Earth Surface (2003–2012), 116(F3).

Serreze, M. C., & Francis, J. A. (2006). The Arctic amplification debate. Climatic Change, 76(3-4), 241-264.

Vaughan, D.G., et al (2013). Observations: Cryosphere. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA

Willis, M. J., et al (2015). Recharge of a subglacial lake by surface meltwater in northeast Greenland. Nature, 518(7538), 223-227

The CryoSat mission is designed to determine fluctuations in the mass of the Earth’s land and the marine ice fields. Its primary payload is a radar altimeter that operates in different modes optimised depending on the kind of surface: Low resolution mode (LRM), SAR mode (SAR) and SAR interferometric mode (SARin). This radar is named SIRAL: Synthetic aperture interferometer radar altimeter [1].

Transponders are commonly used to calibrate absolute range from conventional altimeter waveforms because of it characteristic point target radar reflection. The waveforms corresponding to the transponder distinguish themselves from the other waveforms resulting from natural targets, in power and shape. 

ESA has deployed a transponder available for the CryoSat project (a refurbished ESA transponder developed for the ERS-1 altimeter calibration). It is deployed at the KSAT Svalbard station: SvalSAT. Another  transponder has been deployed by Technical University of Crete  for the Sentinel 3 calibration in the island of Crete.

We are using the transponder to calibrate SIRAL’s range, datation, and interferometric baseline (or angle of arrival) to meet the missions requirements [2]. In these calibrations, we are using 3 different type of data: the raw Full Bit Rate data, the stack beams before they are multi-looked (stack data) in the Level 1b processor, and the Level 1b data itself [3].

Ideally the comparison between (a) the theoretical value provided by the well-known target, and (b) the measurement by the instrument to be calibrated; provides us with the error the instrument is introducing when performing its measurement [4]. When this error can be assumed to be constant regardless the conditions, it will provide the bias of the instrument. And if the measurements can be repeated after a certain period of time, it can also provide an indication of the instrument drift.

This poster presents the analysis and results of this calibration with the Baseline C. The work presented here was initially carried out under an ESTEC/ESA contract, to calibrate CryoSat-2 during the Commissioning phase. It was later extended with an ESRIN/ESA contract, for continue monitoring and including further analysis.

[1] C.R. Francis, “CryoSat Mission and Data Description”, CS-RP-ESA-SY-0059.

[2] CryoSat Science and Mission Requirements Document, CS-RS-UCL-SY-001.

[3] D.J.Wingham, et al.: “CryoSat: A mission to determine the fluctuations in Earth’s land and marine ice fields”, Advances in Space Research 37 (2006) 841–871.

[4] SIRAL2 Calibration using TRP: Detail Processing Model – DPM; ISARD_ESA_CR2_TRP_CAL_DPM_030.

Extensive gravity anomalies associated with lateral density gradients and slow vertical motions of the Earth’s crust result in spatial variations of the ambient atmospheric pressure and, on a long-term basis, influence the local climate. Gravitational deflection of the local vertical in flat Arctic areas with irregular gravity attains 10’’ and the horizontal component of surface gravity is characterized by an acceleration of 0.5 mm/s². This persistent forcing induces low-level air movements towards the centre of positive gravity anomalies, or away from the centre of negative anomalies (see Annex). Under calm weather conditions lasting for several hours the velocity of centripetal flows can reach 3-4 m/s, which is close to the threshold wind speed for snow drifts. Air convergence at the Earth’s surface leads to upward vortex motion of air masses with subsequent cooling, condensation of water vapour, cloud formation and local precipitation, e.g. in the form of snow, and centripetal transport of loose snow along the solid surface may result. The adjacency of negative anomaly facilitates the effect of air uplift and gradient precipitation in the area of positive gravity anomaly. The whirling movement of the near-surface air about the centre of positive gravity anomalies is responsible for the regular shape of ice caps emerging on the flat ground and enhances snow accumulation on top of them. This provides some hints on the spatial asymmetry in the distribution of insular ice caps, their changes and ice flow pattern in the Eurasian Arctic regions with synoptically similar conditions.

The gravigenic non-orographic concept of gradient precipitation and ice cap evolution has yet to be verified and, in the present paper, we try to provide more substantial EO-based evidence to justify the hypothesis about the directional dynamics of Arctic ice caps in anomalous gravity fields. ERS, TDX and Sentinel-1A SAR interferometric models calibrated with ICESat and CryoSat-2 altimetry data were successfully applied to mapping and quantifying glacier elevation changes and ice flow on fifty insular ice caps in the northernmost region of the Eurasian Arctic ranging from Svalbard in the west to the De Long Islands in the east. The resultant maps were collocated and compared with available data on the Arctic gravity field and 25-year long records of daily precipitation obtained from 57 coastal meteorological stations. Free-air gravity anomalies were graphically represented in the reference model geometry using Russian gravimetric maps 1:1000000 (1980s), ArcGP grid (2008) and GOCE gravity field data (Release 3, 2009-2011 and Release 5, 2009-2013).

The main interrelationships determined for topographically smooth and open areas with a very cold, dry climate were summarized as follows. Lateral variations of gravity directly influence the ambient lapse rate, thereby modulating the atmospheric stability and increasing the intensity and frequency of heavy snowfalls over the areas with positive gravity anomalies. Most ice caps are situated within or in the close vicinity of positive gravity anomalies, while the majority of glacier-free islands and peninsulas with relatively large surface areas and significant top heights are situated in areas of low gravity. The main loss of land ice of up to 3.1 km³/a occurred in the seaward margins of ice caps, especially at the fronts of fast-flowing tidewater outlet glaciers and in the area of ice tongues or bridges. The highest rates of positive elevation changes of about 1 m/a were observed in the accumulation areas of large slow-moving ice caps in the vicinity of strong positive gravity anomalies. Passive ice caps were typically found in the areas with normal gravity.

On most ice caps with a regular shape on a flat bed the orientation of the ice flow pattern generally followed the directions of the maximum gravity gradient, and main ice divides were nearly parallel to anomaly isopleths. Sometimes, depending on the mutual configuration of gravity anomaly and glacier morphology, outlet glaciers flowed towards centres of positive gravity anomalies. The elongated areas of snow drift detected in radar coherency images were located within 50 to 80 kilometres distance from the centre of the dominant positive gravity anomaly and were oriented towards the anomaly’s centre. No snow drift was observed between the ice caps. The correlation between the gravity gradient and glacier dynamics decreased in mountainous and humid areas, as e.g. in the south-western part of Svalbard. The aggregate of EO-based indications and contraindications on glacioclimatic trends in Arctic regions with anomalous gravity derived with the aid of satellite gradiometry, altimetry and interferometry (GAIN method) is a meaningful contribution to the theory of glacial isostasy, retrospective glaciological reconstructions and long-term climate projections as it provides a better understanding of the physical relationships between the Earth’s gravity, glaciation and climate.

The spillage of oil in Polar Regions is particularly serious due to the threat to the environment and the difficulties in detecting and tracking the full extent of the oil seepage beneath sea ice.

Currently oil which has been spilt in the ice affected water, tends to disappear beneath the ice and it is not possible to detect and monitor where it has travelled or indeed at which point it will emerge from the edge of the ice into open sea. This poses very significant challenges to any extenuating activities until well after the emerging oil has been located. This project employs cutting edge technology to 'see' the oil beneath the ice and to monitor its spread. This will allow mitigation technologies such as floating restraint barriers and clean-up operations to be ready precisely at the point the oil emerges.

Our study presents results of lab-based trials that eventually leads to the field deployment of hyperspectral imaging technology (HSI) in Polar Regions to detect the oil seepage underneath sea ice.

Hyperspectral imaging is an emerging technology which uses a new type of camera to capture the spectral signature of a scene. It is the next stage in the logical extrapolation of the technology development which took us from monochrome (black and white) imaging to colour imaging. However, whereas these two image types can be viewed by the human eye, hyperspectral images contain detail beyond human vision.

In monochrome imaging each pixel has a single brightness value, in colour imaging each pixel has three values (red, green, blue) associated with it, but in hyperspectral imaging each pixel has a whole spectrum of values. With the correct processing this spectrum can provide valuable information about the subject being viewed such as its moisture, temperature and even chemical content.  Each material reflects a different spectrum or 'spectral signature'. It is well known that the spectral signature of oil is very different from the spectral signature of sea ice.

There are several different types of hyperspectral imaging technology and application of each of them may be beneficial for different tasks. Depending on the type of underlying technology, the cameras can 'see' a different depth into the ice and determine different properties of objects.  Within our study we focus on quantifying the limits of detection of oil beneath ice with two technologically and spectrally different hyperspectral imagers. The goal of this research is to answer the questions of what is the maximum thickness of ice that hyperspectral imaging can detect oil and what is the minimum concentration of oil that can be detected with different HSI systems.

Equipped with state-of-the-art facilities, equipment and signal processing tools, this study provides reliable overview of HSI capabilities for oil spill detection beneath sea ice and by understanding the limits of this detection we will have the knowledge to design a system to be deployed in the field.

Retreat and thinning of Antarctic ice shelves triggers acceleration of tributary glaciers and can potentially cause a significant mass deficit of the Antarctic Ice Sheet with a consequent influence on global sea level. Our current understanding of these processes is largely developed based on observations in the Antarctic Peninsula and West Antarctica, which may not be applicable to East Antarctica.

Dronning Maud Land (DML, 30W – 45E), the Atlantic and Indian Sea sector of East Antarctica, is characterized by many, relatively small ice shelves distributed along its ~1500-km-long coast. Many of these are punctuated by grounded features (ice rises and rumples) that have a significant impact on the ice shelf stability, but details of this mechanism remain poorly known. Possible dynamic coupling between adjacent ice shelves further complicates their sensitivity to climatic and oceanic changes.

In autumn 2014 we started a new project utilizing a range of satellite observations to document the current status (calving positions, surface features, flow speeds, freeboard heights) of the DML ice shelves and estimate their current and past mass balance. We first analyzed Radarsat-2 ScanSAR data collected in the central DML (15W-40E) and compared the ice shelf edge positions with other datasets. It was found that the calving front of most ice shelves appears to be stable over the past decade. Only 8 out of 31 fronts retreated considerably due to calving of icebergs larger than 10km².  Using a Radarsat-2 Fine Wide campaign dataset collected between November 2014 and January 2015, visible features are mapped as proxies of flow stripes, rifts, and crevasses. These images also form the main dataset for a current effort to develop a semi-automated speckle-tracking scheme to extract flow fields of the DML ice shelves. As the vast majority of the coastal DML is subject to movement, we make use of outcrops, ice-divides and slowly moving ice (< 5m/yr based on Rignot et al. 2011) to facilitate accurate co-registration. Once robustness of this scheme is assessed, we will apply the same scheme for ERS, Envisat and ALOS data to retrieve ice flow fields from the past decade(s). The scheduled semi-annual Sentinel 1 campaigns over DML will bring an excellent opportunity for us to continue monitoring the ice flow in the years to come. Ultimately, we will be using our ice flow products together with IceSAT/Cryosat altimetry data and surface mass balance model runs to calculate mass balance of the DML ice shelves.

Glaciers and ice sheets or caps are sensitive indicators of global climate change. Together with thermal expansion it is responsible for sea level rise.At present, ice bodies cover approx.10 % of the Earth’s land surface, majority of which is present inthe polar regions, which is equivalent to three quarter of the world’s total fresh water resources. The retreat of the Antarctic and Greenland ice sheets, as well as that of glaciers worldwide, is a major consequence of global warming.During the last three decades various ice shelves on the Antarctic Peninsula have retreatedand disintegrated. The Larsen B Ice Shelf, which disappeared in 2002 in just over 30 days, is only one example.Several studies around the globe have also been indicated disintegration of differentice shelvesof Antarctic Peninsula such as Prince Gustav Channel (1995), Larsen A (1995), Larsen Inlet (1999), Larsen B (2002) and the Jones Ice Shelf (2003).In 2009the ice bridge which connected the Wilkins Ice Shelf and Charcot Island collapsed.

On the west coast of the Antarctic Peninsula, George VI Shelf is embedded in George VI sound, surrounded by Bellingshausen Sea. Recent observations have shown that the northern and southern boundaries of the George VI Ice Shelf suffer under constant retreat for decades. Therefore a systematic analysis of George VIice shelf and its tributaryglaciersis a paramount interest in order to understand theirbehaviour and environmental interaction.

Ice dynamics,and in particular, glacier velocity is an important control on ice discharge rate. Deriving surface velocity fields of glaciers and ice shelves using satellite images is an efficient,low-costmethod, which has been used since the mid 1980s.

Remote sensing techniques provide an efficient way to collect dataofremoteregions due to its large spatial coverage.Fieldsurveys, time-lapse camera surveys, satellite image feature tracking of optical and SAR data and SAR interferometry (InSAR) are techniques by which velocity of the motion may be determined. However, due to the positional disadvantage of glaciers,fieldsurvey commonly results in sparse spatial coverage. Moreover due to loss of coherence and unavailability of finer resolution DEM, velocity measurements by InSARcannot produce high accuracies. On the contrary, offset tracking in SAR imagery is more useful for measuring flow velocities over longer periods.

Therefore, in this study an attempt has been made to derive the velocity field of some tributary glaciersof George VI ice shelfby SAR feature tracking technique using ERS-2 C-band data. Velocity fields have derived within 35 days temporal interval over the period from 2002 to 2011. The SAR data have been selected during the ablation period. Before feature tracking, both SAR images have been co-registered within an accuracy of 1/5^th of a pixel usingcross correlation intensity algorithm.After the co-registration the estimation of the shift between displaced features have been estimated. Velocity in the range fromapprox. 10 to 70 meter during 35 days have been obtained for some of the tributaries of George VI sound. However, unrealistic velocity profile has been observedalong the main George VI sound dueto the unavailability of sufficient displaced features.Though this technique has gained its reliability but due to the presence of ionospheric disturbances sometimes erroneous offset estimation in azimuth direction can be identified. It has been observed that longer wavelength radar signals are mainly affected by such ionospheric disturbances. It is thus assumed that the velocity estimate can be refined by using TerraSAR-X data which are less affected by ionospheric disturbances.

Sea ice motion strongly influences the distribution of sea ice on different spatial and temporal scales. The coverage and quality of high-resolution Synthetic Aperture Radar (SAR) images has strongly improved during the last years, which allows to derive sea ice motion with high spatial (1 km) and temporal resolution. Such high-resolution ice drift data are not yet provided. The approach is to exploit recent improvements in computer vision by using state of the art feature tracking algorithms. A computational efficient, open source feature tracking algorithm, called ORB, is adopted and tuned for sea ice drift retrieval from Sentinel-1 SAR images. The best suitable setting and parameter values have been found using four representative Sentinel-1 image pairs. A new quality measure for feature tracking algorithms is introduced utilising the distribution of the resulting vector field. The performance of the algorithm is compared with two other feature tracking algorithms (SIFT and SURF). Applied on a test image pair acquired over Fram Strait, the tuned ORB algorithm produces the highest number of vectors (6920, SIFT: 1585 and SURF: 518) while being computational most efficient (66s, SIFT: 182s and SURF: 99s using a 2,7GHz processor with 8GB memory). For validation purpose, 350 manually drawn vectors have been compared with the closest calculated vectors and the resulting root mean square distance is 609.9m (equivalent to 7.5pixel). All test image pairs show a significant better performance of the HV channel. On average, around 4 times more vectors have been found using HV polarisation. All software requirements necessary for applying the presented feature tracking algorithm are open source to ensure a free and easy implementation. Comparison with a pattern matching algorithm from AWI has shown the respective advantages and upcoming work will include the combination of the 2 algorithms.

Microwave sensors onboard polar orbit satellites are commonly used for sea ice monitoring at high latitude : radiometers are routinely used for this application and scatterometers have also shown they can contribute significantly to it. Since 1991, numerous scatterometers data at C and Ku-bands are available since ERS-1 with NSCAT, QuikSCAT, OSCAT, HY, and ASCAT sensors. Here, we will show how the scatterometers sensors could be used to build Earth Observation data time series for sea ice monitoring for both Arctic and Antarctic areas.

Backscatter data enable to discriminate sea ice from open ocean areas, in particular scatterometers are useful to detect new ice, even at the early stage of growth which is not possible using radiometers.Backscatter data from scatterometer can also be used for sea ice type detection (first year from multi-year sea ice in the Arctic), results and differences between C-band and Ku-band sensors results will be presented.

Moreover, sea ice displacement maps can be built in central Arctic from backscatter data, examples will be shown using the merging of scatterometer and radiometer data. We will also focus on the benefit of the use of the two ASCAT sensors that are presently available (onboard MetOp-A -since 2007 and -B, since 2012), in particular for the Antarctic area.

This presentation will enhance i) the need of scatterometer data for sea ice application with many examples of the inferred parameters in particular from the U.S. QuikSCAT/SeaWinds and the E.U. MetOp/ASCATs scatterometers ii) the need of the continuity of scatterometers missions and iii) the benefit of the combination of sensors and datasets (scatterometers, scatterometers with radiometers) for a long-term observation of the polar areas.

These data are routinely processed at IFREMER/CERSAT and available for the scientific community. The ASCAT calibrated reprocessed data (2007-2014) are now available and have been processed at IFREMER/CERSAT, added to the SeaWinds/QuikSCAT dataset (1999-2009), they will provide an exceptional basis for future analysis and synthesis of long-term variations of the sea ice in both Arctic and Antarctic polar areas.

Modeling the Dynamic Ocean Topography (DOT) and determining the circulation pattern of the ocean currents has always been a challenging task in the Mediterranean Sea. This is mainly due to a) the limited number of satellite altimetry data in the area, given land intrusion by the isles and islands, b) the limited number of gravity-related data, so that the geodetic determination of the DOT was deemed as low-accuracy and c) the nature of the Mediterranean circulation itself, being mainly the result of small-scale gyres and eddies as well as cyclones and anti-cyclones. The only satellite altimetry missions that offered a deeper insight to the Mediterranean DOT and circulation where the Geodetic Missions (GM) of GEOSAT and ERS1, which offered dense cross-track resolution, but with lower-accuracy compared to today’s altimetric standards. To tackle the aforementioned limiting factors, two pillars need to be addressed, i.e., improved representations of the long and medium wavelength geoid information and higher accuracy and resolution satellite altimetry data. The mission of GOCE treats the first part of the equation, offering a 1-2 cm geoid accuracy to harmonic degrees 220-240 (~82-90 km). At the same time, ESA’s Crysosat2 mission, given its orbit characteristics and measurement modes, offers a very dense (~7-8 km) cross-track resolution with an accuracy of ~1.5-2 cm for the LRM. Cryosat gives a 5.7-5.8 cm std differences at the crossovers, which is at the same level as that of ENVISAT and Jason2, therefore it can be used for sea level anomaly, geoid and DOT modelling, especially in closed seas like the Mediterranean. The main scope of the present study is to utilize the satellite altimetry data of Cryosat2, from mission start in July 2010 up to the end of 2015 (Cycles 4 to 73) and the latest DIR-R5 and TIM-R5 GOCE-derived global geopotential models in order to determine the DOT and circulation for the Mediterranean Sea. For the DOT determination, filtering is first investigated to model and remove/reduce the effects of the geoid omission and commission errors. Spatial filtering is based on Gaussian, Wiener, cosine arch and boxcar filters, while the spectral filtering was carried out with wavelet (WL) Multi-resolution analysis (MRA). Based on the spectral properties of each WL level relative to a ground-truth DOT, some of these levels were omitted while others have been filtered based on selective filtering. Within the entire DOT evaluation, the RIO_MED synthetic DOT model for the Mediterranean was used as ground-truth, both for the WL-MRA filtering and in order to assess the GOCE-derived DOT models. Other DOT models, like the standard SMDT-MED-2014 and MDT_CNES-CLS13 used for the reduction of altimetric observations by AVISO/CNES have been validated as well.

The Antarctic Peninsula is one of the world`s most affected regions by Climate Change. Long-term remote sensing time series enable to study changes and to reveal information on the underlying processes of the cryosphere as well as the interlinkages with the atmosphere.

The German Antarctic Recieving Station (GARS) at O'Higgins operated by the German Remote Sensing Data Center (DFD/DLR) has acquired data from the two European Space Agency (ESA) European Remote Sensing satellite mission (ERS-)1/2 between 1991 and 2011. Data of other space borne SAR sensors such as ESA`s ENVISAT ASAR, JAXA`s (Japan Aerospace Exploration Agency) ALOS PALSAR, RadarSat-1, DLR`s TerraSAR-X and TanDEM-X or the European mission Sentinel-1 will complement to a dense time series of SAR measurements from the 1990s until today for several regions of the Antarctic Peninsula.

Differential interferometric synthetic radar (DInSAR) methods and intensity tracking are applied inorder to derive important glaciological parameters such as grounding line positions, glacier velocities, surface elevations, ice mass fluxes and glacier mass balances. Additionally, calibrated SAR amplitude images as well as images taken by optical sensors (e.g. Landsat) are used to map glacier extends and to compute changes of glacier areas.

We represent first results of a case study at the Wordie Ice Shelf, located at the south-western side of the Antarctic Peninsula. This ice shelf disintegrated in a series of events during the 1970s and 1980s, so that already in the beginning of the 1990s only disconnected and retreating tidewater glaciers remained. Due to the loss of the buttressing effect of the ice shelf, an increased ice mass discharge has been observed. An increase of flow speeds and elevation decrease have been reported by previous studies – mainly on a bi-temporal basis. However, how long and how exactly in time this process of adaption to the new boundary conditions will last as well as how much ice mass loss and sea level rise is caused by this process is yet not well known. Thus we use dense SAR time series in conjunction with data on surface elevation from photogrammetry and laser/radar altimetry, ground penetrating radar as well as surface mass balance simulations to target more precise estimates as well as data sets that can be better compared with large-scale observations by the GRACE gravimetry mission.

We propose a method to reduce the error generated when computing sea ice deformation fields from synthetic aperture radar (SAR)-derived sea ice motion. The method is based on two steps. The first step consists of using a triangulation of the positions taken from the sea ice trajectories to define a mesh on which a first estimate of sea ice deformation is computed. The second step consists of applying a specific smoother to the deformation field to reduce the artificial noise that arises along discontinuities in the sea ice motion field. This method is here applied to RADARSAT Geophysical Processor System (RGPS) sea ice trajectories having a temporal and spatial resolution of about 3 days and 10 km, respectively. From the comparison between unfiltered and filtered fields, we estimate that the artificial noise causes an overestimation of about 60 % of opening and closing. The artificial noise also has a strong impact on the statistical distribution of the deformation and on the scaling exponents estimated with multifractal analysis. We also show that a similar noise is present in the deformation fields provided in the widely used four-point deformation RGPS data set. These findings may have serious implications for previous studies as the constant overestimation of the opening and closing could lead to a large overestimation of freezing in leads, salt rejection and sea ice ridging.

Satellite based snow cover monitoring is typically performed using optical, active microwave (SAR) and passive microwave sensors. Optical and active microwave sensors are commonly used when estimating snow covered area, whereas passive microwave satellite sensors are suitable for snow water equivalent retrieval in dry snow conditions. A large portion of areas which experience seasonal snow cover are located in the boreal forest region. Without considering the effects of forest canopy on the observed signal, snow retrieval results with all sensor types in these areas would be biased. Various models describing the interaction of electromagnetic radiation with the forest canopy have been developed, but many of these are overly complex with high computational and ancillary data requirements. For operational snow retrieval purposes, simple, invertible models are preferred.

This work aims at increasing the understanding of the effect of forest canopy on remote sensing observations of snow-covered terrain for both microwave and optical regimes, and at quantifying the capability of simple, zeroth-order models in simulating these effects. To achieve these goals, a spatial analysis of various remote sensing data in the northern boreal forest region of Finland was performed. The forest contribution was first analyzed for an airborne optical sensor, then for X- and Ku-band airborne active microwave sensors, an X-band space borne sensor, and finally for airborne passive microwave sensors ranging from X- to Ka-bands. Model parameters for vegetation transmissivity as well as the properties of the underlying surface were optimized by utilizing LiDAR- and Landsat based simplified proxy parameters describing forest canopy closure and stem volume.

The results of this study demonstrated that despite using these relatively simple forest proxies, an optimized zeroth-order model can estimate the extinction properties of electromagnetic signals in a forest canopy to a high degree of accuracy. In particular, model estimates showed a high degree of correlation against HUTRAD passive microwave and AISA airborne optical data, with r² values ranging from 0.7 to 0.9 depending on the observation spectral band and the day of observation (not including 10 GHz band in passive microwave and NIR band in optical). Moreover, in all sensor types, the model estimate for ground/snow contribution was very close to the values observed for open terrain. The SAR model successfully estimated the median of the observations i.e. bias was small, but compared to the optical and passive microwave models, a larger scatter of the observations, typical to SAR sensors, was reflected by higher RMSE and lower correlation with the model.

Forward models applied in retrieval algorithms need to be relatively simple in nature due to both computational and ancillary data restrictions. Compared to higher order models, the presented zeroth order models satisfy both of these requirements. Furthermore, it was shown that a single parameter or proxy value describing the forest properties (stem volume or canopy closure) provided the necessary correlation with the observed signal dynamics, required to fit the models to the observations. This is particularly encouraging as it implies that forest canopy effects can be regulated in retrieval algorithms even if complex multiple scattering effects are neglected. Furthermore, compensation for canopy effects can be made using data which is already available, or can be retrieved from other remote sensing observations.

The Mission Control System (MCS) in use to operate the CryoSat-2 satellite from the European Operation Centre (ESOC) since launch (April 2010) is based on the ESOC develop Spacecraft-Operations-System release 4 (SCOS 2000 or S2K).

This software was initially designed to run on a SUN-Solaris platform and in particular on a Solaris-8 Operating System. Since the end of the GOCE mission (Nov. 2013) CryoSat-2 is the only mission in ESOC still based on such configuration and is no longer covered by the general infrastructure support thus requiring a dedicated maintenance contract.

In order to support operations during the current and possibly future extended phases, a plan on how to overcome hardware and operating system obsolescence and shortcomings in availability of software maintenance had been prepared and deployed.

This poster presents the different possibilities which were considered for the first specific case of CryoSat-2 MCS and which have been selected and implemented.

Emphasis is given to the difficulties encountered during the porting exercise. Two areas in particular: the progressive loss of expertise specific to the CryoSat-2 MCS and the rapidly decreasing availability of hardware replacements, on stock and on the market, as well as a problematic situation with the maintenance of the Solaris-8 OS.

Particular attention is given to the migration to a virtualised environment running on a hardware common to several ESOC missions.  This solution allowed reducing the risk of physical hardware obsolescence, increasing efficiency and exploiting inter-mission synergy.

The technical implementation challenges are presented together with the testing and validation methodologies undergone.

The transition phase to the new system is also documented to describe the interesting approach developed to minimise the impact on a flying mission in terms of both spacecraft operations and ground interfaces between the operations centre and the external parties.

The AltiKa altimeter on-board SARAL was launched by CNES and CSIRO in February 2013 on the same 35-day orbit than previous European altimeters, Envisat and ERS1-2. The altimeter operates in Ka-band, a higher frequency than previous altimeters operating at Ku-band frequency which leads to decrease radar penetration within the snowpack. Launched in 2010 by the European Spatial Agency, Cryosat-2 is a Ku-band radar altimeter operating mostly in SAR mode over sea ice.  

Radar altimeters at Ku-band frequency have been used for more than a decade to estimate sea ice freeboard and make pan-Arctic sea ice thickness estimates. Despite recent improvements of the freeboard technic, informations such as penetration depth at Ku-band fequency and snow depth are missing to make better estimations of ice-thickness.

The collocated information of multi sattelite altimetric radar operating at different fequencies over Arctic sea ice provides thus a unique opportunity to monitor snow depth properties. In particular, the analyzis of height difference at crossover points between SARAL-AltiKa and Cryosat-2 together with Operation Ice Bridge snow depth data provides informations on both Ku-band and Ka-band radar penetration depth  as well as a proxy for snow depth estimates.

The study is supported by TOSCA CNES SICKAyS funding

During the last few years, several studies have highlighted the sensitivity of L-band observations to snow-covered areas. In Antarctica, a very good knowledge of the sensitivity is essential to develop applications and contribute to a better understanding of the ice sheet, a key element of the climate system. Moreover, Antarctica has been identified as a potential satellite calibration target.

 L-band radiometers measurements collected over the Dome C area demonstrated that the brightness temperature is relatively stable at vertical polarisation (standard deviation lower than 1 K at annual scale), while is more variable at horizontal polarisation. From November 2014 to March 2015, a significant event has been observed from both the DOMEX ground campaign and the SMOS satellite. The brightness temperature at horizontal polarisation shows a continuous increase followed by a sharp decrease (about 5 K at 42 degrees of incidence angle) within a few days. No such variations have been observed at vertical polarisation, which suggests that this event could be due to near surface changes. Here we use atmospheric and snow in situ measurements to investigate these variations. When the decrease in brightness temperature of horizontal polarisation occurred, a corresponding increase in snow surface density from 150 kg m^-3 to 300 kg m^-3 was measured. This could be related to a storm passing on Dome C, which may have compacted light snow or removed it from the surface uncovering old denser snow layers.

Electromagnetic models (DMRT-ML, WALOMIS) have been used to reproduce this decrease in L-band brightness temperature at horizontal polarisation. For that, in situ measurements have provided snow temperature and density profiles, as well as snow surface density. Simulations of time-series from January to June 2015 are performed changing only the density of the top snow layer. First results show a good agreement between modelled and measured brightness temperatures and confirm the major role of the surface density. However, when snow surface density is low (about less than 200 kg m^-3), the model skills are still unsatisfying, suggesting that more investigations are needed to understand and improve the simulations.

DTU Space have collected surface elevation observations of the Arctic sea ice and land ice since 1998 using laser scanning and radar altimetry from a small fixed‐wing Twin‐Otter aircraft. The observations provide unique datasets for studying ongoing changes, and support the analysis of satellite measurements of ice sheet changes. The majority of the campaigns have been sponsored by the European Space Agency, ESA, as part of the CryoSat Validation Experiments – CryoVEx. These have been internationally coordinated efforts to collect coincident space‐borne, airborne, and in‐situ data for pre‐ and post‐launch validation studies, with several aircraft and international in‐situ ground teams participating, both in Greenland, Arctic Canada, and Svalbard.

The methods and campaigns are outlined together with examples of results.The campaigns focused on five main validation sites: Devon ice cap (Canada), Austfonna ice cap (Svalbard), the EGIG line crossing the Greenland Ice Sheet, as well as the sea ice north of Alert and sea ice around Svalbard in the Fram Strait. Selected tracks were planned to match CryoSat‐2 passes and a few of them were flown in formation flight with the Alfred Wegener Institute (AWI) Polar‐5 carrying an EM‐bird. The poster will outline the methods and campaigns, as well as show examples of the results.

The ESA CryoSat mission is the first space mission to carry a radar altimeter that can operate in Synthetic Aperture Radar (SAR) mode. It thus provides the first opportunity to test and evaluate, using real data, the significant potential benefits of SAR altimetry for ocean applications.

The objective of the CryoSat Plus for Oceans (CP4O) project was to develop and evaluate new ocean products from CryoSat data and so maximize the scientific return of CryoSat over oceans.  The main focus of CP4O has been on the additional measurement capabilities that are offered by the SAR mode of the SIRAL altimeter, with further work in developing improved geophysical corrections.

CP4O has developed SAR based ocean products for application in four themes: Open Oceans, Coastal Oceans, Polar Oceans and Sea Floor Topography.  The team has developed a number of new processing schemes and compared and evaluated the resultant data products.  This work has clearly demonstrated the improved ocean measuring capability offered by SAR mode altimetry and has also added significantly to our understanding of the issues around the processing and interpretation of SAR altimeter echoes.

This paper presents an overview of the major results and outlines a proposed roadmap for the further development and exploitation of these results in operational and scientific applications, with particular focus on their relevance for Sentinel-3.

 The “CryoSat Plus for Oceans” (CP4O) project has been supported by ESA (Support To Science Element) and CNES.

In this presentation an overview of the CryoSat products currently available, their latency and the way to access them is given. Data visualisation tools for CryoSat data are described, togther with directions where to find all relevant documention and information to fully exploit the CryoSat data. In addition the evolutions engineering activities are briefly discussed.

With the launch of ESA’s Earth Explorer CryoSat, it has been possible to prove the achievable performance of Synthetic Aperture Radar (SAR) altimetry. Cryosat carries as main payload a Ku-band pulsewidth-limited radar altimeter, called Synthetic Interferometric Radar ALtimeter (SIRAL): the first radar altimeter in the low Earth orbit able to operate in SAR mode.

The main advantages of a SAR altimeter is that, by exploiting the coherence of the emitted pulses, a narrower antenna is synthetized and an improved along-track resolution is obtained with respect to conventional pulsewidth-limited altimeters. In turn, for each scattering area on the Earth surface, more observations are gathered so that speckle reduction is obtained by multilooking. Recalling that the echo received by a radar altimeter is generated by the coherent reflections from random scatterers on ground, many individual backscattering constituents randomly combine resulting in speckle. The speckle is usually modeled as a multiplicative noise and it can be reduced by averaging together (i.e. multilooking) statistically independent samples of the received signal.

 

As a figure of merit to evaluate the residual speckle on the multilooked power waveform, the Equivalent Number of Looks (ENL) is usually adopted [1]. The ENL is defined as the estimate of the effective number of statistically independent looks for the backscattered power and it is smaller than the actual number of single look echo waveforms that have been averaged together [1].

In this work we have derived theoretically the ENL for a multilooked power waveform as function of the echo delay. This way, the speckle that affects the waveform can be theoretically computed for CryoSat and similarly for any other SAR altimeter.

It is worth recalling here the ENL is written as ENL(k)=(E[W(k)]^2)/(Var[W(k)]), where k is the range bin, W(k) denotes the random variable for the k-th bin of multilooked waveform, while E[] and Var[] denote the expectation operator and the variance operator, respectively. Being the multilooked waveform given by the average of N single look echoes and modeling the power of the single look sample for a given range bin as  independent and identically distributed exponential random variables due to the nature of speckle, the theoretical ENL(k) has been numerically computed exploiting a closed form expression for the single look waveforms (a simplified and approximated expression derived SAMOSA model [2] has been used).

The derived theoretical ENL has been validated by comparison with the experimental ENL measured on CryoSat Level1b SAR waveforms acquired over ocean. The theoretical ENL has been proved to be in good agreement with the experimental ENL, being able to predict the decay of ENL in the trailing edge and an oscillation in correspondence of the peak of the waveform. Some differences between the theoretical ENL and the experimental ENL occurs especially at the beginning of the leading edge of the waveform and are reasonably due to the simplified single look waveform model that has been used in this preliminary computation of the theoretical ENL.

 

[1] Wingham, D.J.; Phalippou, L.; Mavrocordatos, C.; Wallis, D., "The mean echo and echo cross product from a beamforming interferometric altimeter and their application to elevation measurement," in Geoscience and Remote Sensing, IEEE Transactions on , vol.42, no.10, pp.2305-2323, Oct. 2004

[2] Ray, C.; Martin-Puig, C.; Clarizia, M.P.; Ruffini, G.; Dinardo, S.; Gommenginger, C.; Benveniste, J., "SAR Altimeter Backscattered Waveform Model," in Geoscience and Remote Sensing, IEEE Transactions on , vol.53, no.2, pp.911-919, Feb. 2015

In the last years the possibility to use L-band space-borne radiometers for monitoring the Cryosphere has been investigated using data available from ESA’s SMOS and NASA’s Aquarius and SMAP missions. The interest in L-band for snow and ice applications relies on the very low absorption of ice at L-band, about one order of magnitude lower than for the higher frequencies, and the low scattering by particles (grain size, bubbles in ice) that are very small compared to the wavelength. As a consequence, in dry snow and ice the extinction is low and the penetration depth is very high (hundreds of meters) which open new opportunities to probe the soil or water under the ice, or the internal layers of the ice-sheet. The CryoSMOS project, which was funded by ESA as Support To Science Elements (STSE), aims at investigating this topic by testing the capabilities of SMOS in the monitoring of Antarctica, excluding sea-ice. As a starting point, SMOS data were in-depth analyzed in order to investigate spatial and temporal Tb signatures, which can be related to geophysical parameters of the ice sheet and ice shelves. In particular, it has been observed that Tb can show temporal dynamic trends in the region of the ice shelves and near to the coast where the snow could be wet, while it is more stable in time, but presents significant spatial features in the inner parts of the continent. Moreover, small but significant Tb variations are observed also in the internal part if observed at H polarization. Four case studies, which are in-depth analyzed within the project, have been considered: the possibility to estimate the temperature profile of the ice sheet; the capability of investigating bedrock topography and/or geothermal heat flux; the characterization of the ice shelves; the monitoring of surface properties. For each study case the SMOS data have been first interpreted by using different electromagnetic models which use as inputs data collected on the ground, when available, or from glaciological models. Simulated and measured Tb is in general in good agreement confirming that most of the observed Tb spatial and temporal signatures can be theoretically explained. Model analysis also shows that a better knowledge of dielectric permittivity of ice (especially of its imaginary part which is indeed very small) is required to further improve the results. Starting from this, inversion algorithms have been developed in order to derive geophysical parameters from SMOS data. A preliminary validation has been performed in few tests sites where detailed information is available. The results demonstrate the potential of the method, which will be further extended to the whole Antarctica. 

The Churchill Marine Observatory (CMO) has recently been funded ($32M) as a globally unique, highly innovative, multidisciplinary research facility located in Churchill, Manitoba, adjacent to Canada’s only Arctic deep-water port.  The CMO will directly address technological, scientific, and economic issues pertaining to Arctic marine transportation and oil and gas exploration and development throughout the Arctic.   CMO will include an Oil in Sea Ice Mesocosm (OSIM), an Environmental Observing (EO) system, and a logistics base.  OSIM will consist of two saltwater sub-pools designed to simultaneously accommodate contaminated and control experiments on various scenarios of oil spills in sea ice.  The EO system will be located in the Churchill estuary and along the main shipping channel across Hudson Bay and Strait.  The EO system will provide a state-of-the-art monitoring system and will be used to scale process studies conducted in OSIM to Hudson Bay and the larger Arctic marine environment.  The logistics base will underpin all CMO research.  CMO will position Canada as a global leader of research into the detection, impacts, and mitigation of oil spills in sea ice.  Knowledge gained through CMO will strengthen Canada’s technological capacity to protect the Arctic environment.  Partnerships with indigenous organizations will ensure knowledge exchange; the private sector will provide market-driven uptake of technology; and various levels of government will transfer knowledge into policy and regulation.  CMO will also act as a full service in situ validation site for existing and planned Earth Observation satellites.  ArcticNet investigators have an opportunity to engage in the science of the CMO using either/both and OSIM or EO components.  The facility will begin operations in 2017.

Climate change is well underway in the polar regions of our Planet. Space programs (such as ESA) are providing key earth observation technologies required to understand the changes currently underway.  Surface validation of these space assets is a key requirement for the scientific and operational use of remote sensing data. In this poster we present the characteristics of an ongoing partnership between Aarhus University (Denmark), Greenland Climate Research Centre (Greenland) and the University of Manitoba (Canada) who have joined forces to study the impacts of climate change on the Arctic marine system. The partnership merges over 350 scientists from 20 different research centres and coordinates field campaigns in Greenland and Canada including fieldwork at two unique artificial sea ice microcosms: i) the Sea Ice Environmental Research Facility (SERF) located on the University of Manitoba campus; and ii) the new Oil in Sea Ice Mesocosm (OSIM) and Environmental Observing (EO) system located in Churchill, Manitoba.  The EO system will provide a state-of-the-art monitoring system and will be used to scale process studies conducted in OSIM to Hudson Bay and the larger Arctic marine environment.  Scheduled to open in 2017, OSIM and EO system will act as an in situ validation site for existing and planned Earth Observation satellites.  We describe the nature of this partnership, indicate how Earth Observations of the cryosphere play a role in the research agenda and describe the outreach programs that inform the public and policy makers about the results emerging from this international partnership.

The Greenland ice sheet is rapidly shrinking with the fastest retreat and thinning occurring at the ice sheet margin and near the outlet glaciers. The changes of the ice mass cause an elastic response of the bedrock. Ice mass loss during the summer months is associated with uplift, whereas ice mass increase during the winter months is associated with subsidence. The German TerraSAR-X satellite has systematically observed selected sites along the Greenland ice sheet margin since summer 2012. Here we present ground deformation observations obtained using an InSAR time-series approach based on small baseline interferograms. The deformation data reveal the seasonal variations and net uplift.  Relative variations in the seasonal amplitude for different sites along the ice sheet margin points to spatial variations in ice loss. The combination of ground deformation observations and independent observations of  ice volume changes from airborne and spaceborne altimeters places constraints on the firn density of the lost ice volume.

The rapid decrease of Arctic summer sea ice make the regular commercial transport along the Northeast Passage possible nowadays. Near real-time satellite derived sea ice concentration AMSR2 and SSMIS are commonly used in the manned ship-route planning and numerical model forecasting, however the accuracy and applicability of those data haven’t been systemically evaluated. In this paper, ship-based sea ice observations from Xuelong and Polarstern along Northeast Passage in summer of 2012 are used to assess the concentration and extent from AMSR2 and SSMIS. The results show that AMSR2 agree well with ship-based SIC (R=0.84), but a little overestimation in high concentration (mean bias=0.03). SSMIS have a poor relationship 0.56 with ship-based SIC and obviously underestimate the high concentration larger than 30% (mean bias= -0.14). Further analysis imply what widespread exist of melting pools lead to the large bias of SSMIS in high concentration, which maybe a systemic algorithm error. AMSR2 perform better than SSMIS in detecting sea ice concentration and extent, so AMSR2 should be firstly considered when using them in the operational Arctic sea ice services.

The mountain glacier changes are believed to currently provide significant responses to global climate change and strongly influence human welfare in this arid or semi-arid region, where water supplies are predominantly from glacier melt. Meanwhile, glacier mass balance has direct contribution sea level rise, declining water resources, runoff and disaster of glacier lake outburst. So the accurate estimation of mass balance at high spatial and temporal resolution is very important. Although traditional ground-based techniques exist for measuring glacier mass balance directly and inter-annually, they tend to be labor-intensive, expensive and provide very limited spatial coverage. Synthetic Aperture Radar (SAR) observations offer direct means of monitoring changes in surface elevation over the mountain glacier and can achieve centimeter-level accuracy.

In this study, we developed a method to utilize observations of the glacier surface deformation to derive the thickness changes, and then calculated the mass balance using SAR observations. Ths study area local in Koxkar glacier (41.42︒N-41.53︒N and 79.59︒E-80.10︒E) in Tianshan Mountains, China, which is a typical Tuomuer-type glacier originating from Mt. Koxkar (6,342 m a.s.l.), and flows southeast to the terminus of 3,020 m a.s.l. The glacier extends 25.1 km in length and covers an area of 83.56 km². The equilibrium line occurs at 4,300 m a.s.l. in the icefall from whose foot a 15.5-km-long, debris-mantled glacier tongue appears. The supraglacial debris covers an area of about 19.5 km², which accounts for 83% of the total ablation area, with thicknesses ranging from less than 0.01 m on the upper reach of the ablation area and on ice cliff faces to more than 3.0 m near the glacier snout.

We reveal that the mass balance of the ablation area of the Koxkar glacier (from 3800 m to 3100 m) in melting season and accumulating season. Due to the gap of the SAR observations, we only derive the mass balance of the ablation area of the Koxkar glacier (from 3800 m to 3100 m). The results show that the mass balances are from -1471.2 mm w.e. to 103 mm w.e. corresponded to 3800 m to 3300 m elevation in 1999. The paper demonstrates the feasibility of the presented method to obtain and analyze the mass balance of the mountain glacier.

The ice loss of glacier has been the dominant mass contributor to sea level changes during twentieth century [1]. It is important to monitor thickness changes of mountain glaciers for their contributions to calculate ice volume loss and mass balance [2]. Although traditional ground-based techniques exist for measuring glacier thickness changes directly and inter-annually, they tend to be labor-intensive, expensive and provide very limited spatial coverage. So the remote sensing data are an attractive source of information to complement in situ measurements [3]. Differencing multi temporal digital elevation models (DEMs) generated from space-borne or air-borne observations becomes one of the effective methods to monitor the spatial patterns of glacier thickness and volume changes [4]. These methods include generation of DEMs from radar interferometry, for example, the Shuttle Radar Topography Mission (SRTM) DEM acquired in Feb, 2000 [5], or from spaceborne or airborne photogrammtery [6], or from laser altimetry [7]. However, the accuracy of the results derived from these methods depends mainly on the DEMs’ accuracy. Because of the low precision of these DEMs, the estimation of glacier thickness change is difficult to satisfy the evaluation of ice loss with precision. Over the past decades, the spaceborne synthetic radar aperture interferometry (InSAR) technique has proved to be an effective remote sensing tool to map ground deformations [8]. Using the DInSAR method, pairs of SAR images can be processed to obtain the high spatial resolution maps of surface deformation with large spatial coverage and the precision achieved on the order of centimeters [9]. Because the current SAR satellite systems have relatively short revisit period, DInSAR has the capacity to resolve time-dependent deformation. During the recent years, studies based on DInSAR have so far mostly concentrated on obtaining the glacier movement, and little has been done on deriving the glacier thickness changes using the one component of surface deformation along the line of sight (LOS).

The goal of this study is therefore to use the differential phase derived by DInSAR algorithm to obtain the mountain glacier thickness change and analysis the spatial variations of the Dongkemadi glacier thickness changes. Using this method the glacier thickness changes can be monitored in cm-level accuracy. The results shows the glacier thickness changes of KG are divided by the central axis into the left thickening side and the right thinning side from 3500 m to 3700 m. The accumulation zone of Dongkemadi glacier shows continue to thicken. Net loss occurred mainly at the north-eastern glacier.

The Himalayan glaciers, the largest body of ice outside of the polar icecaps are a source of water

for major rivers of Asia, such as the Indus, Ganga and Brahmaputra. There are growing concerns

about the impact melting of glaciers in the Himalayas, which will affect densely populated

communities in downstream river basins. There is huge uncertainty about how snow and glacial

melting in the Himalayan region will continue to affect climate change and its impact on

ecosystems and well-being of mankind. Due to scarcity of ground measuring stations, remote

sensing is playing a major role in the monitoring and assessment of the Himalayan cryosphere.

The present study demonstrates the utilization of Oceansat-II OSCAT scatterometer data for the

assessment of ice surface melting over the Himalayan region. The enhanced resolution (4.45 km)

Ku-band Sigma-0 data of OSCAT (2011) and QuikSCAT (2001) scatterometers have been

utilized in the present study. K-means clustering algorithm was applied to identify the icy area

using temporal variations of Sigma-0 (fig. 1). It is observed that the presence of liquid water due

to the melting of snow/ice significantly reduces the backscattering (fig.2). The temporal change

observed in the Sigma-0 has been compared with the changes observed in the MODIS Land

Surface Temperature (LST). It is observed that the increase in temperature due to summer

warming is associated with the decrease in backscattering from the snow cover ice surface. Onset

dates for the melting and re-freezing were also identified by investigating the decrease/increase

of Sigma-0 by the threshold value derived in the project.

The first ever demonstration of OSCAT scatterometer over Himalayan region indicates the

potential of enhanced resolution OSCAT data for the regular monitoring of surface melting and

changes detection studies.

The European Space Agency’s (ESA) climate change initiative (CCI) sea ice project has established a round robin data package (RRDP) for sea ice concentration (SIC) algorithm inter-comparison and evaluation. The dataset is openly available. It contains an extensive collection of collocated microwave radiometer data and other geophysical parameters with relevance for computing and understanding variability of the microwave radiometer SIC over open water and 100% sea ice. It covers the Northern and Southern hemispheres, thin ice, first-year ice, multiyear ice and all seasons including summer melt.

The open water data (SIC=0%) are collected in selected regions outside but relatively close to the ice edge in the Atlantic and the Pacific on both hemispheres. High resolution sea ice drift fields derived from consecutive and overlapping SAR scenes have been used to identify regions of sea ice convergence. During winter these regions are assumed to have ~100% sea ice cover (either due to the convergence or due to refreezing of eventual remaining openings). Similarly to the open water data all ice data have been resampled to coarse resolution. A thin ice dataset has been established using ENVISAT ASAR data to identify areas of near 100% ice concentration, and SMOS to quantify the ice thickness, and a Summer dataset with melt-pond fractions has been derived from MODIS data. 

The database consists of TB data measured by:

AMSR-E, SMMR and SSMI microwave radiometers,

Simulated data; using a radiative transfer model for snow and ice; using modified Wentz radiative transfer models for the atmosphere and the open ocean. The simulated data contain in addition to the brightness temperatures at SMMR, SSM/I and AMSR channels also geophysical parameters snow depth, physical temperature and atmospheric parameters

SMOS brightness temperatures

NWP data from ERA Interim

ASCAT backscatter at 40 degree incidence angle

This study is part of a Sea Ice Concentration algorithm evaluation of the European Space Agency (ESA) Climate Change Initiative (CCI) project. The project is developing capacity to construct a sea ice climate record from satellite data including sea ice thickness from radar altimetry (ERS 1-2, ENVISAT and eventually Cryosat 2 and Sentinel - 3) and sea ice concentration.

Snow and Ice surface temperatures (IST) are important boundary conditions for atmosphere and sea ice models in the high latitudes. Several thermal infrared satellite products are now available, e.g. from the ESA DUE project Globtemperature, the EUMETSAT Ocean and Sea Ice Satellite Application Facility (OSI-SAF) and with the Copernicus Marine Environmental Services. Due to the geophysical conditions of the snow and ice, the validation of satellite ice surface temperature observations and their SI traceability is very challenging and there is a need for establishing fiducial reference measurements and for protocols on best practice procedures when validating the satellite IST products This work is undertaken in the ESA project: Fiducial Reference Measurements for validation of Surface Temperature from Satellites (FRM4STS).

In this presentation, we will show results from the FRM4STS project with examples from field campaigns on the Sea ice off Qaanaaq, a High-Arctic settlement in Greenland. The instrumentation includes an Automatic Weather Station (AWS), ice mass balance buoys and narrow and broadband thermal infrared radiometers and ice mass balance buoys. These observations have been used to validate the satellite products and assess the uncertainties related to different components in the satellite vs. in situ comparisons, such as: the temporal and spatial differences and the vertical gradients within the snow and sea ice. Initial ideas will be presented on the protocols and guidelines for inter-comparisons of the thermal infrared radiometers. In addition, the modelled satellite uncertainties derived within the Horizon 2020 project EU Surface Temperatures for All Corners of Earth (EUSTACE), will be validated. To aid in the assessment of the uncertainty budget, results from a simple sea ice thermodynamic and infrared emission model will corroborate the results obtained from the field observations.

 

Nadir-looking dual-frequency microwave radiometers have been initially embarked on satellite mission to perform the wet tropospheric correction for altimeter range over ocean. To evaluate ice sheet responses to global warming, it is necessary to investigate geophysical parameters whose allow the monitoring of the ice sheet evolution. In order to retrieve the snowpack properties (snow grain size, density, accumulation rate,...), we explore the dual-frequency microwave radiometer variable ( Brightness Temperature (Tb)) on the Antarctic ice sheet on-board ERS, ENVISat and AltiKa In combination with the altimeter waveform parameters (Backscatter coefficient, Trailing edge Slope(TeS) and Leading edge Width(LeW)). We compare the radiometer brightness temperature to calculations with the DMRT- ML radiative transfer model which simulates brightness temperature in vertical and horizontal polarizations. This combination allow a more accurate retrieval of snowpack properties. We also show temporal change within +/- 2 K/yr with clear spatial distribution over the Antarctic continent. This work will provide a higher accuracy of the estimation of the topography by better describing the volume echo and in turn a better constraint for ice sheet glaciology modeling and contribute to monitoring the ice sheet surface mass balance.

Remotely sensed surface temperatures of ocean and sea ice are of great interest as this is one of the factors that govern the exchange of heat at the surface.  Whereas remote sensing of sea surface temperatures is a well-established technique, ice surface temperature (IST) retrievals  are at a more novel stage.

This presentation will present the results from a user case study within the ESA DUE project Globtemperature, where satellite ice surface temperature products have been used from Modis, Metop and Iasi from August 2012 to Septemper 2013. The different satellite products have been validated against in situ observations and a multisensor level 4 analysis has been constructed, to provide a gap free field of sea and sea ice surface temperatures for the Arctic. The level 4 product, which is very similar to the product delivered within the Copernicus Marine Environment Services, has been validated against in situ observations from thermal infrared radiometers, ice mass balance buoy and International Arctic Buoy Program (IABP) observations.

In addition to the improved understanding of the temperatures at the surface, the satellite based product has been assimilated into a coupled ocean and sea ice model. The Danish Meteorological Institute (DMI) runs an operational coupled ocean and sea ice model (HYCOM+CICE+ESMF) based on a model system mainly developed by the Naval Research Lab.  This runs in a regional configuration that covers the Arctic and the Atlantic oceans to 20 degrees south of equator and it is forced by ERA-INTERIM in hindcast mode and the operational ECMWF in forecast mode. The results from the assimilation experiment and the impact of ingesting the sea and sea ice surface temperature products into the model will also be presented.

Since April 2015, the CryoSat ice products are generated with the new processing baseline, which represents a major processor upgrade and the baseline for the second CryoSat Reprocessing Campaign; Baseline-C. Several improvements have been implemented, some of which include the SAR retracker optimized for Freeboard retrieval, new tuneable land-ice retracker, and a coarse slant correction which is applied directly on the stack data in conjunction with the window delay alignment. The resulting waveforms show more power and the trailing edge is modified, leading to improved L2 geophysical parameters. This poster provides a high level overview of the various changes and evolutions implemented as part of the Baseline-C processor upgrade in order to improve CryoSat L1 and L2 data characteristics and exploitation over Polar Regions. In addition an overview of the main outcomes from specific IDEAS+ Quality Control performed on the operation Baseline-C products is also presented.

Launched in 2010, the polar-orbiting CryoSat satellite was designed to measure the changes in the thickness of polar sea ice and the elevation of the ice sheets and mountain glaciers. To reach this goal, the CryoSat data products have to meet the highest performance standard and are subjected to a cycle of constant improvement of the associated Instrument Processing Facilities. Following the switch to Baseline-C there are already several planned evolutions for the next Baseline, based on recommendations and feedback from the Scientific Community, Expert Support Laboratory (ESL), Quality Control Centres and Validation campaigns. Some of the proposed evolutions, to be discussed with the scientific community, will include improvements in the processor for Antarctic and Arctic sea ice, optimisation of freeboard quality in SARin mode, better exploiting the potential of SARin data over flat-to-slope transitory land-ice areas, tuning of the land-ice retracker, the switch to netCDF format and the resolution of residual and new anomalies arising from Baseline-C. This poster presents some known processor anomalies and also some of the potential evolutions that are planned and foreseen for future Baselines.

Up to date and frequent glacier outlines are needed for many applications within glaciology. Earth observational data provide a good suitable means for extracting outlines over large areas, and are widely used for semi-automated glacier mapping worldwide. Clean glacier ice can be mapped robustly and accurately using band ratios between the Red or Near Infrared (NIR) spectral channels and the Shortwave Infrared (SWIR) channels and a user-defined threshold. Several methods have been put forward to semi-automatically debris-covered ice, often relying on topographic parameters, the thermal band or SAR coherence data. As a long-term goal the glaciology community should work from semi-automatic glacier mapping methods to fully-automatic glacier mapping. Here, we show work in progress concerning the creation of a New Norwegian glacier inventory using automatic thresholds within an object-based environment. Although there are problems associated with the automatic selection and application of thresholds for multispectral imagery, our results are promising, and could help move towards a workflow that can map clean glacier ice without the need for user-given thresholds. In order to include debris-covered ice and exclude some turbid lakes, some semi-automatic post-processing was necessary that makes use of the ability to assign classes due to spatial, hierarchical and contextual information when working at an object level. Our workflow seems to be transferable between different regions of the Norwegian mainland, but future work should focus on automatically classifying debris-covered ice within an object-based environment, and should trial the methodology in other glacierised regions.

During the Arctic freeze-up period, L-band radiometer data from ESA's Soil Moisture and Ocean Salinity (SMOS) mission are well suited to derive sea ice thickness up to half a meter. The sea ice remote sensing group at the University of Hamburg has developed an iterative retrieval algorithm for the ice thickness, which uses the brightness temperature intensity and is based on a thermodynamic sea ice model and a three-layer radiative transfer model. The used brightness temperature measurements have a resolution of about 35 km and a daily coverage of the polar regions.

The recently launched Soil Moisture Active Passive (SMAP) satellite by NASA also carries an L-band radiometer with orbital parameters and footprint areas that are comparable to those from SMOS. This offers the opportunity for a comparison of measurements from both sensors. Thus, in this contibution we present a detailed comparison of brightness temperature measurements from SMOS and SMAP over the polar oceans. We also evaluate the potential of combining the measurements from both sensors in order to improve the current sea ice thickness product.

Since SMAP has a fixed scanning geometry with an incidence angle of 40°, while SMOS measures over the whole incidence angle range of 0 to 65°, is necessary to interpolate the SMOS measurements to 40° before the comparison. For this purpose we use a two-step regression, which first estimates the brightness temperature intensity at nadir and then uses this value to fit seperate functions to the horizontally and vertically polarised brightness temperatures.

The brightness temperature components are compared over different ice types and over the open ocean, as well as for different seasons. Preliminary results suggest a moderate level of agreement over sea ice with very high correlations but biases of up to 2 K in the cold seasons and even slighly
larger values during the melting season. Differences are much larger over water, however, with biases exceeding 10 K for the horizontally polarised brightness temperatures. We further discuss the implications of these differences for a joint brightness temperature product from SMOS and SMAP over
the polar oceans and for the resulting sea ice thickness retrieval.

In addition, we discuss and compare the different treatments of radio frequency interference (RFI) for both sensors. For SMOS, RFI contaminated measurements are commonly identified and discarded, which can cause substantial data loss, especially close to the coasts of the Arctic Ocean. To resolve this RFI problem, SMAP contains a dedicated subsystem for the detection and mitigation of RFI.

The main payload of CryoSat is a Ku band pulsewidth limited radar altimeter, called SIRAL (Synthetic interferometric radar altimeter), that transmits pulses at a high pulse repetition frequency thus making the received echoes phase coherent and suitable for SAR processing.

This allows to reach an along track resolution that is significantly improved with respect to traditional pulse-width limited altimeters. Due to the fact that SIRAL is a phase coherent pulse-width limited radar altimeter, a proper calibration approach has been developed. In fact, not only corrections for transfer function, gain and instrument path delay have to be computed (as in previous altimeters), but also corrections for phase (SAR/SARIn) and phase difference between the two receiving chains (SARIN only). To summarize, SIRAL performs regularly four types of internal calibrations:

CAL1 in order to calibrate the internal path delay and long-term power drift.

CAL2 in order to compensate for the instrument IF transfer function.

CAL4 to calibrate the interferometer.

AutoCal, a specific sequence used to calibrate the gain and phase difference for each AGC setting.

After about 5 years of operational activity of the CryoSat satellite, the performance of the SIRAL instrument are revealed to be in line or better than the expected one.

In fact the calibration products, that have been designed to model a wide range of imperfections of the instrument, can be analyzed to highlight whether and how the instrument is changing over the time also as function of its thermal status. It is worth underlining here that each variation of the instrument measured by the calibration data is compensated in the Level1 processing.

 

Inspecting the temporal evolution of the calibration data, SIRAL has been verified to be stable during its life. In this poster, as first an overview of the SIRAL instrument, its operating modes and the calibration strategy will be given. Then, the performance of the SIRAL will be presented together with the outcomes of the stability analysis on the calibration data, in order to verify that the instrument has reached the requirements and that it is maintaining the performance over its life. Moreover, in-depth analysis of the calibration corrections  revealed how the instrument depends on its temperature and on the not-sun-synchronous orbit of the satellite.

The high-latitude regions of the earth are characterized by a vast number of water bodies, mostly lakes. Water in general and lakes in particular have the ability to store heat. The monitoring of the ice-cover variability of lakes in high latitudes is therefore a good indicator for changes related to global warming and its effects on the Polar Regions. Furthermore the knowledge of ice growth and decay is important for environmental protection, resource development and safety of operations. E.g. frozen lakes and rivers are often used as transportation routes during the winter period and flow releases from hydroelectric generation facilities can better be managed to reduce the risk of ice jam related flooding.

Remote Sensing has the capacity to provide accurate high resolution information of the sea and land surface in an automated and standardized way. In particular satellites equipped with Synthetic Aperture Radars (SAR) enable regular mapping and monitoring. Their all-weather and day and night observation capability are important advantages in the Arctic due to high cloud coverage rates and low illumination during the winter period.

In 2007 the German Earth Observation satellite TerraSAR-X was launched that provides high resolution SAR data. Since 2010 it is flying together with its twin satellite TanDEM-X in close formation enabling single pass interferometry. The primary mission goal is the generation of a global digital elevation model in outstanding quality and resolution that allows classifying the shoreline and land’s topography at an unprecedented level of detail.

On April 3rd, 2014 ESA launched the first of two Sentinel-1 satellites. The main purpose of this mission is to acquire and deliver SAR data for global monitoring and observation. It ensures continuity of SAR systems operating in C-band frequency.

Objective of the presented work is to assess the applicability of multi-frequency SAR systems, in particular Sentinel-1 and TerraSAR-X  data for the monitoring of lake and river ice growth and decay. S1 data acquired in Interferometric Wide Swath mode with VV/VH polarization were processed and analyzed for the whole winter period 2014/2015. Additionally dual polarized SAR data (HH/VV) of TerraSAR-X are available. Sentinel-1 has the capability for large area mapping while TerraSAR-X provides high-resolution data for detailed regional investigations.

All SAR data are automacially processed to Kennaugh elements, radiometrically calibrated and ortho-rectified using the TanDEM-X DEM. The classification differentiates between open water and up to eight ice classes over the whole freezing period.

The test site is located in Northern Finland and ranges from the Swedish border in the West to Russian border in the East. It covers three rivers, numerous small lakes, but also wide-stretched lakes like Porttipatha Lake and Lokan Tekorjavi. The latter is one of the biggest reservoirs in Europe with an extension of approximately 400 square kilometers. Four meteorological stations provide weather data that are used for verification toghter with optical imagery.

Parameters and geophysical corrections provided in L1, L2I and L2 CryoSat data products are currently checked using the Quality Control for CryoSat (QCC) tool integrated at the Payload Data Ground System (PDGS). This tool provides configurable product quality control functionality for CryoSat data products according to a predefined Test Definition File. A detailed analysis has recently been performed in order to tune these thresholds, as the current thresholds were defined at the start of the CryoSat mission. The study was extended to the recently released CryoSat Ocean Products: IOP and GOP. The updated test thresholds have now been implemented in a new version of the QCC tool which is operationally screening all CryoSat Baseline-C Ice products and the CryoSat Ocean Products. This poster presents the key points from this analysis, summarising the main statistical results and providing a high level overview of the quality trends of CryoSat products, including plots of the temporal change in variables seen over time in operational Baseline-C data.

One of the major evolutions planned for both CryoSat Ice products (FBR, L1, L2, L2i) and Ocean products (L1, L2) is the switch to NetCDF which is described as a far more user-friendly format in comparison to the current Earth Explorer format (EE). This poster provides information on the new format, rationale from other missions as well as some examples and details of the concerned CryoSat data products.

The EE format was designed at the time of the Cryosat 1 mission by using ENVISAT product formats as template with the purpose to maximise the reuse of decoding/analysis tools developed for that mission. The Earth Explorer is however no more suited for today’s evolution and maintenance needs.  ESA therefore decided to face a migration from the EE to the netCDF format, which is the standard “de-facto” for the product formats of the most advanced and modern Earth Observation missions. The new format for Cryosat products is called CONFORM (CryOsat Netcdf FORMat) and this migration is an ongoing activity: the first Cryosat products in netCDF would be distributed to the users around Q4 of 2016.

The design of the CONFORM is performed by considering the Climate and Forecast conventions as well as the guidelines used in other ESA missions. One of the objective will be to ensure good homogeneity and harmonisation between CryoSat and previous (reprocessed ERS1/2, Envisat-Phase F), new (Sentinel-3) and future (Sentinel 6) ESA altimetric missions. At the same time the new design could also contribute to the definition of new conventions and standard for some particular fields specific to this first ESA ice-oriented mission.

Iceberg (IB) detection is significant for navigation in the areas where icebergs may exist. Here we present an algorithm using SENTINEL-1a (S-1a) Extra Wide Swath (EW) mode dual-polarized L1 GRD medium resolution (GRDM) data with HH/HV polarization combination. The applied algorithm is a non-parametric so-called Constant False Alarm Rate (CFAR) algorithm and this version also utilizes SAR segmentation. Because segmentation has been utilized, the segments with brighter backscattering with respect to their background can be extracted and the background distribution can be computed outside of these bright segments, instead of the usual CFAR approach of using a radius/distance parameter around the center point for the background. The algorithm was applied on both the SAR channels and the CFAR results of the channels were combined afterward to yield the best possible detection. For comparison, also traditional CFAR with Gaussian and Gamma distribution assumption were run and the results were compared to the non-parametric CFAR algorithm results. Our algorithm also computes multiple features related to each detected object. The features describe the segment (object) shape and contrast with respect to its background.
To evaluate the algorithm performance, icebergs were detected and manually located by FMI ice analysts in a set of SAR imagery and supporting LANDSAT imagery, and the recognition rates for the CFAR methods were computed. The data used in this study was in the Greenland waters (south, west and east of Greenland).

In preliminary tests It was found that all three tested IB detection methods provide similar results in comparison to each other. However, the parametric Gauss/Gamma methods shows higher target detection rate, including a high number of false detected targets. The non-parametric method shows a lower rate of detected targets than the parametric ones, when the two polarization channels are combined, or in the HH channel alone. However, the HV channel seems to detect a similar number of targets as with the parametric methods, but the number of detected targets is much closer to the number of marked targets by the ice analysts, including a low number of false detected targets, which were later on re-counted by the ice analyst, as valid targets. While the parametric methods show a high rate of target detection, but a much higher false detections rate, which cannot be at this point validated against the available data sets. The nonparametric method can be adjusted by segmentation parameters and by the applied CFAR thresholds. Also more different parametrizations for segmentation and CFAR will be studied to improve the algorithm performance compared to the parametrization used in the tests performed this far. The algorithm will also be applied with different parametrization over open water and sea ice, to detect icebergs either floating in open water or within sea ice. To enable different classification in sea ice and open water, these two classes are first separated based on the HV channel SAR backscattering and texture features for HH and HV channels. Also AIS (Automatic Identification System) data to locate ships in the area will be included in the analysis, and difference of the ship and iceberg segments will be compared based on the segment shape and contrast characteristics.

Cryosat-2 offers the first ever possibility to perform coastal altimetric studies using SAR-Interferometry as well as SAR altimetry. With this technological leap forward Cryosat-2 is now able to observe sea level in very small water bodies and also to provide coastal sea level very close to the shore.

We perform an investigation into the retrieval of marine gravity in several of the fjords in eastern Greenland. Among the fjords are the Scoresbysund Fjord which is the largest and deepest fjord in the World. In the mariginal zone of Greenland the SAR-in is mainly used because of the huge topographic changes as Cryosat-2 is designed to map the margins of the ice-sheet.

For retrieval of marine gravity and also mean sea surfaces, the main important parameter is the spatial density of the sea level data. With only a few points in the fjords from Envisat, the new retracked Cryosat-2 SARin data offers a huge step forward in terms of data quality and data availability. We employ data from the first 5 years (summers) of Cryosat-2 to quantify the improvemnt that be achieved.

By comparing the data with and without the off-nadir correction we can furthermore study the improvement that can be expected from the SAR-in w.r.t. SAR data from gravity field retrieval in complex coastal regions.

Due to its high albedo and low heat conductivity (about one magnitude
smaller than that of sea ice [Massom et al., 2001]) snow on sea ice plays an
important role for polar radiative processes and heat transfer. Consequently
snow properties are an important input parameter to correctly quantify heat
(and also water) transfer in the ocean-sea ice-atmosphere interaction in cli-
mate studies and in data assimilation in global climate and numerical weather
prediction models. Furthermore, accurate snow depth products are necessary
to retrieve sea ice thickness from altimeter freeboard measurements.
Because in-situ measurements are only sparsely available and clouds and
missing daylight limit usage of optical and infrared satellite sensors, satellites
carrying passive microwave sensors offer the only possibility to continuously
retrieve snow depth on Antarctic scale.
The first algorithm for large scale retrieval of snow depth on sea ice was intro-
duced by Marcus and Cavalieri in 1998 for SSM/I (Special Sensor Microwave
Imager) and later adapted to AMSR-E (Advanced Microwave Scanning Ra-
diometer for EOS) data by conversion of AMSR-E brightness temperatures
to SSM/I equivalent brightness temperatures [Comiso et al., 2003, Brucker
and Markus, 2013].
However, in the Antarctic different processes (flooding, melting and refreez-
ing) cause highly variable snow properties. These together with ice ridges
make snow depth retrieval particularly difficult and as a consequence often
the accuracy of the retrieved snow depth is not known.
To improve this situation, in the frame of the ESA sea ice climate change
initiative (SICCI) project, the existing snow depth products from the stan-
dard NSIDC AMSR-E [Cavalieri et al., 2014] and Cryosphere Science Re-
search Portal SSM/I [Cryosphere Science Research Portal] retrieval are com-
pared with visual ship-based observations from an extended ASPeCt data
set [Worby and Allison, 1999, Worby et al., 2008] on a monthly basis and by
Antarctic sectors. By comparing the AMSR-E vertically polarised brightness
temperature gradient ratio of the 18.7 and 36.5 GHz channels with ASPeCt
snow depth observations, the algorithm is re-evaluated and adapted to be
directly applicable to AMSR-E and AMSR2 brightness temperatures. In ad-
dition, for the first time an uncertainty estimation based on Gaussian error
propagation is included. This will allow to distinguish uncertainties result-
ing from those of the input parameters and those arising from the limitations
of the algorithm itself. Finally the new AMSR-E snow depth product and
its uncertainties are compared to independent reference data sets such as
Operation IceBridge, buoy and in-situ measurements.

References

L. Brucker and T. Markus. Arctic-scale assessment of satellite passive
microwave-derived snow depth on sea ice using Operation IceBridge air-
borne data. Journal of Geophysical Research: Oceans, 118(6), 2892–2905,
2013. doi: 10.1002/jgrc.20228.

D. J. Cavalieri, T. Markus, and J. C. Comiso. AMSR-E/Aqua Daily L3 12.5
km Brightness Temperature, Sea Ice Concentration, and Snow Depth Polar
Grids. Version 3. June, 2002 to October, 2011. Boulder, Colorado USA:
NASA National Snow and Ice Data Center Distributed Active Archive
Center, 2014. http://dx.doi.org/10.5067/AMSR-E/AE SI12.003.

J. C. Comiso, D. J. Cavalieri, and T. Markus. Sea Ice Concentration,
Ice Temperature, and Snow Depth Using AMSR-E Data. IEEE Trans-
actions on Geoscience and Remote Sensing, 41(2), 243–252, 2003. doi:
10.1109/TGRS.2002.808317.

Cryosphere Science Research Portal. http://neptune.gsfc.nasa.gov/csb/index.php?section=52.
URL http://neptune.gsfc.nasa.gov/csb/index.php?section=52.

T. Markus and D. J. Cavalieri. Snow Depth Distribution over Sea Ice in
the Southern Ocean from Satellite Passive Microwave Data. Antarctic Sea
Ice: Physical Processes, Interactions and Variability, Antarctic Research
Series, 74, 19–39, 1998.

R. A. Massom, H. Eicken, C. Haas, M. O. Jeffries, M. R. Drinkwater,
M. Sturm, A. P. Worby, X. Wu, V. I. Lytle, S. Ushio, K. Morris, P. A.
Reid, S. G. Warren, and I. Allison. Snow on Antarctic Sea Ice. Reviews of
Geophysics, 39(3), 413–445, 2001. doi: 10.1029/2000RG000085.

A. Worby and I. Allison. A Technique for Making Ship-Based Observations
of Antarctic Sea Ice Thickness and Characteristics: PART I Observational
Technique and Results. Research Report 14, pp. 1-23, Antarctic Cooper-
ative Research Centre, University of Tasmania, Hobart, Tasmania, Aus-
tralia, 1999.

A. P. Worby, C. A. Geiger, M. J. Paget, M. L. Van Woert, S. F. Ack-
ley, and T. L. DeLiberty. Thickness distribution of Antarctic sea ice.
Journal of Geophysical Research: Oceans, 113(C5), C05S92, 2008. doi:
10.1029/2007JC004254.

Cryosat-2 was intended to measure surface elevation changes of the world’s land-ice, including glaciers, ice caps and ice sheets. The on-board instrument used for measuring this is a Ku-band radar altimeter. A question posed at the onset of the mission was, how will penetration of the radio waves into the surface affect the measurement of surface elevations? This in turn leads to the question, if penetration is variable over time, what implications does this have for estimates of surface elevation changes?

Here we show time series of elevation changes estimated from the Cryosat-2 SARIn mode around the margins of Greenland. The results indicate that there is indeed variable penetration which varies in an expected manner with latitude and elevation, according to the seasonal changes in the surface conditions. Thus, signal penetration is an issue when measuring surface elevation changes with Cryosat-2. However, seen in the perspective of a multi-annual time series, it also provides valuable additional information on the surface conditions and progression of seasonal changes. This is demonstrated for some areas studied where the monthly Cryosat-2 “height” change correlates strongly with the seasonal surface change and provides additional information on the extent and magnitude of melt. Our results show the potential of Cryosat-2 for deriving both seasonal and multi-annual surface elevation changes for parts of Greenland which have been shown in previous work to be losing mass to the oceans.

Glacio Isostatic Adjustment (GIA) has, until recently, been estimated using forward models that attempt to determine how the solid Earth responds to changes in ice-ocean loading through time. These models require knowledge of spatially-varying Earth rheology, including mantle viscosity, and ice load history, both of which have large uncertainties for Antarctica. Recent advances in GIA models include consideration of three-dimensional variations in Earth rheology and consideration of power-law rheologies. Such GIA models predicta remarkably different patterns of uplift over Antarctica when compared to predictions using one-dimensional Earth models, such as a shift in the uplift maximum from the Ross to the Wedell Sea (van der Wal et al., 2015). However, large uncertainties still remain inthe ice loading history models (A. et al 2014 and van der Wal et al., 2015) and substantial differences are found between Antarctic reconstructions.

An alternative approach is to use observations of crustal motion from GPS, combined with mass trends from GRACE to invert for GIA. However, this is an undetermined problem which requires assumptions on the density profile of the ice column for which numerical models have been commonly used (Gunter el al., 2014). Here, we present a novel solution to the inverse problem using state-of-the-art methods in statistical modelling of spatio-temporal processes. Specifically, we combine observational data, including satellite radar and laser altimetry, GRACE, GPS and InSAR, with prior information on the process’ spatial and temporal smoothness to solve, simultaneously, for ice mass trends and GIA (results on ice mass trends for Antarctica will be presented elsewhere at this conference). In addition, we assess the most recent forward and inverse GIA solutions on a basin-scale for Antarctica based on a comparison with a set of GPS-observed vertical velocities.

Cryosat is the first ESA satellite mission dedicated to the study of the Cryosphere.
Its primary goal is to monitor the thickness of land ice and sea ice and help explain the connection between the melting of the polar ice and the rise in sea levels and how this is contributing to climate change. The main instrument is the SAR/Interferometric Radar Altimeter SIRAL-2
It operates in threee modes:

Low-resolution mode (LRM) like a conventional altimeter, on land ice or sea which are composed of few rough surfaces,

SAR mode (Delay Doppler Altimetry) operating a high resolution measurement on sea ice,

SAR interferometer (SARIn) mode operating on rough surfaces like on the sea ice/land limit.

This 3 mode capability allows to investigate in detail the advantages of each mode for the detection of icebergs. The method of detection of icebergs using conventional altimeters data was developed by Tournadre et al(2008) and has been used to estimate the icebergs climatology (Tournadre et al 2012). It has also been applied to the Cryosat LRM data within the ALtiberg project (Tournadre et al 2015). The presentation will focus on the iceberg detection and characteristics determination using DDA and SARin data.

The method of detection of icebergs using conventional altimeters data was developed byDelay Doppler Altimetry (DDA), proposed by R.K. Raney (1998), offers improved altimetric precision and better along-track resolution than conventional pulse limited altimeters. DDA altimeters have a high pulse repetition frequency to ensure pulse-to-pulse coherence, leading to an along-track resolution about 300 meters, improved signal-to-noise ratio and enhanced altimeter ranging performance. The delay doppler maps (DDM), the beam stacking capabilities and the improved along-track resolution offers new possibilities for the detection and the determination of the characteristics of both icebergs and ships. Using Cryosat SAR L1b and L2 date, a method of detection of target emerging the sea has been developed. Within the pseudo LRM waveforms obtained with  the range correction applied, the signatures of iceberg are bright spots  that can also be easily detected by classical image processing methods. The advantage of this method is to allow the determination of the size of the echo as well as the detection of collocated multiple echoes. Using the estimate of the iceberg's position the 85 Hz DDM  are then used to estimate the precise iceberg area by averaging the DDM taking into account the iceberg displacement in both range and doppler. This allows to compute an image of the iceberg backscatter at high resolution.
The new ESA SARVATORE project processing  allows the production of higher resolution waveforms that include more than 300 range bins in the waveforms noise part. (i.e. above the mean sea level). These waveforms can be used to improve the  detection by increasing more than twofold the detection swath of the altimeter. Several examples of such detection of icebergs  are analyzed and presented.
DDA altimetry offer improved capabilities of both iceberg and ship compared to classical LRM altimetry as it is demonstrated using a limited Cryosat SAR dataset. In the perspective of Sentinel 3 it can be of importance for the scientific community whose interest for icebergs and its impact on both the southern ocean circulation and ecosystems has increased during the recent year to develop an operational processing chain design to detect icebergs.
The SAR interferometer mode uses the phase difference between returning radar waves: in order to measure the arrival angle, a second receiver antenna is activated so that the radar echo is received by two antennas simultaneously. When the echo comes from a point not directly beneath the satellite there is a difference in the path length of the radar wave, which is measured by the phase difference. Simple geometry provides the angle between the baseline joining the antennas and the echo direction. An iceberg being a target emerging from the sea surface its echo lies within the noise part of the waveforms where the power received by the two antennas is incoherent. The phase difference has thus a zero mean as well as zero coherence. Within this mode, an iceberg can be simply detected using either a parabola detection in pseudo LRM not corrected for range or a bright spot detection within the range corrected waveforms. The signature are also characterized by a high (>0.8) coherence between the two antennas and a non zero phase difference. The phase difference is  converted to incidence angle using the interferometric equation and thus, knowing the satellite altitude,  to distance from nadir. As the range of the iceberg's signature (i.e. the range bin where it is detected) depends on the distance from nadir and the freeboard elevation,  the icebergs freeboard is estimated by solving a simple equation. The iceberg area  can also be estimated using the availbale DDM. The SARin mode allows to estimate all  iceberg's characteristics: area, freeboard, surface backscatter with an unprecedented precision. Using the limited available data set of SARin data an operational algorithm of iceberg detection and analysis is presented.

Eurasian lakes and internal seas are an integrator of climate processes and an indicator of existing or potential climate changes. Variability of ice and snow regime is important for their physical, chemical and biological properties, and for human activity (navigation, transport, fisheries, tourism etc).

We present studies of ice and snow cover of the Caspian and Aral seas, Baikal, Ladoga and Onega lakes using synergy of simultaneous active and passive satellite microwave observations (TOPEX/Poseidon, GFO, Jason-1 and -2, ENVISAT, SARAL/Altika). An ice discrimination approach is presented and evolution of ice conditions from historical data and satellite observations is analysed. We present results of our field studies on lakes Ladoga, Onega, Baikal (Russia) and Hovsgol (Mongolia) and their relation with radar altimetry observations and optical satellite imagery. We also address formation of giant (diameter 5-7 km) rings on Lake Baikal ice. We analyse the timing of and duration of their existence as well as associated ice and water column structure, and suggest a new explanation of their origin.

This research has been done in the framework of the CNES TOSCA "Lakes", CNRS PICS "BaLaLaICA", RFBR 13-05-91051, ERA.NET RUS Plus S&T #226 “ERALECC”, Swiss-Russian "Lake Ladoga: life under ice", FZP 1.5, ANR "CLASSIQUE", IDEX Transversalité 2013 InHERA, and FP7 MONARCH-A projects, as well as by GDRI "CAR-WET-SIB" and French-Siberian Centre for Education and Research.

Sea ice thickness (SIT) data is important for climate modeling and ship navigation in the polar regions. Thickness of thin sea ice is provided by satellites using L-band passive microwave radiometers.For this study we investigate two SMOS algorithms for thickness of thin sea ice, one using mainly the intensity at low incicence angles from 0° to 40° [1], and one using intensity and polarization difference at higher incidence angles between 40° and 50° [2]. Both methods have been validated, yield similar results and are being used to produce daily thickness maps of thin sea ice up to about 0.5 m thickness. At L-band (1.4 GHz) the atmosphere has a low impact on the brightness temperatures, but other factors such as Faraday rotation, ascending/descending overpasses inhomogeneities and extraterrestrial radiation sources such as sun glint and galactic radiation can change the SMOS brightness temperatures and retrievals based upon them. Although the sea ice thickness retrieval has less strict brightness temperature accuracy requirements than that for the sea surface salinity, an investigation of the impact of these unwanted influences, together denoted as geophysical noise, on the retrieval is required.

Microwave radiation passing through the ionosphere can undergo a rotation of the polarization vectors. As the sea ice thickness retrieval algorithm uses polarization difference as one of the input parameters, correcting for the Faraday rotation is important. The algorithm uses the Faraday rotation data provided in the L1C data product which is based on a forecast state of the ionosphere. A comparison was done with the retrieved SIT values using a higher accuracy Faraday rotation data, and with no correction at all, to see if the impact on the retrieval is significant. For most cases the error from Faraday rotation is under 1 K meaning less than 1 cm in error for retrieved SIT.

The retrieval algorithm uses daily means for the gridded brightness temperatures. Differences between the brightness temperatures recorded by SMOS over the same areas during ascending and descending overpasses and their impact on the retrieval are determined. Due to SMOS having a large area snapshot and the RFI filter removing complete snapshots when at least one pixel is compromised, the unequal distribution of RFI compromised snapshots between ascending and descending overpasses can induce biases in the daily mean brightness temperature. Moreover, Areas with high brightness temperature gradient, such as ice edges or coastlines also show different brightness temperature values between ascending and descending overpasses.

The extraterrestrial microwave radiation sources at L-band are the sun and galactic radiation. Sun glint is a factor influencing the retrieved sea surface salinity [3], with brightness temperature values reaching up to 10⁷K in the L-band. Although SMOS has an orbit which minimizes sun glint contamination, reflection and scattering over sea ice might be possible. Galactic radiation has several sources, one being uniform, cosmic background radiation, while the other two, hydrogen emission lines and continuum radiation are variable. Thus the incoming radiation varies with day of the year and position of the satellite on the orbit. The Bidirectional Reflectance Distribution Function (BRDF) is investigated for sea ice, for the two cases of specular and Lambertian reflextion.

 

References

[1] Kaleschke, L., X. Tian-Kunze, N. Maaß, M. Mäkynen, and M. Drusch (2012), Sea ice thickness retrieval from SMOS brightness temperatures during the Arctic freeze-up period, Geophys. Res. Lett., doi:10.1029/2012GL050916.

[2] – M. Huntemann et al., “Empirical sea ice thickness retrieval during the freeze-up period from SMOS high incident angle observations”, The Cryosphere, 2014, pp. 439-451

[3] – Simon H. Yueh et al., “Error Sources and Feasibility for Microwave Remote Sensing of Ocean Surface Salinity”, IEEE Transactions on Geoscience and Remote Sensing, 2001, pp. 1049-1060

The National Ice Center (NIC) is a U.S. Government agency that brings together the Department of Defense - Navy, Department of Commerce – National Oceanic and Atmospheric Administration (NOAA), and the Department of Homeland Security – U.S. Coast Guard (USCG) to support coastal and marine sea ice operations and research in the Polar Regions and other subpolar ice infested waters as well as provide global snow cover analysis.   The main users of NIC products are the NOAA National Weather Service, U.S. Navy submarine force, U.S. and Canadian Coast Guard icebreakers, U.S. Military Sealift Command on re-supply missions to Antarctica and Greenland, and NOAA research vessels operating near the sea ice cover in both hemispheres.  Time-series of weekly or bi-weekly ice analysis charts, daily marginal ice zone and ice edge products, and tactical support annotated imagery and forecasts are generated by NIC expert analysts supporting global sea ice domain awareness. 

Near real-time observations from a variety of satellite sources in addition to in-situ observations and numerical weather prediction model output are required to produce robust sea ice and snow analyses and forecasts.  Still, access to all-weather, day and night high-resolution synthetic aperture radar (SAR) wide coverage data streams is required in order to provide the best sea ice analysis and tactical support possible.  The increasing access by the NOAA National Environmental Satellite, Data, Data, and Information Service (NESDIS) to ESA's Sentinel-1A SAR data and the use of this exceptional data stream in NIC operational analysis and other products will be discussed.  

Svalbard-type glaciers, also called polythermal glaicers, are still melting inside during the winter time. The water runoff from the polythermal glaciers usually refreeze into ice in the foreland. So some icing areas can be seen in front of the glaciers in spring time. There is a tipical ice-pond appeared every spring in front of Pedersenbreen, Svalbard. Obviously this ice-pond represents the water runoff due to the sub-ice melting inside Pedersenbreen through winter. Once the volume of these ices is calculated, the amount of sub-ice melting in the winter time can be inferred.

In April 2014, a RTK-GPS measurement was carried out around the ice-pond in front of Pedersenbreen.  From which the area of the ice-pond was calculated into 0.0787 km², and the perimeter was 1.343 km at that time. The total area of Pedersenbreen is 6.2 km² at the same period, so we can infer that the mean net mass loss is about 2.4~3.6 cm in the winter time, assuming the mean ice thickness of the ice-pond at 2~3 m. While the mean net mass balance of Pedersenbreen is about -0.32 m per-year between 1990 and 2009, the sub-ice melting accounts probably for 15% or more of the net mass balance of a whole year. With the high precise measurement of ice-free foreland topography, the contribution of the sub-ice melting will be calculated more accurately.

With the climate change, the sub-ice melting will probably change in the future. Then the icing amount will also be dynamic. So, monitoring the icing dynamics is a possible way to disclose the interaction between the glaciers and the local climate.

The estimation of lake water level variations with satellite altimetry is a challenging task because the majority of altimeter waveforms of smaller lakes are contaminated by land. The smaller the surface footprint size of the satellite altimeter the less the altimeter waveforms are corrupted. In this study altimeter measurements from the three satellite missions Envisat, CryoSat-2, and SARAL (Satellite with ARgos and ALtiKa) are used to determine long-term water level time series of smaller lakes. Therefore, we have to deal with different satellite missions and sensor characteristics. While the satellites Envisat and SARAL/AltiKa follows the same orbit with a ground track coverage every 35 days, the CryoSat-2 satellite flies on a different orbit with a ground track repeat cycle of 369 days and a sub-cycle of 30 days. Thus, Envisat and SARAL/AltiKa measurements have a higher temporal resolution and CryoSat-2 measurements have a higher spatial resolution. The Envisat altimeter has a significantly larger puls-limited footprint size (up to 10 km) than the AltiKa altimeter (up to 8 km). For the SAR (Synthetic Aperture Radar) mode the SAR Interferometer Radar Altimeter (SIRAL) on board the CryoSat-2 satellite has a remarkable smaller pulse-Doppler-limited footprint size of ∼0.3 km along-track and ∼1.7 km across-track. Depending on the observation technique and surface conditions the waveforms variegate from quasi-specular over rivers, to multiple peaks over land and brown like waveforms in the center of the lake. In this study, the waveforms of three different satellite missions Envisat, CryoSat-2, and SARAL/AltiKa are retracked and combined in an optimal way to obtain improved long-term water level time series of smaller lakes, such as lake Tonle Sap in Cambodia or lake Vänern in Sweden. Beside the retracking the inter-mission offsets have to be considered. We will use Cryosat-2 Baseline B products as well as Baseline C products in order to analyze the improvements. Furthermore, we will study the accuracy and reliability of the different data sets over small lakes. The focus will be on SAR measurements which should be superior to classical sensors. These new findings can be transferred to the upcoming Sentinel-3 Mission.

Rock glaciers (RGs) are creeping permafrost landforms consisting of mixtures of unconsolidated rock debris and ice, and generally form in high mountains. Monitoring the dynamic of RGs is valuable for inferring the distribution of mountain permafrost and understanding the potential effects of ongoing climate changes on such a landform. RGs are widely distributed in high mountains in western China (e.g., Tien Shan, Tibet Plateau, Himalaya). However, they have been ignored for more than 20 years with only a limited number of geomorphologic studies conducted in the 1980s and early 1990s.

With availability of the snow-free ALOS Phased Array type L-band Synthetic Aperture Radar sensor (PALSAR) images acquired between 2007 and 2010, we applied a variety of SAR techniques to investigate the dynamics of the RGs in Eastern Tien Shan. These included the conventional two-pass differential interferometric synthetic aperture radar (DInSAR) technique to map the surface displacements of RGs in the line of sight (LOS) direction. For the RGs with aspects in nearly north-south direction, we performed the multiple aperture interferometry (MAI) to map the displacements in satellite flight direction (i.e., along track). Additionally, we applied the pixel offset tracking (POT) technique to obtain the flow displacements of RGs in cases of significant loss of interferometric coherence. We also calculated time series of surface motion at selected high coherence points using the small baseline subset (SBAS) inversion method.

Based on the surface displacements measured by the DInSAR, MAI and POT, we make an inventory of the active RGs for the study area of about 70 km×100 km. We found more than 15 active RGs, which are located between 3070 m and 3760 m, with a mean elevation of about 3494 m. Most of the identified active RGs are tongue-shaped, with a mean length of 892 m and mean width of 239 m. Our measurements show that the annual motion speed of the active RGs ranges from 10 to 90 cm/year.

Our study demonstrates the use of the satellite SAR interferometry to investigate the dynamics of the RGs, which contributes a good data source for understanding the kinematics, evolution of the RGs at the lower boundary of permafrost distribution. We will use of climate data (i.e., ground and air temperature, precipitation), to interpret the temporal changes of RGs and to assess the relationships of RGs flow with regional climate and local conditions.

Going above and beyond its ice-monitoring objective, CryoSat ,is also a valuable source of data for the oceanographic community, improving the existing altimetry dataset by reaching closer to the north and south poles thanks to its unusually high-inclination orbit. Since April 2014, the Cryosat-2 ground segment delivers dedicated L1b and L2 Ocean products with accurate geophysical parameters such as the Sea-Level Height, the Significant Wave Height, and the Wind Speed. These Ocean products are available 2-3 days after the sensing (IOP) and in delayed time with the most accurate orbit and range corrections (GOP, 30 days). These new data help to bridge the gap between previous ocean-oriented altimetry missions and the future Sentinel-3 mission being developed for Europe’s Copernicus program.Based on the outcomes from 2 years quality control and validation activities, R&D projects as well as the scientific exploitations of CryoSat data over the ocean, ESA intends to upgrade the CryoSat Ocean processing chain with a set of evolutions of highest interest among which: improved processing of the SAR and SARin altimetry data with the SAMOSA retracker, new ocean products and parameters available in real-time, improved wet tropospheric correction from data combination (Dcomb), upgraded geophysical models and new quality indicators. These new and improved ocean products would help scientists to better address key oceanographic issues, ranging from regional sub-mesoscale to open-ocean large-scale processes.

Permafrost and seasonal frost affects large parts of the earth surface and therefore is an important variable in climate research and related disciplines. The phase change near the surface can be detected with microwave data. Coarse resolution data provide good sampling rates what allows to study freeze/thaw cycles. Many applications do however require better spatial detail. The objective of the study is to assess the spatial heterogeneity of freeze/thaw using C-band and the potential added value of SAR data, especially with respect to Sentinel-1.

 

This study compares two data sets with different spatial and temporal resolutions considering terrain and landcover of permafrost regions. The data sets used were produced within the ESA DUE Permafrost project and cover a spatial resolution of 1 km (ENVISAT ASAR GM) and 25 km (ASCAT instrument on board the MetOp-A satellite) and a temporal resolution of daily, weekly and monthly. These products are based on different  methods. The SAR datasets are based on an temporal edge detection approach. The ASCAT product is an application of an empirical threshold-analysis algorithm to the backscatter data designed for global application. It is known that the thematic accuracy is lowest during the transition periods.

 

The study area extent is limited to the extent of the SAR data. This includes the Laptev Sea Coast, the Yamal region, central Yakutia, Alaska North Slope and MacKenzie region. The overlapping period is limited to 2007 to 2010. In addition in situ records available from the permafrost monitoring networks (GTN-P) are utilized.

 

The information available from the SAR data is the day of year of thaw and freeze-up in the autumn. The ASCAT dataset provides a surface status for each day. The fraction of frozen ground in SAR is assigned to the ASCAT grid and transformed to daily time slices for comparison. The number of acquisitions used in the single SAR maps (which is crucial for the accuracy of the product) are included in the assessment. Results show different patterns for spring and autumn which are assessed using ancillary information. The subgrid patterns also impact the agreement with in situ temperature records. Results are discussed with respect to the used wavelength (C-Band), noise (specifically an issue of the SAR product), the methods which have been used to produce the SAR and scatterometer products, and eventually the applicability of the coarse resolution dataset in complex terrain. 

CryoSat was launched on April 8^th 2010 and is the first European ice mission dedicated to the monitoring of precise changes in the thickness of polar ice sheets and floating sea ice. CryoSat is also the first altimetry mission operating in SAR mode and continuous improvements in the Level1 Instrument Processing Facility (IPF1) are being identified, tested and validated in order to improve the quality of the Level1B products.

The current IPF, Baseline C, was released in operation on April 1^st 2015. The second CryoSat reprocessing campaign was jointly initiated, aimed at reprocessing all previous data, since July 10^th 2010, in Baseline C. Reprocessed data will be released to the users on a monthly basis, as soon as generated and validated.

The purpose of this poster is twofold: first, to summarize the main L1B quality improvements introduced by baseline C processor and second, to present the refinements introduced in the reprocessing campaign.

The main quality improvements introduced in baseline C involved the generation of the SAR/SARIn waveforms (stack weighting), the storing of the SAR/SARIn waveforms in the L1B product (doubled the number of samples), the improvement of the stack characterization parameters, the removal of datation and range biases and the accuracy of the attitude information. For the latter, a dedicated processor has been developed.

The reprocessing campaign introduced a further refinement involving the calibration corrections: the long-term calibration models. While it was not possible to introduce such corrections in operation, the calibration models have been generated and used for the whole period covering the reprocessing campaign.

This poster will describe and show examples of the baseline C L1B quality improvements plus the calibration results from the reprocessing campaign. Moreover, a dedicated analysis of the new Star Tracker products will be shown, using the re-processed data since the beginning of the mission.

Sea ice is an important component of the climate system. It is a key component in the heat exchange between the ocean and the atmosphere. It is also a sensitive indicator of climate change. Accurate information about the extent and thickness of the sea ice is therefore crucial in understanding the changing climate. Furthermore, with the rapid melting of the sea ice, ship traffic and offshore activity is expected to increase in the Arctic. In order to maintain the safety of vessels in Arctic waters, accurate and timely information about the sea ice is of great importance.

With the gradual reduction of the sea ice minimum area along with the changing climate, an increasingly larger part of the sea ice consists of seasonal first year ice. A service estimating the ice thickness of the thin, seasonal sea ice would therefore be of high value.

ESA’s Sentinel-3 satellites, under the Copernicus program, will provide a new opportunity for monitoring sea ice. The Sentinel-3A satellite is scheduled for launch in December 2015, while the second satellite, Sentinel-3B, is planned to follow 18 months later. Among other instruments the satellites will carry the Sea and Land Surface Temperature Radiometer (SLSTR). This sensor will continue the legacy of ENVISAT AATSR. But with the improved swath width and the twin satellite constellation, SLSTR will provide greatly improved coverage.

The thermal bands of SLSTR are very suitable for estimating the surface temperature of ice and snow, which may be used to estimate the thickness of thin sea ice by applying the model by Yu and Rothrock (1996). Here, the heat balance on the ice surface is described as a sum of component heat fluxes. With estimates of the surface temperature, along with atmospheric data (air temperature, wind speed) it is possible to model the individual heat fluxes to describe the thermal balance.

For cold winter conditions, the heat balance is strongly influenced by the heat transfer from the water surface to the top surface of the ice. This can be assumed to follow a linear dependency on the ice thickness. Inverting this relation allows us to estimate the thickness of the ice.

We will here present an automatic algorithm for producing sea ice thickness charts from Sentinel-3 SLSTR data. The spatial resolution of the product will be similar as for the thermal bands of SLSTR, 1 km, allowing view of the leads in the ice.

We will also present and discuss some first results, and we will discuss the future possibilities for sea ice thickness monitoring with Sentinel-3.

For the Greenland ice sheet, the period from 2007-2015 is characterized by a large variability in climate conditions, with record-breaking melt in 2010 and 2012 and minor net mass loss in 2013.   

As a part of the Programme for Monitoring of the Greenland Ice Sheet (PROMICE) funded by the Danish Energy Agency, repeat airborne LiDAR surveys have been conducted in 2007, 2011 and 2015. The surveys were conducted around the entire margin of Greenland ice sheet. In addition to LiDAR measurements, ice-penetrating radar measurements were carried out in 2007 and 2011, to estimate ice sheet thickness along the flight-path. As the only one of its kind, the repeat surveys have been performed in late summer, to coincide with the end of the balance year. With the third survey successfully conducted in 2015, a unique opportunity for evaluating elevation changes at a consistent and tailored circum-Greenland fluxgate consist with this PROMICE dataset.  

Here, we combine available satellite observations of elevation changes of the Greenland ice sheet, with the PROMICE dataset, to estimate the state of the Greenland ice sheet seen from the PROMICE fluxgate. The PROMICE LiDAR snapshots will guide the contemporary observations from satellites to target the balance year.          

SnowPEx is an is an international collaborative effort for evaluation and comparison of continental to hemispherical satellite snow products funded by the European Space Agency (ESA) under the Quality Assurance framework for Earth Observation (QA4EO).  Here we explain the protocol for snow product validation using in-situ data and show results from in-situ validation of Snow Extent (SE) products. 14 different snow extent products from different providers in Europe, Japan, U.S. and Canada are participating in the validation exercise. The products feature daily snow information in different resolutions and projections, the investigated time period varies from one yearly season up to five, depending on the product. The reference in-situ data set is a compilation of a massive set of snow observation from different institutes and research groups in Europe, Asia and Northern America. Satellite snow products as well as in-situ data have been converted to a consistent format defined by the SnowPEx team, enabling the easy handling of the datasets. The validation of in-situ measurements aims at a better understanding of the accuracy of each product, and importantly, tries to recognize the temporal/spatial regions where the accuracy differs from a product’s general performance. The in-situ observations are distributed so that the product performance over different land cover and different snow zones can be investigated.

 

The validation work follows the SnowPEx protocol, producing standardized measures of the performance of snow products. To accomplish this, all products and in-situ data are treated as binary (snow/no-snow) information. The SE products featuring SCF are converted to binary using different thresholds for SCF: SCF=25% and SCF=50%. Both ways are analyses separately in the validation work. Originally binary products are used directly in the validation. For in-situ data typically including measurements on Snow Depth (SD), binary conversion relies on three different thresholds: SD=0%, SD=2cm and SD=15cm, which are separately analyzed. The in-situ observation is labelled ‘snow’ the threshold is exceeded, otherwise it is labelled ‘no-snow’.

After these conversions the in-situ observations are associated with temporally and spatially matching product pixels. To diminish the possible inaccuracies of the product geo-location, the in-situ/product pairs are accepted for the analyses only if for the in-situ location, the overlaying pixel and its three nearest neighbors provide the same binary information. The generated dataset includes thousands of cases, from which several validation measures are calculated, amongst them at least Hit-rate, False alarm rate, Precision and F-score. These are stratified according to the Sturm snow zones and additionally according to the different land-cover classes (forested plains, non-forested plains, forested mountains, non-forested mountains, as indicated by the ESA GlobCover data). Since it is expected that the performance of a snow product varies during the year, the binary measures are produced for four three-month periods: October-December, January-March, April- June, July-September. This approach benefits the interpretation of the results and helps us to comprehend when and where each product performs the best.

Earth observation from space is the most suitable approach to determine land-ice mass changes and contribution to sea level rise at a global scale. Interferometric SAR altimeters like CryoSat-2 are one of the most promising tools for this purpose, but precise applications over glaciers and ice sheets can be hampered by variable signal penetration and backscatter in snow and ice, particularly during the transition from a cold winter snowpack to a melt-affected summer snowpack. In this study, we use dense networks of GPS surface profiles to validate CryoSat-2 elevation estimates from ESA's Level 2 product and alternative processing techniques using Level 1b data. We consider surface elevations from both the point-of-closest-approach (POCA) and the full swath of the interferometric mode. We investigate the variability in signal penetration by analysing the seasonal evolution of derived surface elevations in respect to climate parameters. The main study targets are Austfonna ice cap on Svalbard and three ice rises in Dronning Maud Land, East Antarctica, where extensive ground-truth data are available. We find that the results of POCA and swath processing complement each other very well: POCA locations concentrate along ice divides whereas the swaths give good coverage in gentle slopes which are not resolved by POCA. A combined approach has the potential to improve current DEMs in coastal Antarctica and increase the coverage of elevation-change measurements at the margins of ice sheets and glaciers where most of current mass losses take place. We test this potential for the entire glacier region of Svalbard and for ice rises and promontories along the coast of Antarctica. We account for uneven spatial sampling and make assessments of regional volume change and mass balance.

The isotope-geochemical research was carried out in the different parts of the Tavan Bogd mountain massif during the SPbSU expeditions. This massif is located on the border between Russia, Mongolia and China. Field work was carried out in Mongolia in 2013 and 2014 and in Russia in 2015.

The task was to obtain the isotopic characteristics (deuterium and 18O concentrations) in different parts of glacio-nival systems. The main objects of the research were: precipitations, glacier runoff and snow-firn thickness. More than 300 samples had been collected during 3 years. Samples from the last expedition are being analyzed at this moment.

The handling of river water samples presents data about the isotopic content of melting glacier water and its changes according to the river flow. The average melting water d18O was -17,3 ‰. The analysis of the water samples showed, that the isotopic content of river water didn’t change along 30 km from the edge of the glacier. It may relate to the big runoff from glacio-nival objects in the mentioned part of the river basin.

Stable isotope records in snow and firn may present information about accumulation and about the origin of precipitations which take part in accumulation. For this kind of research the samples were collected from 2 snow-pits at an altitude 3400 meters in Kozlov Glacier (Mongolia) accumulation zone and from 2 snow-pits at an altitude 3400 – 3650 meters in Russian part of this massif. The depth of these snow-pits varied from 2,6 to 1,2 meters, samples were taken after every 5 cm.

In the accumulation zone positive temperatures and snow melting occur during the ablation season. Such conditions are unfavorable for isotope-geochemical research, because isotopic homogenization may happen. However, it was found out, that the fluctuations of water stable-isotope records partly show preserved seasonal variations. Therefore, it is possible to establish, during which season the maximum accumulation takes place and what is the main precipitation source. 

 The tentative supposition was formulated: the autumn snowfalls take the biggest part in the accumulation, as spring and summer precipitations melt in ablation season. The assumption was based on the records, which showed that isotopic easy firn occurs in the upper part of snow-pits.

 At this moment the further analysis of the collected samples is being proceeded and new data is expected to be found out.

Following the disintegration of the Larsen A and Larsen B ice shelves in 1995 and 2002, respectively, the stability of Larsen C gained a lot of attention as next potential candidate. The George VI ice shelf is situated at the other side of the Antarctic Peninsula and is also one of the largest shelves in this region. They are both exposed to an increase in surface air temperature and strong bottom melt in recent years.

Those are the two main factors which are expected to favor an ice shelf collapse. Increasing air temperatures can lead to stronger surface melt and subsequent hydro-fracturing, while excessive bottom melt rates account, combined with calving, for nearly the entire ice shelf mass loss. New L-band measurements from ESA's SMOS mission at the George VI and Larsen C ice shelves show a common unique pattern. It can be characterized in the cold season by low brightness temperatures (Tb), high polarization differences and, most striking, a substantial decrease of the Tb in the last five years which occurred in case of the Larsen C mainly near the grounding line. The interpretation of passive L-band measurements from ice shelves is challenging and has yet not been done in literature.

Summer time surface melt can influence the winter emissions in two ways: 1) Melt water can drain into the snow which acts as insulator when the temperature drops. The snow temperature, even tens of meters below the surface, can therefore be significantly higher than the annual mean air temperature if melt conditions are reached regularly in summer. The warm snow will therefore emit more radiation. But then there is also a higher probability of ice layer formation from melt water. Those layers can reflect the radiation from below and thereby reduce the surface emissions recorded by SMOS. The penetration depth at L-band in fresh water ice is of the order of hundreds of meters. It is therefore also possible that we can see a contribution from below the shelf. Seawater has a low emissivity, because of which a thinning of the ice shelf can be associated with a reduction in Tb. This relation could be amplified in case of high oceanic melt rates, as the warmest ice is in general found at the bottom of the shelf. At least at the Larsen C ice shelf marine ice could also influence the L-band signal. This ice is formed from seawater which is frozen to the bottom of the shelf and is expected to be mostly opaque and to have a much higher emissivity than water. A retreat of marine ice could therefore also result in a reduction of the SMOS Tb.

An incoherent numerical emission model for layered snow with smooth interfaces has been adapted to ice shelves and the mentioned processes have been implemented. A statistical stratification model is used to represent the variability in near surface density profiles. The emission model can reproduce the recent changes in SMOS measurements by increasing the density stratification, however the general low Tb at the shelves is not captured in a reasonable range of parameters. This discrepancy could, for example, indicate that the used representation of the absorption coefficient is too high, but further investigations have to be done. It is worth mentioning, that all described processes likely to lead to the observed reduction in SMOS Tb can be linked to a potential weakening of the ice shelf stability.

The Arctic Ocean is under profound transformation. Observations and model predictions show dramatic decline in sea ice extent and volume. Despite its importance, our understanding of the pacing of Arctic sea ice retreat is incomplete largely due to a paucity of observations.

The launch of the Soil Moisture and Ocean Salinity (SMOS) mission, in 2009, marked the dawn of a new type of space-based microwave observations. Although the mission was originally conceived for hydrological and oceanographic studies [1,2], SMOS is also making inroads in the cryospheric sciences. SMOS carries an L-band (1.4 GHz, or 21-cm wavelength), passive interferometric radiometer (the so-called MIRAS) that measures the electromagnetic radiation emitted by the Earth’s surface, at about 50 km spatial resolution, full polarization, continuous multi-angle viewing, large wide swath (1200-km), and with a 3-day revisit time at the equator, but more frequently at the poles.

A significant difference of the L-band microwave radiometers with respect to higher frequency radiometers, such as SSMI/AMSR-E/AMSR-2, is that the former can also “see through the ice.” That is because ice is more transparent (i.e., optically thinner) at 1.4 GHz than at higher frequencies (19-89 GHz). In radiometric terms, the brightness temperature measured by an L-band antenna radiometer does not correspond to the emissivity of the topmost surface layer but of a larger range of deeper layers within the ice (about 60 cm, depending on ice conditions). Thanks to that increased penetration in the medium, L-band radiometers can provide, for the first time, thin ice thickness from space [3, 4].

A novel radiometric method to determine sea ice concentration (SIC) is presented. The method exploits the distinctive radiative properties of sea ice and seawater when observed at low microwave frequencies and from a range of incidence angles, to discern both media. The Bayesian-based Maximum Likelihood Estimation (MLE) approach is used to retrieve SIC. The advantage of this approach with respect to the classical linear inversion is that the former takes into account the uncertainty of the tie-point measured data in addition to the mean value, while the latter only uses a mean value of the tie-point data. When thin ice is present, the SMOS algorithm underestimates SIC due to the low opacity of the ice at this frequency. However, using a synergistic approach with data from other satellite sensors, it is possible to obtain accurate thin ice thickness estimations with the Bayesian-based method.

Despite its lower spatial resolution relative to SSMI or AMSR-E, SMOS-derived SIC products are little affected by the atmosphere and the snow (almost transparent at L-band). This new dataset can contribute to complement ongoing monitoring efforts in the Arctic Cryosphere.

 

[1] Font, J.; Camps, A.; Borges, A.; Martí andn-Neira, M.; Boutin, J.; Reul, N.; Kerr, Y.; Hahne, A. & Mecklenburg, S. SMOS: The Challenging Sea Surface Salinity Measurement From Space Proceedings of the IEEE, no. 5, 2010, 98, 649 -665

[2] Kerr, Y.; Waldteufel, P.; Wigneron, J. P..; Delwart, S.; Cabot, F.; Boutin, J.; Escorihuela, M.J..; Font, J.; Reul, N.; Gruhier, C.; Juglea, S.; Drinkwater, M.; Hahne, A.; Martin-Neira, M. & Mecklenburg, S. The SMOS mission: New tool for monitoring key elements of the global water cycle Proc. IGARSS 2010, no. 5, 2010, 98, 666-687

 [3] Kaleschke, L., Tian-Kunze, X., Maaß, N., Mäkynen, M., and Drusch, M.: Sea ice thickness retrieval from SMOS brightness temperatures during the Arctic freeze-up period, Geophys. Res. Lett., doi:10.1029/ 2012GL050916, 2012.

[4] Huntemann, M., Heygster, G., Kaleschke, L., Krumpen, T., Mäkynen, M., and Drusch, M.: Empirical sea ice thickness retrieval during the freeze up period from SMOS high incident angle observations, The Cryosphere Discuss., 7, 4379–4405, 2013.

In the past years, supraglacial lakes came more into the focus of research due to their potential to drain on short time scales and deliver large amounts of water to the base of the glacier, enhancing basal sliding. Supraglacial lakes are categorized into three types, those that drain through cracks and moulins, those that drain at the surface and refreezing lakes. We use TerraSAR-X imagery to study the formation of supraglacial lakes on the 79°N Glacier in Greenland. For this purpose we study the radar backscatter intensity and its variation across lake sites. Additionally to that, we use airborne laser scanner data for surface topography and optical imagery from the board camera from flights performed in 2013 and 2015. Beside the seasonal evolution in 2015, we compare lake type, size, area and shapes.

 

We present the sea ice thickness and freeboard products of the ESA Climate Change Initiative (CCI) project. The products include gridded Arctic sea ice freeboard and thickness for the ice covered seas in the northern hemisphere. The freeboard and subsequent thickess esimates are derived from the radar altimeters ERS-2, Envisat and Cryosat-2 satellites during cold winter months from November to March since November 1995. The spatial coverage of the estimates is limited by the inclination of the satellite platforms (measurements were made only south of 82 N for ERS-2 and Envisat and 88 N for Cryosat-2).

We describe the products and present the technical challenges of constructing a consistent thickness time series from the three different altimeters. We present and justify the technical choices made, such as the chosen retracking scheme, as well as the used auxiliary data products such as the snow load, and sea ice concentration estimates used in the processing. We shall give estimates of sea ice thickness development in chosen areas for the 20 year timespan 1995-2015.  

As the Greenland ice sheet warms, a change in the structure of the upper snow/firn occurs. This change further induces changes in the reflective properties of the firn seen from satellite radar altimetry. If not identified as changes in the reflective properties of the firn, these may be interpreted as actual surface elevation changes seen from the satellite radar altimetry (Nilsson et al., 2015).

Here, we investigate how to correct the elevation change observed from the ESA Cryosat-2 radar altimetry mission to derive elevation change of the air/snow interface of the Greenland ice sheet. The elevation change of this “real” physical surface is crucial, if the goal is to derive Greenland mass balance as done for LiDAR missions.

The investigations look into waveform parameters to correct for the observed bias between Radar and LiDAR observations when using Croysat-2 level-2 data. Based on the knowledge gained by analyzing the elevation change derived from the inclusion of various waveform parameters, we pinpoint the challenges associated with the using Croysat-2 observation in mass balance studies. As for mass balance studies utilizing LiDAR observation (ICESat), a strong firn-modeling component is needed; here the information gained from running such models may also provide input for the correction of radar elevation in to surface elevation.    

 

Reference:

Nilsson, J., Vallelonga, P., Simonsen, S. B., Sørensen, L. S., Forsberg, R., Dahl-Jensen, D., Hirabayashi, M., Goto-Azuma, k.,  Hvidberg, C. S., Kjær, H. A.  and Satow, K. (2015). Greenland 2012 melt event effects on CryoSat-2 radar altimetry. Geophysical Research Letters, 3919–3926. doi:10.1002/2015GL063296

In this study we present an algorithm that automatically segments sea ice and extracts the sea-ice floe size distribution (FSD) from high-resolution Synthetic Aperture Radar (SAR) imagery. The algorithm implemented various image analysis methods. We implemented Kernel Graph Cuts to segment SAR imagery into water/ice regions, and then implemented the watershed to split touching ice floes followed by rule-based post-processing. We selected two typical summer cases for the algorithm validation, in which both very high-resolution visible and SAR imagery co-exist so that reliable ground truth data can be constructed. The validation reveals promising results as well as limitations of the automated algorithm. The algorithm tends to fail to detect small floes (mostly less than 100 m), compared to the ground truth data, however for floes larger than 100 m the algorithm-derived FSD is very consistent with the ground truth data. The algorithm is implemented to TerraSAR-X SAR imagery, and can be extended to other SAR imagery such as Sentinel-1.

Global snow cover is an important environmental parameter, as is influences hydrology, vegetation, radiation balance, and the living space of humans and animals. Snow is an essential source for freshwater in many regions of the world and at the same time, snow cover depends on precipitation and temperature during the snow season. As climate and weather varies, also the amount as well as the onset, duration, and offset of snow cover changes throughout the years. It is important to analyze this variability in order to identify possible trends, but also to predict the impact of the snow cover situation on local freshwater availability, floods, or the influences on vegetation.

The Global SnowPack is a set of global snow cover parameters that are suited to perform these analyses. It is based on time series of global, daily snow cover information derived from medium resolution remote sensing data. The aim is to include every possible data source available to offer high quality products for the whole globe which range back until the early 80s. Such a long time series is required in order to derive possible trends from the data. AVHRR, MODIS, MERIS, and the upcoming Sentinel-3 data are only some examples for sensors that are/will be included in the Global SnowPack products. These products consist of the overall snow cover duration per hydrological year (different for Northern and Southern Hemisphere), early season snow cover duration and late season snow cover duration. The processing steps that are performed to derive these products include temporal and spatial interpolation to estimate the snow cover status below clouds, during polar darkness, and during data gaps, as well as a snowline detection which is depending on the topography of an area. Landsat data is used to perform accuracy assessment.

The presentation will give a quick overview of how the processing steps are implemented before some examples are given on how the Global SnowPack datasets look like and what they can be used for. The locations of these examples will be situated in Central Asia, Western USA, and Europe. In Central Asia, snow cover changes have severe impacts on the local hydrology and the freshwater availability; in the Western USA, California has experienced several snow cover seasons with extraordinary low snow cover durations; in Europe, the snow cover characteristics are highly variable between the years and show no distinct trend in any direction.

POLAR ICE is a 2.5 year FP7 project supported by the European Commission. Key aims are to: (1) provide state-of-the-art ice information products, including new developments for ice pressure, ice forecasts and ice thickness, (2) deliver these products to end users through a single integrated system, taking account of end users’ needs (bandwidth limitations, regions of interest, subscribing to favourite products), (3) providing a visualisation system that allows a broad range of products to be displayed in a unified manner.

This presentation will include a discussion of the project’s experience in practically demonstrating the POLAR ICE system and products with a range of end users involved in a variety of different operations in the Arctic and Antarctic.

Mass changes seen by Gravity Recovery And Climate Experiment (GRACE) satellites have improved our knowledge about climate change. GRACE is by far the best space-borne sensor for mass change detection. In the last decade there have been many attempts to quantify the rate of ice mass loss in Greenland from GRACE observations. Every contribution in this direction reported a different number to the rate of ice mass loss. One of the reasons for difference was different filtering scheme. It is widely accepted that filtering introduces leakage error and change the signal. Thus, leakage analysis is important for trend studies. However, there is no contribution discussing change in the trend signal due to filtering. The methods used to improve the filtered solutions are either model dependent or follow an iterative scheme. Models have uncertainties and thus a model dependent approach increases the GRACE error budget. The iterative schemes work for a simulated environment, but their efficiency is highly vulnerable to changes in the inputs and simulation setup. We use a data driven approach to recover the signal loss due to filtering. The approach used is developed through a mathematical understanding of convolution on the sphere and we give a mathematical relation for change in trend due to filtering. We validate the method in a closed-loop environment, and discuss the importance of leakage analysis for trend studies. We find that due to filtering the ice mass loss rate over Greenland is underestimated by 13.6 gigaton yr^-1. Although our contribution does not resolve every uncertainty, but it certainly improves the \textsc{grace} product quality.

The Greenland ice sheet is a vast ice-covered plateau extending for about 1,700,000 km2 over 80% of the whole Greenland surface. It represents the second largest ice body of the planet after the Antarctica ice sheet and its characteristics are significantly affected by temperature changes. The knowledge of its properties can substantially contribute to better understand the arctic ecosystem and its responses to climatic changes. Melt phenomena have strongly increased in the last years, leading to modifications in the characteristics of the snow pack. Previous studies of the Greenland ice sheet led to the definition of different snow facies, depending on the amount of snow melt and on the properties of the snow coverage itself. C. S. Benson divided the ice sheet into four zones [1]: melt does not occur in the dry snow zone, which is situated at the highest altitudes at the center of the Greenland plateau. It is surrounded by the percolation zone, where a limited amount of melting occurs, leading to the generation of larger snow grains and to the formation of small ice structures, such as lenses and pipes, within the snow pack. The wet snow zone is located further down slope towards Greenland’s coasts, where previous year accumulation saturates with water during summer melt. Outer coastal regions are classified as ablation zone, where the previous year accumulation completely melts during summer, resulting in a surface of bare ice and surface moraine. Up to now, such facies have been located using microwaves radar sensors by estimating the backscatter levels of the reflected signal. An incident radar wave is able to penetrate the snow-pack, depending on the sensor parameters and on the characteristics of the ground. Interferometric synthetic aperture radar (SAR) acquisitions over Greenland ice-sheet are therefore subjected to volume decorrelation. The intensity of such phenomena can be associated to different dominant scattering mechanisms on ground and can therefore help to classify the type of snow coverage. In this paper we present an approach for classifying the different snow facies of the Greenland ice sheet by exploiting X-band TanDEM-X interferometric SAR acquisitions. TanDEM-X data is particularly suitable for this analysis due to the single-pass bistatic acquisition mode which
does not suffer from temporal decorrelation. As far as spaceborne SAR sensors is concerned, this data set is unique. Large-scale mosaics of radar backscatter and volume decorrelation, derived from the interferometric coherence, represent the starting point for applying an unsupervised classification method based on the c-means fuzzy clustering algorithm. The detected snow facies can then be related to the well-known Benson’s facies. Results show a good agreement with external snow melt data and independent multi-temporal TanDEM-X data. Moreover, given the proper location of the different facies, we estimate the X-band penetration depth into the whole ice sheet and finally, we analyze TanDEM-X interferometric time series, acquired over dedicated test sites located within the different snow facies, in order to monitor their evolution in time.

There are many advantages to the use of satellite radar altimeters for measuring height change on glacial ice caps: in particular frequent, reliable coverage is achieved independent of cloud or light conditions, and this cannot be matched by any other system. The problem lies with the bias that exists between the surface elevation and the height derived from the radar altimeter, and the fact that this bias can change with the processing method, and with the variable penetration of the microwave radiation depending on the conditions of the snow, firn and ice. For example, for many ice caps we anticipate a lower bias between the surface and Cryosat detected elevations under summer melting conditions than in winter when relatively transparent dry snow lies above the denser layer created during the previous melt season.

In this presentation we compare elevations derived from the SARIn mode of ESA’s Cryosat 2 satellite with elevations obtained from surface GPS transects and airborne laser altimeters flown by ESA in the CryoVex campaigns, and NASA in the IceBridge campaigns. Cryosat height results from various ‘point-of-closest-approach’ and ‘swath mode’ algorithms will be compared with surface heights for various areas in Greenland and also from Austfonna and the Devon Ice Cap, areas which were picked for Cryosat calibration and product validation.

The mean height difference, the standard deviation of the mean, and the shape of the distribution of the histograms of the height differences will be shown and compared for different areas and zones on the ice caps. Results will be discussed and interpreted based on conditions on the ice cap and the problem of comparing essentially point surface measurements with the radar altimeter ‘heights’ which are averaged over a complex shaped area and may contain a significant subsurface contribution.

Notwithstanding the complication of the variable bias it will be shown that the excellent temporal and spatial coverage possible with the Cryosat 2 SARIn mode allows a new and useful picture of both seasonal and year-to-year height change. In some cases, even for areas around Greenland, it is possible to estimate the height loss associated with summer melt as a function of elevation.

The periodic change of glacier outlines is a prerequisite to observe retreat/surge rates, which is used to precisely measure glacier volume and mass changes. It may be used to monitor natural hazards like GLOF events etc. Though, debris and clouds complicate the automatic periodic delineation of glacier outlines using optical sensors. It leads to immense work load by manually editing the glacier outlines. Interferometrically derived SAR coherence can be used as an alternative to overcome these limitations. It is a quantitative measure of cross-correlation of the observed surface. The coherence value of 1 signifies the stability, while 0 shows complete temporal decorrelation of the surface.

The fully-automatic processing chain includes thresholding of InSAR coherence (<0.2) and slope, morphological operations, raster to vector conversion and comparison with available secondary data. We tested the method to delineate the glaciers boundaries in Karakoram as it is known for many stable and surge type glaciers which have debris covered tongues.

In this study, InSAR coherence is calculated from ALOS PALSAR HH Fine (L-band) and TerraSAR-X Stripmap mode (X-band) in different time periods. The temporal baseline for ALOS PALSAR is 46 days and for TerraSAR-X, it varies from 11 days to 3 months depending on the data availability and stable interferometric phase. The topographic phase is simulated from reference DEM like SRTM and ASTER GDEM and further subtracted to get the differential interferogram. The thresholding of InSAR coherence and slope as well as morphological operations enable us to extract the glacier outlines in vector format.

We compare the results with existing glacier inventories from GLIMS database and manually edited outlines from Rankl et al. (2014). We observe that this tool is very sensitive to glacier change and is able to outline debris-covered tongues in a very precise way. The TerraSAR-X (multi-looked) with 4m of spatial resolution often shows better results in the ablation area in comparison to ALOS-PALSAR of 10m resolution but is more susceptible to interferences. The establishment of a fully-automatic processing chain together with a high-resolution L-band sensor like Tandem-L could be a step to a cost-extensive real-time monitoring of glaciers.

Reference

Rankl M, Kienholz C and Braun M (2014). Glacier changes in the Karakoram region mapped by multimission satellite imagery. The Cryosphere 8(3), 977–989 (doi:10.5194/tc-8-977-2014)

The problem of global warming and its effects (such as a reduction of the multiyear ice area) is very widely discussed by scientists. In the last decades the ice area reduced to the record low values. Many scientists consider the “greenhouse effect” as the main cause of the ice reducing. Despite the fact that the surface ocean temperature increased almost linearly, the ice cover in the Arctic Basin experienced significant fluctuations in the last 100 years. Apparently the fluctuations of ice cover are influenced by atmospheric circulation, which determines the dynamics of air masses between the northern and southern latitudes.

The estimation of climatic changes in the Arctic ice cover is based now only on changes of the ice area and ice thickness. This approach is not comprehensive enough, because the condition of ice formation and the temporal structure of ice cover variability are not homogeneous in different parts of the ocean. It is necessary to consider also the geographical location of “the hat” of sea ice in the Arctic, the configuration of the ice area, the relationship between areas of different aged ice, the status and configuration of the fields and areas of ridging.

One of the main features of the Arctic ice cover large-scale structure is the ice massifs, the existence and variability of its shapes and sizes are determined by the complex of geographical and meteorological factors.

Wind mode and ocean currents largely impact on the distribution of the sea ice, the formation of the zones of ridges, cracks and clearings, where the young ice forms. The regularities of the young ice distribution may contain an information about surface currents in this area. Therefore, not only thermal, but also the dynamic processes in the ocean are reflected in the spatial structure of the ice cover. The ice movements (as asonal factor) change the distribution of the ice thickness and its entire meso and macrostructure.

Water exchange and freshwater runoff can significantly affect the ice conditions. Desalination of the surface layer changes the physics of frost and growth of the ice and the weakening of the ice removal through the Fram Strait creates more favorable conditions for the multi-year ice formation.

 

The main large-scale circulation elements in the Arctic are Transpolar drift and Anticyclonic circulation (Beaufort gyre). There are time cycles in the dynamic of these elements, but with  a predominance of different frequencies in the Eurasian and Amerasian parts of the Arctic Basin. The variability of atmospheric pressure and wind fields over the ocean leads to the formation of different types of surface water and ice circulation - with developed or weak Anticyclonic circulation or Transpolar drift. Asynchronous variability in the large-scale structures intensity may lead to the anomalies in the drift fields, where the conditions of multi-year ice formation may vary.

The variability of the ice drift under the influence of the global atmospheric circulation plays an important role in the formation of the conditions, that defines the changings of multi-year ice area and the average thickness of the ice in the Arctic Ocean.

To test this concept the authors analyzed the long-term changes in the ice drift fields, basing of the satellite datas and the ice drift fields calculation, and attempted the evaluation of  the relationships between ice drift fields changings and climate changings of the ice cover in the last decades.

CryoSat-2 carries a sophisticated radar altimeter, SIRAL, proposing three different operating modes: SAR and Interferometric modes and a Low Resolution Mode similar to other conventional altimetry missions (Jason-2 for example). CryoSat-2 is designed to acquire continuously while revolving around our planet, switching automatically to its three measurement modes according to a geographic mode mask. Nominally, SAR is operated over sea-ice areas and over some ocean basins and coastal zones. SAR Interferometric (SARIn) mode is used over steeply sloping ice-sheet margins, over some geostrophic ocean currents and over small ice caps and areas of mountain glaciers. It is also used over some major hydrological river basins. Low Resolution Mode (LRM) is operated over areas of the continental ice sheets, over oceans and over land not covered by other modes.

If the performance of SAR altimeter has been thoroughly analyzed over water surfaces, the capabilities of this acquisition mode over land ice have still never been addressed. Beginning of 2014, CNES asked ESA to set sporadic SAR mode acquisitions over Antarctica (Dome C, Vostok lake and spirit zone) to be in the position to process and to analyze the SAR mode measurements respect to conventional altimetry.

The chief objective of this paper is to give an overview of SAR mode performances over Antarctica. We will analyze the Doppler echoes shapes, the elevation accuracy respect to different DEMs, the sensitivity to terrain slopes, the snow penetration impacts. All those characteristics will be compared to LRM ones and a cross comparison with Altika Ka-band altimetry satellite mission will also be presented.

Ice surface velocity is an essential parameter to monitor environmental changes for ice sheets and glaciers. Especially for outlet glaciers, it helps to characterize the dynamics of the ice sheet and aids to quantify the ice export. Furthermore, velocity time series obtained by combining data from various, past and present spaceborne sensors assist in observing the behavior of important glaciers and the ice sheets on a long term scale. The ice surface velocity can be used as input for glacier mass balance calculations.

One approach to derive velocity fields is to track features from time series of repeat pass SAR imagery. We apply the normalized cross-correlation technique on backscattering amplitude images to retrieve surface displacements between two 11-day repeat pass datatakes. We use TerraSAR-X (TSX) single look slant range complex (SSC) products. As the processing is performed in radar geometry, a digital elevation model (DEM) is utilized to geocode the displacement vectors on the earth's surface.

Results have shown, that producing accurately mapped ice velocities requires a high resolution DEM. Therefore, wherever possible, we process a DEM, which is close in time to the TSX data pairs. The DEMs are generated from TanDEM-X (TDM) bistatic acquisitions. The TDM DEMs offer an unprecedented spatial resolution for the investigated area in Northeast Greenland and help to considerately improve the accuracy of ice velocities. This is true especially for polar regions because they could not be covered by SRTM. For areas that cannot be fully covered with valid TDM elevations, the Greenland Ice Mapping Project (GIMP) DEM is used.

In this work we present ice velocities for 8 major outlet glaciers located  along the coastal area in Northeast Greenland (see figure in additional material). These outlet glaciers give insight into the ongoing environmental changes in this part of Greenland. The TerraSAR-X velocities are derived from recent acquisitions during the months February - March 2015.

The TSX velocities will be compared to the recently published Sentinel-1 velocity map for the entire Greenland Ice Sheet available for the winter 2014/2015.

Additionally, we generate time series of absolute elevation for Zacharias Isstrøm from TDM bistatic data and compare the results to the ATM elevations from the Operation IceBridge overflights in this area from 2010 until 2015. This comparison validates the TDM and GIMP DEMs that were utilized in the process of generating the velocity fields.

Mountain glaciers represent an important source of fresh water across the globe. Those reservoirs are seriously threatened by global climate change, and a widespread reduction of glaciers extension has been observed in recent years. Surface processes that promote ice melting are driven by both temperature/precipitation and albedo. The latter is strongly influenced by the presence of impurities within the snow/ice, such as dust, soot etc. The origin of these light-absorbing impurities can be local or distal. During melting season, these impurities can aggregate on the glacier tongue forming characteristic cryoconites, which strongly decrease ice albedo promoting further melting.

In this contribution, we present results from the analysis of hyper- and multi-spectral satellite data (i.e. EO1 – Hyperion and Landsat 8 – Operational Land Imager) and ground hyperspectral measurements (collected during summer 2015 with a field spectrometer on the glacier ablation zone) to characterize ice and snow surface albedo and impurities at the Morteratsch glacier (Swiss Alps). We also compared spectral reflectance with those simulated with the SNow, ICe, and Aerosol Radiative (SNICAR) model in order to estimate the structure and composition of ice and snow.

Observed spectra showed interesting features in spatial albedo distribution at Morteratsch glacier. In particular, the ablation area showed very low albedo values (~ 0.2), and this is probably due to multiple processes such as accumulation of light-absorbing impurities, collapsing of lateral moraine and debris covering. In addition, the presence of surface cryoconites strongly lowers ice albedo and this was clearly captured from ground spectral measurements. Cryoconites albedo was found to be smaller than 0.1 and it creates a feedback process where melt ponds and surface run off further increase the absorption of incident radiation and accelerate ice melting.

Finally, samples of snow, ice and cryoconites were collected on the glacier and analyzed in order to define its geochemical fingerprint. According to these analysis it could be possible to identify the sources of such materials and hence the potential role played by anthropogenic activities in their formation and in the Alpine glaciers darkening.

Current satellite missions, such as ESA Sentinel-2 Multi-Spectral Imager (MSI) presents a higher spatial resolution with respect to Hyperion and OLI imagery, and it is expected to improve the analysis of mountain glaciers.

As part of the ESA (European Space Agency) Data Service Initiative managed by ESA’s Ground Segment Operations Division (EOP-GT), the X-PReSS consortium led by Serco is performing the CryoSat Baseline-C Reprocessing campaign, covering the period between 15-July 2010 and 31 March 2015. The end to end activities from the collection of data, consolidation of new data delivered, and the integration within a completely new reprocessing chain, has been managed for the Star Tracker, Calibration, Level 1 and Level 2 processors. This paper outlines the implementation of this new reprocessing chain, passing from the consolidation of input Level 0 products, through the integration process, generation of the Star Tracker products, generation of the Calibration files, including their usage to generate the new calibration models, and the reprocessing of the entire period of Level 1, Level 2, Level 2 Level 2I (In-Depth) and Level 2 GDR (Geophysical Data Record) products. The final step is also presented; systematic Quality Control of the generated products, which is performed prior to repatriating the data back to ESA in order to make them available to the user community.

The Advanced Microwave Scanning Radiometer 2 (AMSR2) instrument was launched on May 18, 2012 onboard the Global Change Observation Mission 1st - Water "SHIZUKU" (GCOM-W1) satellite. It follows a long line of microwave instruments including the Scanning Multi-channel Microwave Radiometer (SMMR), the Special Sensor Microwave Imager (SSM/I), and the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E). A suite of AMSR2 snow products has been developed for operational use by NOAA, including snow cover detection, snow depth, and snow water equivalent (SWE).

 

Quantitative assessment of the algorithms was performed for a 10-year period with AMSR-E and a 2-year period with AMSR2 data using the NOAA Interactive Multi-Sensor Snow and Ice Mapping System (IMS) snow cover and in-situ snow depth (SD) data as references. AMSR-E snow covered area (SCA) showed a monthly overall accuracy rate of about 80% except in May. Accuracy improves significantly to over 90% when wet snow cases are excluded, and accuracy differences between ascending and descending portions of orbits also decrease. Microwave-derived SCA over dry snow areas can therefore be obtained with accuracy close to optically-derived SCA. Evaluation of results for AMSR-E SD showed a low overall bias of 1 cm and a root mean square error (rmse) of 20 cm. Results for AMSR2-based SCA and SD are similar to those from AMSR-E. Biases and root mean square errors show dependencies on elevation, forest fraction, magnitude of snow depth and snow cover class. AMSR2 SWE has been compared with SNODAS SWE for one month of Jan. 2015. AMSR2 SWE has a bias of 0.02 mm and a rmse of 29.1 mm during this period. Algorithm improvements for snow cover detection and snow depth are being developed and tested. Potential improvements to the snow algorithms will be discussed.

The polar regions are extremely important in terms of their global impacts on weather and climate and functioning of the Earth system. In addition to this, human presence and related activities are increasing in these regions. Focus of this paper is to show the strategy of the COSMO-SkyMed (Constellation of Small satellites for Mediterranean basin Observation) mission in terms of potentiality and opportunity offered to the user community for climate change study, cryosphere environment monitoring and developing of operative services in polar areas.

COSMO-SkyMed is an Earth Observation (EO) Dual-Use (Civilian and Defence) Space System based on a co-planar constellation of 4 mid-sized SAR (Synthetic Aperture Radar)  satellites, fully operational from 2011. It has been designed to face international partnerships and integration of the system itself into a multi-mission framework of cooperating multi-sensor systems. In the polar regions, due to its sun-synchronous orbit, the COSMO-SkyMed constellation can offer opportunities in terms of revisit time and coverage of large areas, like in the case of the north-east and north-west pass covered in only 24 hours.

Sea ice monitoring is required by a wide spectrum of users (Authorities and privately held companies) operating at high latitudes, including for navigation (rivers, lakes and sea) and offshore operations. EO Satellites and in particular SAR instruments represent a reliable tool for ice monitoring because they are able to provide a synoptic view that complements the accurate but low coverage reports from ships and airborne sources. SAR data are able to provide information on the ice coverage, the size and shape of ice floes. In particular, SAR images provide the crucial advantage of a weather-independent, day–night imaging system, in the glacier environments where persistent clouds continue to hamper data acquisitions by visible imagers and where the polar night imposes a prolonged period of darkness. In addition, timely and variable information on sea ice conditions are essential for all operations in ice-covered areas, in fact the safety and efficiency of sea transportation, offshore operations, fisheries and other activities in regions covered by sea ice require high-resolution and ice forecasts. On arctic lands SAR can provide monitoring of land instability due to permafrost thawing and melting. One of the services required is the Ice Charting, were high resolution SAR data could make a significant contribute. These are only few examples of the potentialities of the COSMO-SkyMed SAR data in the framework of the exploitation of EO data on polar regions. In order to support the recent needs for ice sheet applications, a specific acquisition plan has been reorganized and expanded over Antarctica and Greenland in the framework of COSMO-SkyMed Background Mission (BCK Mission). This low priority acquisition plan, started in 2011, represents a great opportunity for user community: it is exploited toperform systematic acquisition plans over specific areas of interest, in order to guarantee measuring continuity and the availability of reference datasets for current and future activities, such as mapping projects, emergencymapping, change detection applications, etc.  The areas of interest have been selected collecting the expression of interest related to specific sites and topics coming from the scientific, institutional and commercial community. 

Furthermore, ASI periodically issues announcements of opportunities devoted to the scientific exploitationof COSMO-SkyMed data (free of charge) for basic and applied Research & Development (R&D). In this context, two open calls were published on the ASI’s website on February 2015: one of their objectives is to facilitate innovative ideas for their synergistic utilization with the ESA (European Space Agency) and international EO Missions.

This paper aims to show the usefulness of COSMO-SkyMed data exploitation on polar regions. 

At the border between Uganda and the D.R. of Congo, the Rwenzoris form a remote and high-altitude mountain range stretching through the East African Rift System. With heights of 4-5 km, they include Africa's third highest peak (Mt. Stanley, 5109 m) as well as some of the last African glaciers. The combined area of the Rwenzori glaciers declined by more than 75% during the 20th century, and halved between 1987 and 2006. This extreme mass loss may have strong implications for the local hydrology, ecosystems and communities, and recent estimates suggest that the glaciers will disappear in the next decade(s). This trend correlates well with similarly dramatic glacier retreats on Mt Kilimandjaro (Tanzania) and Mt Kenya (Kenya) during the same period, and is attributed to increased air temperature or reduced humidity/cloud cover. Despite recent work on the evolution of glacier extent in the last decades, the measured glacier retreat, as well as the interpretation of the driving climatic factors responsible since the 1980’s, remain controversial and are limited to available data.

Our new project, RIDEC (Rwenzori Ice Dynamics and Environmental Changes), is aimed at gaining a better understanding of the dynamics of this recession. Here we report on the current state of the two largest Rwenzori glaciers, Stanley and Speke glaciers, based on results obtained from a panel of remote sensing, geophysical and geochemical methods. Our results shed light on the sensitivity of Rwenzori glaciers to the changing climate and will further be compiled to provide a first estimate of modern, past and future ice budgets in the area of interest.

The Sentinel-1 mission of the European Copernicus Programme operates a C-Band SAR sensor with advanced imaging capabilities, providing improved, long-term observation capabilities for important bio- and geo-physical parameters of the Earth system. Snow cover monitoring is a main application for water management and climate research. We developed, implemented and tested a procedure for retrieving maps of snowmelt area. The algorithm applies a similar change detection method as applied for snow mapping with SAR data of ERS and Envisat. We use Level 1 (swath based) single look complex (SLC) SAR products acquired in Interferometric Wide swath (IW) mode in dual (VV, VH) polarization. The IW mode is the nominal operation mode over land surfaces, with 250 km swath width and 5 m x 20 m nominal ground resolution. As first processing step multi-channel speckle filters are applied to the precisely co-registered snow image and to reference images acquired under snow-free conditions. A next step the ratio of backscatter intensity of the SAR image with melting snow versus the reference images is compiled. A data fusion procedure is applied to combine the VV and VH ratio images for obtaining an optimum feature space for snow segmentation. Terrain corrected geocoding and segmentation is performed for the ratio image, using precise orbit data and a DEM. Post-processing employs a land cover map in order to exclude water surfaces and dense forests. We generated a sequence of snowmelt area maps with 100 m spatial resolution for April to June 2015 covering the European Alps, and maps with 50 m resolution for regional studies. For quality assessment we compare the S1 snow maps with snow extent derived from Landsat images and from medium resolution MODIS images. The snow maps of the different sensors show good overall agreement. Some minor differences are evident along the snow boundaries, with a trend for underestimation of snow extent by SAR in particular in areas of broken snow cover due to high backscatter signals of snow-free patches. The fusion of co- and cross-polarization channels in the snow retrieval procedure yields improvements over single channel data. The retrieval algorithm enables the regular operational production of high resolution snow maps over extended areas. Sentinel-1 IW data provide complete repeat coverage in mid latitudes within about 6 days with a single satellite, and within 3 days with the two satellite constellation. This offers excellent capabilities in support of operational snow melt monitoring and forecasting. 

A new sea ice concentration retrieval algorithm that uses weather corrected polarisation difference in brightness temperature at 89 GHz measured by AMSR-E is developed. Two approaches, the new version of the Artist Sea Ice algorithm with weather correction (ASI2) and the Linear 90 algorithm (Lin90) are studied and compared. Effects of wind, total water vapour, cloud liquid water and surface temperature on the measured brightness temperatures (TBs) are evaluated through a radiative transfer model (Wentz and Meissner 2000). TBs of open ocean and multi-year ice are more sensitive to the atmospheric water due to their low emissivities, whereas that of first year ice are more influenced by the surface temperature. These weather effects are corrected by simulating changes in TB caused by the atmospheric water absorption/emission and wind roughened ocean surface scattering using climate reanalysis data fields as atmosphere profiles. New tie points are computed for the new algorithm, and the corrected TBs are then input to the ASI2 and Lin90 algorithms to retrieve ice concentration.

The corrected TBs and ice concentrations are compared with a validation data set consisting of TBs measured by AMSR-E over open ocean and consolidated ice. This dataset was developed in the context of ESA Climate Change Initiative Sea Ice project . Lower standard deviation of TBs and retrieved ice concentration is interpreted as a positive correction effect. The ASI2 and Lin90 algorithms yield similar results. Two reanalysis products are tested as input atmospheric profile, ERA-Interim by European Centre for Medium-Range Weather Forecasts (ECMWF) and Arctic System Reanalysis (ASR) by Bromwich et al (2012). The spatial resolution of ERA-Interim is about 80 km, while that of ASR is 30 km. The latter utilizes a high resolution version of the Polar Weather Forecast Model (PWFM) that is optimized for the Arctic, and yields lower bias in ice concentration. Over open water, the correction decreases the standard deviation of ice concentration from over 30% to about 20%, and lowers its bias to within 5% in all months. Over consolidated ice, the correction effect strongly depends on the chosen ice emissivity. The standard deviation of first-year ice emissivity can reach 0.12 at 85V and 0.11 at 85H during melt season (Willmes et al 2014), but such variability is not represented by the monthly ice emissivities used in this study. The corrected ice concentration show little difference in winter, and has much higher bias (up to -26.1%) and standard deviation (up to 24.7%) in summer. Seasonal tie points are being researched to better represent the variability of ice emissivity.

The ASI2 algorithm is applied to AMSR-E Level 2A re-sampled swath data of 2008. At 89 GHz, the Level 2A data has a spatial resolution of 6 km. New weather filter based on corrected TBs at 19V, 19H and 37V are applied to screen out the residual cloud impacts over open water. Resulting ice concentrations are compared to those of the NASA Team 2 and Bootstrap algorithms, and to MODIS images near the ice edge. More comparison will be done with the OSI-SAF and SICCI ice retrievals. The ice area computed by ASI2 agrees well with NASA Team 2 and Bootstrap in winter, but is underestimated in summer, in accord with the poor performance of weather correction during the melt season. Overlaying ice concentration contours of ASI and ASI2 with MODIS images shows that ASI2 resolves a more realistic ice gradient across the ice edge. In conclusion, the new algorithm can effectively reduce the weather effects at regions of low ice concentration, and is able to retrieve ice concentration lower than 10%. Its application on closed ice pack depends on a better understanding of the ice emissivity variability, and the research is on going.

Despite the importance of precipitation for the water cycle and for the surface mass balance of the Antarctic continent, its quantification remains uncertain and little is known about its microphysics.

Measuring precipitation in Antarctica is difficult because the inhospitable environmental conditions lead to significant logistical challenges and technical limitations. On the coastal areas, affected by strong katabatic winds, an additional challenge is to discriminate at the ground level between precipitation and horizontal transport due to wind-blown snow.

From the end of October 2015 until the end of February 2016, several precipitation measurement devices will be operated around the base of Dumont D'Urville, on the Antarctic coast in the framework of the project APRES3 (http://apres3.osug.fr/).

A scanning X-band (9.41 GHz) polarimetric Doppler radar will be used to explore the horizontal and vertical structure of precipitation, to characterize precipitation microphysics (thanks to the dual-polarization and Doppler capabilities), and possibly to identify layers of wind blown snow. A second system, a Micro Rain Radar (MRR) profiler operating at 24 GHz, will complement the information about the vertical structure of precipitation with continuous vertical profiles.

Two in-situ measurement devices will be deployed to sample the characteristics of precipitation at the ground level. A wind-shielded weighting gauge will provide snowfall quantities while a Multi Angle Snowflake Camera (MASC) will be used to collect high resolution images of snowflakes and wind-blown snow that will help to understand snowfall microphysics and to constrain the radar-based retrievals.

The measurements collected during the field deployment period and the early results will be presented.

Gravity models are powerful tools for mapping tectonic structures, especially in the deep ocean basins where the topography remains unmapped by ships or is buried by thick sediment. We combine new radar altimeter measurements from CryoSat-2, Jason-1, and AltiKa with existing data to construct a global marine gravity model with significantly better accuracy and spatial resolution than previous models.  All the altimeter waveform data are double-retracked to remove the correlation between sea surface height (SSH) and significant wave height (SWH).  For the standard waveforms, the double retracking improves the range precision by a factor of 1.5.  The 20 Hz equivalent noise level of the waveforms is calculated for each altimeter at 2 m SWH.  Ranked from best to worst are AltiKa (22 mm), CryoSat-2 LRM (43 mm), Jason-1 (46 mm), CryoSat-2 SAR (50 mm), Envisat (52 mm), Geosat (57 mm) and ERS-1 (62 mm).  Altika has the lowest noise because it operates at a higher frequency offering a sharper radar pulse, a higher pulse rate and a smaller antenna beamwidth.  By far the best spatial coverage is achieved by CryoSat-2, which provides more than 60 months of non-repeat coverage.

All of these new data are providing a much sharper image of the tectonics of the deep ocean basins and continental margins.  During this talk we will give a tour of the new tectonic structures revealed by CryoSat-2, Jason-1 and AltiKa and speculate on the tectonic views of the ocean basins in 2018 and beyond.

Permafrost landscapes are very sensitive to climate change. An adequate modelling of their water regime is important for an accurate evaluation of fresh water fluxes into the Arctic ocean and emission of CO2/CH4 from thawing soils. Watershed of the river Pur , situated in the north of the Western Siberia, is a complex periglacial region abundant in water bodies that belong to the vast wetlands.

Objective of this study is to compare the performance of two distributed hydrological models - WATFLOOD and ECOMAG - to simulate water outflow and watershed water balance through the validation at in situ (discharge) and space (watershed water storage) observations. The integrated water storage anomalies from GRACE mission were used to estimate the accuracy of the water storage simulations. A prototype product of the future NASA-CNES Surface Water and Ocean Topography mission (SWOT) was elaborated from high resolution optical Landsat 8 sensor and ENVISAT RA-2 altimeter. This prototype is used to evaluate the consistency of the model simulations of the seasonal water storage variability in the surface pool, sensitive to the permafrost thawing.

This research has been done in the framework of the SWOT-TOSCA CNES, IDEX Transversalité 2013 InHERA and ANR "CLASSIQUE" projects, as well as by GDRI "CAR-WET-SIB II" and French-Siberian Centre for Education and Research.

Accurate measurements of ice thickness acquired at high temporal frequency are important for the improvement of operational sea and lake ice forecasting systems. Retrieval of ice thickness is challenging in high latitude regions, at a time when such measurements are increasingly being requested by operational ice centres. The majority of current sea ice forecasting systems assimilate data from passive microwave sensors with coarse spatial resolution (tens of km) and lake ice forecasting has received much less attention. Ice thickness observations are not typically directly assimilated.

This study, currently being conducted as part of the Marine Environmental Observation Prediction and Response Network (MEOPAR) project, aims to improve retrieval algorithms for the estimation of sea ice and lake ice thickness using data from the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA’s Aqua (EOS PM) and Terra (EOS AM) satellites. The accuracy of ice thickness retrievals are investigated based on MODIS Ice Surface Temperature (IST) using a heat balance equation and snow parameterization implemented in a 1-D thermodynamic lake ice model (Canadian Lake Ice Model, CLIMo). Great Slave Lake, a large lake located in the Mackenzie River basin in Canada’s Northwest Territories, is used as a test site due to the availability of in situ snow and ice thickness measurements required for evaluation of satellite retrievals. This allows examination of one of the main sources of uncertainty in retrievals algorithms; the snow depth parameterization. Retrieved ice thicknesses are compared with those obtained from the Advanced Microwave Scanning Radiometer—Earth Observing System (AMSR-E) and in situ measurements from Canadian Ice Database (CID) for the period 2002-2014. The accuracy of ice thickness estimates is improved when using the snow parameterization rather than an empirical relationship between snow depth and ice thickness, with biases of 0.05 m and 0.02 m for ice thickness from MODIS and AMSR-E, respectively.

Further evaluation of the retrieval algorithms are also planned over the Great Lakes of North America and over the Beaufort Sea, where airborne EM thickness measurements have been acquired through the MEOPAR project, with the intent of improving our ability to forecast changing ice conditions. 

The Sentinel-1 satellite constellation, comprising after full deployment 2 satellites, opens up new opportunities for operational monitoring of the Earth’s surface, including ice sheets and their outlet glaciers. The Interferometric Wide Swath (IWS) mode of Sentinel 1 is the standard operation mode over land surfaces and inland ice. It applies the novel Terrain Observation by Progressive Scans (TOPS) acquisition technology, providing a spatial resolution of about 3 m and 22 m in slant range and azimuth, respectively, with a swath width of 250 km. With these powerful imaging capabilities, in combination with a coordinated acquisition strategy, the Sentinel-1 constellation has the potential to become the main source for regular and comprehensive monitoring ice motion over Antarctica and Greenland.

 

The Sentinel-1 acquisition planning for Greenland includes an annual ice sheet wide campaign with 4 repeat acquisitions for each track. The first campaign was held in January to March 2015. For generation of the ice sheet wide velocity map we applied an iterative offset tracking algorithm using coherent and incoherent image cross-correlation. The full spectrum of flow velocities is mapped by using variable sizes of the matching windows. We present the first ice sheet wide Sentinel-1 velocity map of Greenland, with 250 m pixel spacing. The Sentinel-1 ice velocities agree very well with velocities derived from high resolution TerraSAR-X Stripmap data which are available for several outlet glaciers. In addition to ice sheet wide SAR data acquisitions on campaign basis, continuous monitoring of the margins throughout the year is planned, using 6 tracks that are acquired at every cycle. This allows observing short term variations of the ice flow of the outlet glaciers. We developed and implemented an automated system for generating velocity time series with 12 day sampling along the main flow lines of the glaciers, starting at the calving front. Examples for the short term velocity variations are shown for several outlet glaciers of Greenland.

 The first Sentinel-1 ice sheet wide acquisition campaign for Antarctica (with polar gap) took place from May to October 2015. About 45 tracks, each with 4 consecutive repeat observations, were collected and processed to produce ice velocity maps by applying the same offset tracking procedure as for Greenland. We show the first Antarctic ice sheet wide velocity map (at 200 m pixel spacing) derived from Sentinel-1 data. On several glaciers these products are compared with velocity maps derived from TerraSAR-X and TanDEM-X data. We report also on short term variations of the ice flow in 12 day intervals over Pine Island Glacier, Thwaites Glacier, and several glaciers of the Antarctic Peninsula throughout a period of more than 1 year starting in October 2014.

Pine Island Glacier in West Antarctica is sensitive to changes in its oceanic and atmospheric environment because of its geometrical configuration, and it is known to be in a state of ice dynamical imbalance. Today, it is retreating, which eventually leads to continuous thinning and speedup of the glacier. Fluctuations in the glacier’s surface mass balance affect our ability to observe these dynamics, for example by affecting the penetration depth of the radar altimeter signals within the near-surface snowpack. We have collected an extensive record of ground-based and airborne observations along a 900 km traverse of the Pine Island Glacier as part of the UK NERC iSTAR-D project, to develop an improved understanding of the factors that affect satellite observations. The data set includes ground-penetrating radar, Ku-band phase-sensitive radar, Ku-band ASIRAS airborne radar, snow density measurements based on neutron scattering, and shallow and deep ice cores. We combine these measurements to determine patterns of snow accumulation across the traverse, as well as their variation in time: Reflectors in the uppermost 10s of metres of the glacier are tracked along the traverse in the various radar data sets and – assuming that each reflector stems from the same event – dated according to the ice core chronologies and annual layer counting in the density–depth profiles from neutron scattering. From these data, we then compute the mean accumulation rate along the traverse using each dated layer and their depth. The data allow us to evaluate the impact that fluctuations in surface mass have on the CryoSat-2 signal over the Pine Island Glacier.

CryoSat-2 was launched in late 2010 tasked with monitoring the changes of the Earth’s land and sea ice. It carries a novel radar altimeter allowing the satellite to monitor changes in complex terrain, such as smaller ice caps, glaciers and the marginal areas of the ice sheets.

Here we present on the development and validation of an independent elevation retrieval processing chain (JPL/DTU) and respective elevation changes based on ESA’s L1B data. Overall we find large improvement in both accuracy and precision over Greenland relative to ESA’s L2 product when comparing against both airborne data and crossover analysis.  The seasonal component and spatial sampling of the surface elevation changes where also compared against ICESat data.

The retrieval processing chain presented here does not correct for changes in surface scattering conditions and appears to be insensitive to the 2012 melt event. This in contrast to the elevation changes derived from ESA’s L2 elevations, which where found to be sensitive to the effects of the melt event and can be directly attributed to differences in retracking procedure. This points out the importance of retracking for radar waveforms for altimetric volume change studies.

Using the newly developed JPL/DTU pipeline we provide an updated estimate of the volume change of the Greenland ice sheet for the period of 2010-2015. The derived volume change shows close agreement with GRACE derived estimates for the same period, but differs with previous CryoSat-2 studies.

Compared to the western and southern parts of the Greenland ice sheet, where narrow tidewater and land terminating glaciers dominate, the north-eastern part is characterized by a major ice stream, the North-East Greenland Ice Stream (NEGIS), which drains about 8% of the ice sheet via three outlet glaciers. While the mass loss of the ice sheet was most prominent in the southern and western areas in the past decade, within the past few years also the north-eastern part showed changes. Therefore, a continuous monitoring of surface velocities of the large outlet glaciers Nioghalvfjerdsfjorden (79 NG) and Zachariae Isstrøm (ZIS) is intended. For this purpose we make use of Sentinel-1a data, aiming at near real time velocity measurements. Surface velocities are calculated for every 12-day and 24-day repeat pass by means of offset intensity tracking. Here we make use of Terrain Observation with Progressive Scans SAR (TOPSAR) Single Look Complex (SLC) data and precise orbit information. In order to achieve robust velocity fields and to close data gaps, offset tracking results are stacked on a monthly basis resulting in a continuous time series of velocity measurements since December 2014.

Two decades of satellite observations have documented substantial mass loss from the Greenland Ice Sheet in response to ocean- and atmospheric-driven melting. During this period, ice mass has fluctuated over a range of timescales, from seasonal cycles in surface melt to decadal and longer responses to the surrounding climate. Mass loss is also highly variable in space, as a result of the distinct geometries and boundary conditions of different glaciological catchments. Resolving changes in ice sheet mass at both a high spatial resolution and temporal sampling frequency is necessary for understanding the timescales and drivers of glaciological change, but has remained a challenge for all satellite geodetic techniques. While radar altimetry has the potential to sample elevation change at a high (~ monthly) temporal frequency and with moderate (~ 5 km) spatial resolution, observations of Greenland Ice Sheet elevation change and mass balance have been challenging because of the kilometer scale resolution of pulse limited elevation measurements and the impact of highly variable snowpack conditions upon the backscattered echo. In this study we investigate the ability of CryoSat, which offers higher spatial resolution around the ice sheet margins, to derive estimates of Greenland Ice Sheet elevation change. We use an along-track approach to map ice sheet elevation changes between 2011 and 2015 and evaluate our results using ICEBridge airborne altimetry. Particular attention will be given to the influence of changing snowpack conditions on the received echo, using the RACMO regional climate model to simulate changing melt conditions across the ice sheet. We will investigate the impact of the extensive 2012 summer melt event on the relative contributions of surface and volume backscatter, and the corresponding impact upon the altimeter derived estimates of surface elevation change.

Extreme weather events significantly affect social well-being and economics. The 2008 Southeast China snow and ice storm, lasting from middle January to middle February, affected 21 out of China’s 34 provinces and regions with heavy snows, ice and freezing rains. As China’s worst winter in 5 decades, the storm caused extensive damage and transportation disruption, displaced nearly 1.7 million people, and claimed 129 lives. The massive accumulation of snow and ice due to the month-long precipitation and cold temperature slightly changed the gravity field of the Earth, and was sensitive to the Gravity Recovery and Climate Experiment (GRACE) satellite measurements. Here we employed an improved energy balance approach to invert for sub-monthly mass changes resulting from the snow/ice storm over the study region. The results are compared and validated against nowcast data sets from an ensemble of atmospheric and hydrologic models, and with in situ precipitation data.  We concluded that the reprocessed GRACE solution indeed captured storm-induced abrupt mass transports at sub-monthly temporal resolutions over storm-affected region, implicating that such data set if available at appropriate latency, may be able to constrain numerical weather forecasting models.

Two-pass Interferometric SAR (InSAR) using short temporal baseline (1- and 3-day) ERS-1 and -2 SAR data is applied to the observation of vertical ice movement, interpreted as resulting from the influx and drainage of water through sub-glacial and/or englacial cavities. The Malaspina is the largest piedmont glacier in the world and is characterised by slow or stagnant horizontal ice flow, allowing phase changes due to vertical ice movement to dominate. InSAR analysis has been augmented using fine spatial resolution Digital Surface Models (DSM) from airborne InSAR and interpretations of sub-glacial drainage patterns are presented.

Since the launch of ERS-1 in 1991, polar-orbiting satellite radar altimeters have provided a near continuous record of ice sheet elevation change, yielding estimates of ice sheet mass imbalance at the scale of individual ice sheet basins. One of the principle challenges associated with radar altimetry comes from the relatively large ground footprint of conventional pulse-limited radars, which limits their capacity to make reliable measurements in areas of complex topographic terrain. In recent years, progress has been made towards improving ground resolution, through the implementation of Synthetic Aperture Radar (SAR), or Delay-Doppler, techniques. In 2010, the launch of CryoSat heralded the start of a new era of SAR altimetry, although full SAR coverage of the polar ice sheets will only be achieved with the launch of the first Sentinel-3 satellite in December 2015. Because of the heritage of SAR altimetry provided by CryoSat, current SAR altimeter processing techniques have to some extent been optimized and evaluated for water and sea ice surfaces. This leaves several outstanding issues related to the development and evaluation of SAR altimetry for ice sheets, including improvements to SAR processing algorithms and SAR altimetry waveform retracking procedures. Here we will outline SPICE (Sentinel-3 Performance Improvement for Ice Sheets), a 2 year project which began in September 2015 and is funded by ESA’s SEOM (Scientific Exploitation of Operational Missions) programme. This project aims to contribute to the development and understanding of ice sheet SAR altimetry through the emulation of Sentinel-3 data from dedicated CryoSat SAR acquisitions made at several sites in Antarctica. More specifically, the project aims to (1) evaluate and improve the current Delay-Doppler processing and SAR waveform retracking algorithms, (2) evaluate higher level SAR altimeter data, and (3) investigate radar wave interaction with the snowpack. We will provide a broad overview of the project, together with any preliminary results arising at this early stage.

Three quarters of the Antarctic ice sheet’s outer margin are confined by ice shelves and glaciers floating freely at the sea. These areas are separated from the grounded ice sheets by grounding zones where ice dynamics change abruptly as the ice detaches from its bed. Mapping ice behaviour in this area is important in order to understand present day ice sheet mass balance and for improving models of potential future change, as ice-flow across the grounding line depends on glaciological parameters like ice rheology, basal properties, and grounding line configuration. Ice dynamics near the grounding line are most precisely measured in all three spatial dimensions by SAR interferometry yielding accuracies for displacement in the cm to sub-cm range. Depending on the radar wavelength and the tidal regime, the fringe pattern is dominated by the tidal signal or mixed with horizontal displacement and superimposed topographic contribution. Differential SAR (DInSAR) is therefore a pre-requirement in order to separate the various phase components. Phase decorrelation in polar snow and lack of stable scatterers are the main restrictions for DInSAR applications in grounding zones.

We studied grounding line dynamics by X-, C-, and L-band SAR in various Antarctic areas with different climatic, tidal, and glaciological regimes. There is now a range of data  easily or freely available from various satellites (ERS-1/2, Envisat, Sentinel-1, TerraSAR-X, ALOS-1/2), providing different revisit times, radar frequencies and bandwidths. All sensors are differently susceptible to volume and surface decorrelation, and depending on ice dynamics and the amplitude of the tidal wave they are differently sensitive to glaciological parameters. A comparison of the performance of these sensors allows us to come up with restrictions and guidelines for suitable imaging configurations and temporal baselines, varying with glaciological boundary conditions, snow conditions, and tidal regimes.

For Antarctic ice shelf conditions, we found the decorrelation to be primarily driven by the repeat pass time. E.g., 11-day TerraSAR-X DInSAR can be successfully applied in many conditions, as well as 12-day Sentinel-1 repeat pass interferometry. As expected the 14-day repeat pass data in L-band (ALOS-2) show more often phase coherence, but the repeat pass time is not as favourable in both diurnal and semi-diurnal regimes in order to map the tidal flexure. For many smaller outflow glaciers with significant velocity gradients across glacier the deformation restricts phase coherence, which is relatively independent of the applied wavelength and repeat pass time (if longer than 1 day). This is also the case in high accumulation and windy areas where the repeat pass time needs to be very short (like for the 1-day ERS-1/2 interferograms) because of the short decorrelation times. In this contribution we showcase our experience, and come up with guidelines to using the presently available data, and an outlook for the expected application of the Sentinel-1/2 satellite configuration to measure grounding line ice dynamics.

A significant increase in sea ice area has been observed in the Antarctic ocean over the past few decades, contracting the expectations during a period of climate change. However, the total sea ice volume, and therefore the sea ice mass balance, is not well known, because of the significant limitations in measuring sea ice thickness from space. One of the restrictions in the Southern Ocean is the relatively thick snow cover over a relatively thin layer of sea ice, which exacerbates the conversion of the satellite measured freeboard to total thickness. Improved information on the snow cover is therefore a necessary requirement in the challenge to measure sea ice volume.

In November 2013 we have conducted validation measurements for snow depth and snow density on level fast ice in McMurdo Sound. At the same time, overlapping TerraSAR-X satellite data in stripmap and scansar mode have been acquired in HH and VV polarisation. Based on this data set, we evaluated the sensitivity of various parameters to snow cover. Amongst these parameters are the mean backscattering coefficient and its incidence angle dependence, the polarimetric phase difference, and the phase coherence between HH and VV channels. In addition we analysed the waveforms of coincident CryoSat-2 overflights.

During the time of the field work the snow was still dry. The snow depth was in general very small, and we could use a Landsat-8 satellite image acquired at the same time to classify snow and snow free areas in support of our field data. Although the dry snow was found to be almost transparent in X-band, we found evidence that TerraSAR data can be used to discriminate between snow and snow free areas if the snow cover is higher than 5 cm. Furthermore, slightly higher incidence angle dependence was found in areas with more snow. Despite these findings we were not able to establish a strong relationship to quantify the snow depth. The largest snow depths were around 20 cm in some areas, but neither CryoSat-2 waveforms nor backscattering coefficients showed strong evidence hinting towards significant snow depth sensitivity. However, the ensemble of data allows the robust classification of snow free areas at high spatial resolution, which is complementary information during times when no optical satellite data are available. Overall we interpret our results as another encouraging step toward improved estimations of snow depth on sea ice.

Operational thin sea ice thickness data has been produced at the University of Hamburg, based on the brightness temperatures at 1.4 GHz (L-Band) measured by ESA's Soil Moisture and Ocean Salinity (SMOS) Mission. The iterative sea ice thickness retrieval takes account the variations of bulk ice temperature and ice salinity in the radiation model. However, 100% ice coverage is assumed, which causes significant underestimation of ice thickness if this condition is not met. Furthermore, in many operational applications a higher resolution of a sea ice product would be preferable instead of using the relatively coarse resolution (12.5 km) of the existing SMOS product. In this study, sea ice concentration with a resolution of 3.125 km derived from the Advanced Scanning Microwave Radiometer AMSR-2 on-board of JAXA's satellite GCOM-W1 has been used to correct open water effect in the SMOS sea ice thickness product.

Due to the different resolutions, a first step for the correction is the disaggregation of low resolution SMOS data with the high resolution AMSR-2 data. Whereas the ice concentration and brightness temperature have linear dependence in the idealized case, the dependence between ice thickness and brightness temperature is not linear, which makes it a main challenge to combine AMSR-2 and SMOS data. Additionally, the sub-grid heterogeneity of ice thickness should also be considered.

In our first approach the correction has been made on the sub-grid brightness temperature, using AMSR-2 data to calculate the sea ice brightness temperature. However, in the marginal ice zone (MIZ), where relatively low ice concentration is often expected, AMSR-2 ice concentration tends to underestimate the sea ice fraction, especially in the thin ice-covered regions. The underestimation of ice concentration results in an inconsistency which hampers the correction in the MIZ.

In the second approach we try to correct the ice thickness, which is retrieved under 100% ice coverage assumption, with a parameterized function of ice concentration. The parameters of the correction function have been estimated using validation data and ice thickness retrieval model.

The implementation of the correction approaches shows that in the Central Arctic, both of the methods could overcome the discrepancy which exists in the current retrieval, that is, the significantly low ice thickness in the presence of leads. However, in the MIZ, a gradual transient in the ice thickness distribution can be achieved if we post-correct the retrieved ice thickness with a parameterized function.

A new method, called the fixed full-matrix method (FFM), is used to compute height changes at crossovers of satellite altimeter ground tracks. Using the ENVISAT data in East Antarctica, FFM results in crossovers of altimeter heights that are 1.9 and 79 times more than those from the fixed half method (FHM) and the one-row method (ORM). The mean standard error of height changes is about 14 cm from ORM, which is reduced to 7 cm by FHM and to 3 cm by FFM. With FFM, the accuracies of the first-half and second-half parts of the height-change time series are almost identical. Assisted by the ICESat-derived height changes, we determine the optimal threshold correlation coefficient (TCC) for a best correction of the backscatter effect on ENVISAT height changes. Use of a TCC value of 0.9 yields the best result compared to the cases with the TCC values of 0.5 and 0.7. FFM yields ENVISAT-derived height change rates in East Antarctica mostly fall between -3 and 3 cm/yr, and they are consistent with the ICESat result. The height-change rates along the Chinese expedition route CHINARE show a large spatial variability and will contribute critical information to the mass balance study here.

The Novaya Zemlya located in the Eurasian Arctic Ocean and the North of Russia contains a lot of glaciers on northern part of island. The northern island is the largest island (48,100 km²) in the whole Eurasian Arctic. Nearly 50% of its surface is occupied by the main ice sheet, which is reputed to be the largest mass of land ice (23,800 km²) in Europe. It is important to measure the glacier velocity variation on the Novaya Zemlya because the glaciers play a major role in influencing the dynamics of mass balance of Eurasia Arctic Ocean. L-band SAR offset tracking method is a useful technique to precisely estimate glaciers movements without in-situ measurements. In spite of low coherent areas, the method is allowed to retrieve two-dimensional glacier movements in LOS and azimuth directions, respectively, using intensity cross correlation. Because the knowledge of velocity patterns is important for predicting ice discharge and understanding total mass loss of glaciers in the study area, we have applied the offset tracking technique with ALOS PALSAR, L-band satellite image pairs acquired during the winter (December to January) season from the 2006 to 2009. However, ionospheric streaks frequently occur in polar region by ionospheric path delays on radar signals, particularly with L-band SAR systems, because L-band SAR data is much more sensitive to ionospheric effects along the SAR swath path. To solve the problem, we mitigates the ionospheric streaks on the azimuth offset maps using directional filter along the LOS direction and validates the performance of the ionospheric streaks mitigation technique. The method remarkably reduces the standard deviation of stationary targets in the 46 days azimuth offset maps from 3.01 m to 0.19 m. It means that the method improves the measurement precision from 23.9 m/yr to 1.5 m/yr by about 16 times. Finally, we measure the glacier velocity variations on Novaya Zemlya using transverse profile maps and analyze them by comparing monthly average regional temperature. The results demonstrate that the L-band SAR offset tracking with the ionospheric streaks mitigation technique has a potential to precisely measure variations of glacier velocity.

The physical conditions along the Antarctic Peninsula have undergone considerable changes during the last 50 years. A period of pronounced air temperature raise, increasing ocean temperatures as well as changes in the precipitation pattern have been reported by various authors. Consequently, the glacial systems showed changes including widespread retreat, surface lowering as well as increased flow speeds. During the last decades numerous ice shelves along the Antarctic Peninsula retreated, started to break-up or disintegrated completely. The loss of the buttressing effect caused tributary glaciers to accelerate with increasing ice discharge along the Antarctic Peninsula. Quantification of the mass changes is still subject to considerable errors although numbers derived from the different methods are converging.

We analysed time series of various SAR satellite sensors (ERS-1/2 SAR, ENVISAT ASAR, RADARSAT-1, ALOS PALSAR, TerraSAR-X/TanDEM-X, Sentinel-1) to detect changes in ice flow speed and surface elevation. The aim is to study the reaction of glaciers to the changing climatic conditions, especially the readjustments of tributary glaciers to ice shelf disintegration, as well as to better quantify the ice mass loss and its temporal changes. We applyed intensity feature tracking techniques on time series from different SAR satellites over the last 20 years to infer changes in glacier surface velocities. Variations in ice front position are mapped in conjunction with the velocity data sets. High resolution bi-static TanDEM-X satellite data was used to derive digital elevation models by differential SAR interferometry. In combination with ASTER and SPOT stereo images, changes in surface elevations were determined. Altimeter data from ICESat, CryoSat-2 and NASA operation IceBridge ATM were used for vertical referencing and quality assessment of the digital elevation models. Airborne laser scanning, ground penetrating radar (AWI Polar-5/6, NASA operation ice-bridge) and differential GNSS data from field campaigns support the ice discharge analysis. At the Sjögren-Inlet a total mass loss of -50.9±8.3 Gt and a contribution to sea level rise of 18.7±5.2 Gt were found. The current average surface lowering rate amounts to -2.1 m/a. At Dinsmoor-Bombardier-Edgeworth glacier system the results show an increase in surface velocity from 0.9 m/d in 1996 up to 8.8 m/d in 1999 close to the terminus. Subsequently, surface velocities decreased to 1.5 m/d in 2014. The changes in flow speeds are coinciding with changes in front position. The surface elevation changed by at least -130±15 m between 1995 and 2014 and -40.7±3.9 Gt of ice were discharged. The detailed multi-mission time series analysis will support the imbalance calculation in the research area and the interpretation on how ice shelf disintegration affects the tributary glaciers.

The progressive retreat of glaciers causes important environmental changes in high mountain regions. The appearance of new glacier lakes is one of the most evident effects in recently deglaciated areas: their formation and evolution is strongly dependent on dynamics of glacial masses.

The Italian Alps, as other mountain regions of the world, are deeply experiencing these changes. It is important to improve the knowledge on the phenomenon and to monitor the present-day situation in such a densely populated region where glaciers and lakes are fundamental both for human and natural systems.

At the present-day, having an updated overview of the presence of glacier lakes at the regional scale has become a need for several reasons. In fact, glacier lakes represent important elements for mountain landscape and ecosystem; they are an economic resource (as water reservoir, for tourism and for the production of hydroelectricity); moreover, they can play as risk factors for enabling Glacier Lake Outburst Flood (GLOF).

The present study aims to produce an updated inventory of glacier lakes in the Western Italian Alps (Piemonte and Aosta Valley) and to reconstruct the evolutionary stages of some selected case studies starting from the end of the Little Ice Age (mid-XIX century) until now.

A first large scale overview and a preliminary inventory have been carried out by manual detection and on screen digitizing in a GIS environment using the most recent aerial orthophotos (2012): about 260 lakes and water ponds were identified within the Little Ice Age glaciers extent boundaries.

In order to get fast updates of the inventory (present-day situation) we are applying different semi-automatic methods for discriminating between water and other surface types (e.g. combination of simple bands ratio, NDWI) on multispectral satellite images (Landsat 8, Sentinel-2 and ASTER). The validation will be done comparing results obtained by the semi-automatic classification of satellite imagery taken in 2012 with those obtained by the orthophotos analysis.

The later stage of the research is devoted to detailed mapping of glaciers and related glacier lakes and to their characterization in term of geographical and geomorphological properties.

Multitemporal analysis is in progress on some selected case studies (e.g. Rutor lakes in Aosta Valley and Northern Locce Lake in Piemonte). The purpose is to reconstruct the evolutionary stages of glacier-glacier lake systems (such as glacier maximum extent, appearance of the lake, further enlargement and/or disappearance of the lake), starting from the end of the Little Ice Age until now. For covering the entire period, an integration of different data is required: historical maps published since the mid-XIX century, aerial photographs available since the 1954 (first Italian national flight) and optical satellite images.

Furthermore, the future evolution of these glacier-glacier lake systems will be monitored using high-resolution satellite images from the Sentinel-2. In-situ GPS measurements (proglacial lake boundary) taken during summer 2015 at the Rutor Glacier will be used to validate the results obtained by the analysis of satellite imagery.

Interpretation of whole data complex will offer better understanding of the dynamics between glaciers and related glacier lakes. Results of considerable significance for the assessment of future development of dynamic mountain environment.

Indian Himalayan glaciers are a part of widely spread mountain ranges of Hindu Kush-Karakoram-Himalaya which consist of second largest ice mass outside polar regions. These mountain glaciers not only indicate the changing climate regimes but also affect infrastructural necessities of the large populations living nearby. Therefore, it's utmost requirement to continuously monitor the glacier changes to address such scientific and civic questions.

The glaciers of Indian Himalaya are located in Himachal Pradesh and other national territories. The glaciers in this region majorly receive precipitation from Indian summer monsoon and mid-latitude winter westerlies. Therefore, it's very interesting to monitor glaciers in this region to possibly estimate the extent of these two different climatic regimes and their influence on glacier mass changes. The ground-based measurements of glaciers in Indian Himalaya are limited to a great extent due to tedious terrain, bad weather conditions, army restrictions etc. Therefore, repeat pass multi-mission satellite remote sensing can be used as an alternative tool to observe such changes in different time scales.

In this study, we use repeat pass radar data from German TanDEM-X bistatic mission and interferometrically process this to generate DEM. The along/across track scenes of TanDEM-X in 2012 are mosaicked and differenced from SRTM C band DEM of 2000 to measure the ice thickness change (myr^-1) during 2000-12. The published ice thickness change of Chhota Shigri glacier, the benchmark glacier in the region, during 1988-2010 from field observations is compared with ice thickness change derived from DEM differencing. This enable us to measure the thickness change residual corresponding to varied radar signal (C and X band) penetration into snow, ice and firn. We use this residual to extrapolate for the entire region to precisely measure the ice volume change during observation period. In addition, we measure the short term (2012-13) ice thickness change using TanDEM-X data which doesn't have residual thickness component as same X band radar signal is used.

The hypsometry analysis (25 m elevation bin) of thickness change of ~800 km² of ice covered area during 2000-12 and 2012-13 is shown in the figure attached. It clearly shows significant downwasting at lower elevation but less negative or balanced condition at higher catchments of the glaciers. We use ice thickness change from 2003-2009 ICESat data to potentially validate the measurements from the earlier techniques.

We estimate the volume changes of the entire region with more attention given to 6 glaciers namely Chhota Shigri, Bara Shigri, Parbati, Samudra Tapu, Patsio and Hamtah glaciers located in different basins.

Sea ice thickness is an essential climate variable that is controlling the surface energy balance and other physical processes of polar oceans. Remote Sensing efforts of sea ice thickness with space borne radar and laser altimetry have been intensified in the recent years with new radar altimeter missions, such as CryoSat-2 or SARAL AltiKa. However the main focus of data product generation from radar altimetry is currently sea ice in the northern hemisphere. One key reason for the focus on the northern hemisphere is that better parametrizations for essential parameters such as snow depth and stratigraphy exist than for sea ice in the southern ocean. Antarctic sea ice is characterized by higher rates of precipitation and regular formation of ice layers in the snow as a result of surface flooding. As a result, it is expected that the interaction of the thick snow layer and Ku-Band waves add significantly to the error budget of sea ice thickness retrieval by radar altimetry.
Nevertheless, ice thickness or volume retrievals of Antarctic sea ice are of interest to better understand the inter-annual and decadal variability of Antarctic sea ice extent. In addition, the geographical distribution of sea ice around the Antarctic continent at lower latitudes leads to better coverage by available mission data and to a potentially longer time series of the entire sea ice cover than in the northern hemisphere. We therefore apply a freeboard retrieval algorithm to a time series of CryoSat-2 radar altimeter data over Antarctic sea. The consistency to earlier pulse-limited radar altimetry missions is tested with coincident data from Envisat. The conversion of freeboard to thickness does depend on the interpretation of main scattering horizon and on the choices of auxiliary data products, such as for snow depth. We present the impact of these choices on sea-ice thickness and total ice volume during Antarctic winter in the recent years. Validation data for sea-ice freeboard and ice thickness is used to assess the uncertainty of the derived ice thickness maps and the prospect of a long-term Antarctic sea-ice volume time series will be discussed.

Estimating the contribution of ice sheets to sea level change is a major goal of glaciologists and of high interest for the public. For this purpose we analyse altimeter data of different satellite-borne satellites with a main focus on CryoSat-2 and estimate by this the volume change and as a final product the mass change using a firn densification model. For the assessment of the contribution of ice sheets to sea level change robust, consistent processing, as well as the estimation of uncertainties is important. There are numerous sources for uncertainty, ranging from instrumental errors, different processing approaches towards the interpolation between sparsely distributed data.
This presentation focuses on the present-day ice-volume changes of the Greenland and Antarctic ice sheets. Based on five years (January 2011 to January 2016) of CryoSat-2 data acquisition we derived elevation change maps and finally volume and mass change estimates for both ice sheets. We will present a set of estimates derived from different processing approaches and interpolation methods. Additional we will compare our results to elevation change rates obtained from ICESat data covering the time period from 2003 to 2009. In contrast to our study of 2014 we extended the time series of CryoSat-2 by two years, used the new data release 34 of ICESat and implemented the output of the firn densification models of the Institute for Marine and Atmospheric research Utrecht (IMAU). The new results will be presented and compared.

The Sea Ice Essential Climate Variable (ECV) as defined by GCOS pertains of both sea ice concentration, thickness, and drift. Now in its second phase, the ESA CCI Sea Ice project is conducting the necessary research efforts to address sea ice drift.

Accurate estimates of sea ice drift direction and magnitude are essential for quantification of the role of dynamic sea ice processes contributing to polar sea ice volume and ocean-atmosphere heat exchange. Error-characterised, long-term sea ice drift information are required for assessing the performance of climate simulations and further develops the physical parametrizations. Sea ice drift products are also required to locate regions of convergent and divergent ice motion across spatio-temporal scales from meters / hours to basins / years.

The work in the CCI Sea Ice project includes defining metrics for assessing the accuracy of algorithms, selecting relevant satellite and “ground-truth” data, building the Round-Robin Data Package for testing the algorithms, and finally selection of the most promising algorithm(s) for processing of a new sea ice drift climate dataset. Specific efforts are dedicated to the definition of per-grid-cell uncertainties in the final product. This contribution reviews the motivation for the work, the plans for sea ice drift algorithms intercomparison and selection, and early results from our activity.

Sea Ice Concentration is the main element of the Sea Ice Essential Climate Variable (ECV) as defined by the Global Climate Observing System (GCOS). Sea Ice Extent and Area are key indicators of climate change, and the trends in maximum and minimum sea ice coverage are regularly in focus of the climate research community, the media, and general public. IPCC models are evaluated against the available satellite-based climate data records, and re-analyses assimilate sea ice concentration datasets or use them as boundary conditions. Sea Ice Concentration can consistently be retrieved from the late 1970s to present from the SMMR, SSM/I, and SSMIS instruments, and can benefit from better spatial resolution since the early 2000s with the AMSR-E and AMSR2 sensors. Numerous algorithms and several datasets were developed already, and with such a focus from the EO community, one could think the story is told for long.

In this contribution, we present the recent outcome from R&D efforts conducted jointly by the teams engaged in the ESA CCI Sea Ice and EUMETSAT OSISAF projects. They include corrections for land spill-over contamination in coastal regions, use of atmosphere re-analysis data to correct the satellite signal, self-tuning algorithms, processing of ice concentration from high(er) frequency channels, more reliable uncertainties, and retrieval of snow depth over sea ice. These are all significant improvements upon the methods that were featured in past versions of the datasets released from EUMETSAT OSISAF and ESA CCI Sea Ice. Changes of processing methodologies and system deployment are also outlined. These algorithm and system enhancements will directly feed into the upcoming sea ice concentration datasets that will be released late 2016 by the two projects.

Current study focuses on sea ice type classification  from  high resolution co-polarized interferometric TanDEM-X SAR images acquired over coastal sea in the Gulf of Riga (Baltic Sea). The first objective was to compare and analyze different (1) imaging modes of TanDEM-X mission (bistatic, monostatic), (2) polarimetric coherence combinations (HH-HH, VV-VV, HH-VV) and (3) incidence angle values (23-44 degrees) in the context of sea ice type classification. Second objective was to find/define the most suitable imaging parameters as well as threshold values of polarimetric coherence and backscatter for sea ice type classification.

The analysis was based on regions of interested on SAR images that were selected based on in situ field observations (35 stations) conducted at the time of SAR and optical satellite imagery (MODIS). The in situ data was used for describing backscatter properties and polarimetric coherence of ice types (fast ice, thin smooth ice, pancake ice, water) on SAR image pairs.  

Incidence angle analysis revealed that the best result for ice type discrimination was achieved in case of  SAR image pairs with high incidence angle (>44°) which showed better coherence differentiation between water and fast ice. In case of high incidence angle the water had mean polarimetric coherence values from 0.25 to 0.45 and the corresponding values for fast ice ranged from 0.58 to 0.78. While in case of low incidence angles fast ice and water had mean coherence values in small overlapping range (0.75-0.91).

The analysis of polarimetric coherence values in case of fast ice showed that best discrimination between fast ice and water was achieved in case the coherence was calculated from two different polarizations (VV-HH). The mean polarimetric coherence difference between ice and water was 0.40 in case of VV-HH bands, while in case of other combinations (HH-HH, VV-VV) the difference (separability) was significantly smaller -  0.15. Comparison between monostatic and bistatic imaging modes revealed no significant difference in ice type differentiation from polarimetric coherency data. Both imaging modes followed similar pattern.

Using the in situ data stations backscattering and coherence threshold values for each ice type were calculated based on mean and standard deviation values. An example classified ice map was constructed by implementing the threshold values on SAR image pair (backscatter-HH, polarimetric coherence bands HH-VV, imaging mode- monostatic, incidence angle- 44.9°).

The algorithm for ice type classification was effective at discriminating between ice types, although some areas (23.20%) were unclassified. Observed backscattering and coherence threshold values for thin smooth ice and water were very similar and therefore inseparable with the proposed classification which was based on combining the coherence and backscattering threshold values.

Sea-ice leads promote a very strong exchange of heat and moisture between the relatively warm ocean and the cold winter atmosphere. Consequently, a considerable fraction of new ice is produced in leads, thereby contributing to the seasonal sea-ice mass balance of the whole Arctic. Leads have moreover been recognized as a source of global methane emissions, which makes them a potential driver for greenhouse gases. The recurrence of leads and their spatial distribution are valuable diagnostic parameters for the sea-ice drift, changes in the sea-ice stability and represent an essential habitat for marine mammals and birds. We apply a recently developed method to automatically identify leads from thermal infrared satellite imagery that includes a filter module to mitigate cloud artifacts and the associated potential misclassification of leads. Here, we present the resulting quasi-daily pan-Arctic lead maps for the months of January to April, 2003-2015. The spatial distribution of the associated average lead frequencies reveals a strong inter-annual variability as well as distinct patterns of predominant fracture zones in the Beaufort Sea and along the shelf-breaks in the Siberian sector of the Arctic Ocean.

In this pan-Arctic study, high-resolution MODIS thermal infrared satellite data are used to infer spatial and temporal characteristics of 11 prominent coastal polynya regions as well as to map the monthly distribution fast-ice areas scattered over the Arctic shelf regions. Thin-ice thickness distributions (≤ 20cm) are calculated from MODIS ice-surface temperatures swath-data (MOD/MYD29), combined with ECMWF ERA-Interim atmospheric reanalysis data in an energy balance model for the last 13 winter-seasons (2002/2003 to 2014/2015; November to March). Valuable quantities such as polynya area and total ice production are derived from (quasi-)daily thin-ice thickness composites. Two different cloud-cover correction schemes are applied on the daily polynya area and ice production estimates to account for cloud and data gaps in the MODIS composites. During the investigated period, the average total wintertime accumulated ice production in all 11 polynya regions is estimated with about 2150 km³. The largest contributions originate from the Kara Sea region (~22%), the North Water polynya (~14%) and scattered smaller polynyas in the Canadian Arctic Archipelago, while other well-known polynya regions (Laptev Sea, Chukchi Sea) show smaller contributions with around 6-7%.  Compared to another recently published pan-Arctic polynya study using coarser resolution passive microwave remote sensing data, our estimates are considerably larger although certain differences regarding the observed winter-period and polynya mask areas do exist. The use of high-resolution MODIS data enables the detection of lead-structures in proximity of the polynyas. Ice production in those areas is therefore included in our estimates. Despite the short record of 13 winter-seasons, positive trends in ice production can be detected for some regions of the eastern Arctic and the North Water polynya, while other polynyas in the western Arctic show more interannual variability with slightly negative trends. We assume that distinct atmospheric and oceanic patterns are responsible for these regionally different developments. Although reliable thin-ice thickness validation data is still absent and despite potential ambiguities from still inherent cloud effects in the MODIS data, we think that our study contains the most accurate estimations of circumpolar polynya dynamics and ice production to date. Combined with a high-resolution mapping of monthly fast-ice distributions, this data set is highly valuable for atmosphere- and ocean-modelling applications.

Newly launched satellite sensors, Landsat-8 and Sentinel-2, together promote very high temporal resolution of optical satellite images especially in high-latitude regions. Additionally with SAR-sensors, like Sentinel-1 and Radarsat-2, optical and radar multi-sensory applications will be promoted. While passive optical sensors reveal the reflective properties at the surface, active radar sensors may penetrate the surface revealing properties from a certain volume. The combination of these sensors thus creates new possibilities for glacier mapping, in our case, extracting glacier outlines and surface types.

Optical glacier mapping methods currently rely on a single satellite scene with good mapping conditions from one point in time. The present-day method utilizes a ratio between the red band and the short wave infrared band (red/SWIR). A threshold value is applied to extract snow and ice from the ratio images. Future glacier mapping methods based on optical time-series, builds on this robust method, however without applying threshold values.

Throughout the season, glaciers display a temporal sequence of properties, both in optical reflection and in radar backscatter, as the seasonal snow melts away, and glacier ice appears in the ablation area and firn in the accumulation area. Each pixel on a glacier thus has a specific temporal signature that varies along the glacier and is often different from the temporal signatures of off-glacier terrain. For example, since the radar waves penetrate into the snow, it can be beneficial to combine radar with optical imagery where seasonal snow often is a problem around the glacier perimeter. Active and passive sensor systems reveal different surface properties and thus they can complement each other. The result from both sensor systems are seasonal mean mapping of snow and ice, either using applied statistics on an image stack, or by taking the chronological order into account fitting curves through the temporal data. The power of the optical on-glacier temporal signal, changes between glacier regions and will profit from ancillary data, in this case SAR-data. In this work we merge the two sensor systems and present a preliminary multi-sensor approach for mapping glaciers.

Regular monitoring of glacier parameters world-wide as important climate change indicator is only possible by means of high resolution satellite data. The new Sentinel-2 mission of ESA and Copernicus is equipped with a high-resolution multispectral imager (MSI), provides a new and comprehensive data base for an operational monitoring service on glacier parameters. The first of the two satellites of the Sentinel-2 mission was launched on 23 June 2015. The Sentinel-2 mission offers various advancements versus other high resolution optical satellite missions. Four bands in the visible and near infrared spectral range with 10 m pixel size provide a very valuable data base for mapping glacier outlines, for offset tracking to retrieve ice surface velocity of glaciers, and for discriminating snow and ice areas on glaciers. Additional bands in the VNIR and shortwave infrared spectral ranges with 20 m and 60 m pixel size enable improvements of existing processing chains for generating glacier products from high resolution optical satellite data. As glaciated and mountainous areas are often affected by cloud cover the revisit frequency of 10 days with one Sentinel-2 satellite, and 5 days with two satellites will significantly increase the chances to acquire more often clear-sky images for analysing glacier outlines, snow and ice areas on glaciers and ice surface velocities of glaciers. Thereby, the wide orbital swath width of Sentinel-2 enables the observation of large glaciated areas at the same time. Further, the synergistic usage of Sentinel-1 and Sentinel-2 data can provide new possibilities for analysing glacier parameters.

We tested applications of Sentinel-2 data over glacier areas in the European Alps and in Greenland. We tested the suitability of S2 data for retrieving ice velocity for outlet glaciers near Jacobshaven Isbrae and Disko Island in Greenland during summer with significant surface melt. We used georeferenced level 1C data acquired on 16 August and 8 September 2015. L1C data are rectified image data using a digital elevation model and are generated by the ESA Sentinel-2 ground segment in the commissioning phase of the satellite. We applied an image cross correlation technique on the NIR Band 8 of the S2 Level 1C data to estimate the local displacement of the images. In the stable areas in front of the ice sheet we found an average systematic shift of about 5 meters corresponding to half a pixel. On the outlet glaciers the method provided an almost complete velocity field by tracking features on the melting glacier surface. The optical ice velocity field agrees very well with SAR based ice velocities derived from a S1 IW image pair acquired in the same period, but shows a much better coverage due to signal decorrelation in the SAR images due to surface melting. Further, we tested the capabilities of the S2 Level 1C data for semi-automated mapping of glacier outlines and snow and ice areas on glaciers. We exploited the suitability of the VNIR bands (2, 3, 4, 8) with 10 m pixel size combined with the SWIR bands 11 and 12 with 20 m pixel size to map glacier outlines and snow areas on glaciers from a S2 scene acquired on 13 August 2015 over the Austrian Alps. Two Landsat 8 scenes acquired on 31 July and on 1 September 2015, overlapping the S2 image over the Zillertaler Alps, are used to compare the resulting glacier outlines and to assess the retrieved late summer snow extent on selected glaciers.

Results are presented to demonstrate the capabilities of the Sentinel-2 data for an accurate monitoring of glacier outlines, late summer snow extent on glaciers, and ice motion. Sentinel-2 has great potential to be used as data source for regular observations of glacier parameters world-wide.

CryoSat-2 radar altimeter data has been used to measure Arctic sea ice thickness and volume for the last 5 winters. By far the largest uncertainty in these measurements is the snow loading on the sea ice, which affects the conversion from sea ice freeboard (the height of the ice above the ocean), to sea ice thickness. In past studies we have used a snow climatology to provide the snow loading. However, this climatology uses measurements collected over multi-year ice in the central Arctic between 1954 and 1991, and may not reflect present snow conditions, particularly on the increasing amounts of seasonal ice seen in recent years. The climatology is also not properly constrained below 70°N, where large areas of winter sea ice are located. Here we present new estimates of snow loading on Arctic sea ice from our own dynamic snow load. This combines daily meteorological fields of precipitation and evaporation with daily estimates of sea ice motion to build daily maps of snow loading on sea ice. We present these maps of snow loading, along with a comparison of sea ice thickness results after applying the snow loading climatology and dynamic snow load. The quality of the new results is assessed by comparison with sea ice thickness measurements from fieldwork campaigns. Finally we use 26 years of daily maps to assess the variability of snow loading on sea ice.

River ice is known to affect many of the largest rivers on the world. About 60% of rivers in the Northern Hemisphere experience its significant seasonal effect (Prowse, 2005). The main hydrological consequence of river ice is its influence on river discharge. Ice-induced extreme flow events can result in serious economic issues such as floods and damages in hydro infrastructure. In the last decade, significant progress has been made on river ice research based on Earth observation data, especially those from satellite Synthetic Aperture Radar (SAR) systems. Methodologies and algorithms have been developed to discriminate river ice types using single-polarized SAR data with good results (Puestow et al, 2004; Gauthier et al., 2010). With the launch of fully polarimetric SAR satellite systems further research have been conducted to improve classification accuracy using dual-polarised and quad-polarised data. Although comparison between classification accuracy based on those two data types has been presented in literature (Mermoz et al., 2009), the information how individual classes (representing different river ice types) change is still needed.  

 

The aim of this study is to investigate whether or not reduction of polarisation channels from quad-polarised data to dual-polarised data results in significant loss of information for main ice types detection. The following ice types were selected for this research: smooth ice cover (black ice), ice cover from loosely agglomerated mobile ice (frazil ice) with loosely accumulated underhanging ice dams, consolidated mobile ice (frazil ice and floes).

 

RADARSAT-2 Fine Quad-pol Single Look Complex data were registered for a section of the lower Vistula river in the winter 2013/2014. Dual-polarised datasets were generated by reducing polarisation channels from original fully-polarised data. From quad- and each of combination of dual-polarised data covariance matrix (C) was computed. To reduce speckle effect the C matrices were filtered with Refined Lee filter (Lee et al., 1999). In the first part of experiment unsupervised Wishart classification based on H/alpha decomposition parameters (Lee, Pottier, 2009) was applied. Through this classification method pixels are classified based on their entropy (H) and alpha value as well as statistical relationship with other pixels. Maximum number of output classes is 8, as the H/alpha plane segmentation, however, it may happen that the final classes number is lower if data represent only a part of entropy or alpha range. For this method classification results for quad-polarimetric and dual-polarimetric dataset were significant different. The number of output classes and classes distribution changed. In the second part of the research supervised classification with maximum likelihood algorithm based on covariance matrix elements were applied. In this case several systematic changes in classification results were observed. Preliminary results showed that although number of polarisation channels had been reduced the classification result still presented satisfying accuracy for the selected classes.

 

This study could be conducted thanks to providing RADARSAT-2 data through Science and Operational Application Research (SOAR-2) opportunity.

 

References:

Prowse, T. D. (2005). River‐ice hydrology. Encyclopedia of hydrological sciences.

Gauthier, Y., Tremblay, M., Bernier, M., & Furgal, C. (2010). Adaptation of a radar-based river ice mapping technology to the Nunavik context. Canadian Journal of Remote Sensing, 36(sup1), S168-S185.

Puestow, T. M., Randell, C. J., Rollings, K. W., Khan, A. A., & Picco, R. (2004, September). Near real-time monitoring of river ice in support of flood forecasting in eastern Canada: towards the integration of Earth observation technology in flood hazard mitigation. In Geoscience and Remote Sensing Symposium, 2004. IGARSS'04. Proceedings. 2004 IEEE International (Vol. 4, pp. 2268-2271). IEEE.

Mermoz, S., Allain, S., Bernier, M., Pottier, E., & Gherboudj, I. (2009). Classification of river ice using polarimetric SAR data. Canadian Journal of Remote Sensing, 35(5), 460-473.

Lee, J. S., & Pottier, E. (2009). Polarimetric radar imaging: from basics to applications. CRC press.

Lee, J. S., Grunes, M. R., & De Grandi, G. (1999). Polarimetric SAR speckle filtering and its implication for classification. Geoscience and Remote Sensing, IEEE Transactions on, 37(5), 2363-2373.

The albedo is a fundamental component of the surface energy budget of glaciers, determining the amount of net solar radiation available for melt. However, its temporal and spatial distribution on the glacier surface is often overlooked and in melt models a constant value of albedo is assumed for the whole glacier, leading to unrealistic estimations of ice melt.

Remote sensing approaches rely on observations from the MODIS sensor, whose spatial resolution is however too poor to characterize variations on a local scale. In this study, we retrieve the albedo on the surface of the Forni Glacier, Stelvio National Park, Central Italian Alps, using satellite images from Landsat 5 TM, Landsat 7 ETM+ and Landsat 8 OLI sensors at 30 m resolution, and investigate its spatial distribution and temporal evolution over a period of 10 years. The Forni Glacier is the widest Italian valley glacier and a darkening of its ablation tongue has been observed in recent years, therefore our study also serves the purpose of validating in-situ observations.

The albedo retrieval method chosen in this study was first introduced by Klok et al. (2003), and it includes correction steps for all the processes that influence the relationship between the satellite signal and the albedo: radiometric calibration, atmospheric correction using the 6S radiative transfer code, correction for local topographic effects and for the anisotropy of the reflected radiation over the hemisphere. The original procedure was developed for use with Landsat 7 ETM+; we re-calibrated it for Landsat 5 TM, taking into account sensor decay, and for Landsat 8 OLI, in view of the differences in the spectral bands compared to its predecessors.

An Automatic Weather Station (AWS), installed in 2005 on the eastern part of the ablation tongue was employed as an ideal data set to validate modelled albedo values against field measurements and good correspondence was found between the two. The albedo values for the ablation tongue in general are typical for debris-rich ice. An attempt was also made to generalize the procedure in an automatic way, through GIS software scripting capabilities, to extend the albedo retrieval to all the glaciers in the Central Italian Alps. Such an automatics procedure can allow obtaining more accurate estimations of ice melt at a regional scale. Application of the procedure to imagery of the Sentinel-2 satellite is also under testing, and, in view of the higher spatial resolution, could in the future capture the spatial variability of surface albedo with greater accuracy.

Snow on sea ice affects the radiative balance and knowledge of snow depth it is required when determining sea ice thickness from freeboard measurements based on altimeter observations e.g. from CryoSat-2. Remote sensing of snow depth based on microwave satellite observations was first suggested by Marcus and Cavalieri (1998) for Antarctic sea ice based on SSM/I (Special Sensor Microwave / Imager) data. It is based on the the gradient ratio of the 18 and 37 GHz channels, vertically polarized, i.e., the difference between the two brightness temperatures, normalised by their sum, denoted (GRV18,37). Transferring the procedure to Arctic ice conditions has been attempted, but turned out to be difficult because it is hampered by several complications such as the presence of more distinct ice types in the Arctic, and more sea ice deformation. During the various flights in the context of the Operation Ice Bridge (OIB) campaigns, also snow depth on sea ice has been observed with a snow radar and laser altimeter. Here, we use OIB snow depth data of the years 2009-2011 (about 500 observations) for a systematic investigation of all gradient ratios between daily averages of the AMSR-E (Advanced Microwave Scanning Radiometer for EOS) channels near 6, 10, 18, 23, 37 and 89 GHz, both polarizations for correlation with the snow depths. As initial result we find the strongest correlations with values around -0.7 for the vertically polarized gradient ratios at 6 GHz channel with those at 10,18 and 23 GHz and for GRV(18,10). A similar study is performed for AMSR2 and OIB data 2012-1015 (about 300 observations). Here we find even stronger correlations up to -0.88 at the same channel combinations. These are promising results in view of systematically retrieving snow depth on Arctic sea ice. A multilinear regression and a more detailed physical analysis of the data will be presented, also distinguishing cases of first-year and multi-year ice and discussing potential systematic influences in the OIB data.

The Barents Sea is one of the fastest changing marine regions of the Arctic, and indeed earth; expressed in an above-average increase in air and ocean temperatures, and decrease in sea ice area. The Barents Sea carries a lot of the observed variability in wintertime Arctic sea ice.  In addition to the Arctic climate implications, sea-ice information is increasingly important for commercial reasons as the Barents Sea becomes the focus of activities such as shipping, fisheries and the petroleum industry.

Little is known about the sea ice thickness in the Barents Sea: in summer, when most field campaigns take place, the Barents Sea is often ice free; in winter field campaigns are sparse. Ice thickness from satellites only recently became available, and validation of these new methodologies is still going on. Here we present extensive ice thickness datasets from Barents Sea field campaigns in 1994-1996, 2003, and 2014, with emphasis on the most recent dataset.

Ice thickness data were acquired with a helicopter-borne electromagnetic sounding instrument (EM-bird) in the Barents Sea, in March 2014. In total eight EM-bird helicopter flights were performed from R/V Lance as part of the campaign between 19 and 26 March 2014. Two flights on 20 March under flew CryoSat-2 tracks. 2014 was an unusual year in that large parts of the Barents Sea that commonly are sea ice covered in winter were almost ice free in mid-February. Ice growth started again about three weeks before the campaign, and the ice encountered in the field campaign in March was thin, new ice; the product of in-situ thermodynamic ice growth rather than being ice transported into the Barents from the Arctic Ocean. For 2014 coincident ice thickness estimates from the CryoSat-2 and SMOS satellites will also be analysed.

The EM-bird data from March 2014 is compared to a previous EM-bird campaign from the same region acquired 11 years earlier in March 2003 during an R/V Polarstern cruise; and to ice thickness from draft measured by upward looking sonars (ULS) between 1994-1996. In 2014 there was no ice at all in the region where the flights took place in 2003, which already highlights the strong variability of the Barents Sea ice cover. We set both years in the wider context using ice concentration and extent, and drift data, from satellite remote sensing, showcasing the variability in ice extent and thickness that is possible in the region.

We also turn special focus on the two Cryosat underflights in 2014. Sea ice thickness retrieval from CryoSat-2 is difficult over thin ice, but we find some valid freeboard values from the waveforms in these orbits, and explore the possibilities for ice thickness retrieval.

The Arctic is an extremely challenging region for the use of remote sensing for ocean studies. One is the fact that despite 25 years of altimetry only very limited sea level observations exists in the interior of the Arctic Ocean. However, with Cryosat-2 SAR altimetry the situation is changing and through development of tailored retrackers dealing with presence of sea ice within the radar footprint, we can now develop sea surface height and its variation in most of the Arctic Ocean.  We have processed 5 years of Cryosat-2 data quantified as either Lead or Ocean data within the Cryosat-2 SAR mask in the Arctic Ocean. By carefully reprocessing and reedited conventional altimetry from ERS-1/ERS-2 and Envisat we have now been able to derive a 25 year time series (1991-2016) using far more remote sensing data in the interior of the Arctic Ocean than ever before.

Along with gradiometer observations from the ESA GOCE mission we are now able to derive a mean dynamic topography of the Arctic Ocean with unprecedented accuracy to constrain the ocean circulation.  We present both a new estimation of the mean ocean circulation and new estimates of large scale sea level changes based on satellite data and perform an estimation of the freshwater storage increase over the last decade using temporal gravity changes from the GRACE satellite. 

The main objective of SONARC Project is to develop a sea ice monitoring and forecasting system to support safe operations and navigation in Arctic seas. SONARC exploits Synthetic Aperture Radar (SAR) data from satellites as a major component of this system. From 2015 Sentinel-1 is the main provider of SAR data, which are delivered every day in the near real-time for monitoring of sea ice and other environmental parameters. In the framework of the project the following algorithms will be developed: (1) sea ice classification of different ice types and open water; (2) sea ice drift retrieval with sufficient resolution to map mesoscale ice motion and deformation fields; and (3) iceberg detection by combined use of SAR and high-resolution optical images. Furthermore, SONARC integrates SAR data with data of Automatic Identification System (AIS) from vessels operating in sea ice areas. With AIS positions combined with SAR images it will be possible for ship captains to find sailing routes through open leads or thin ice and avoid areas with ridges and other difficult ice conditions. For planning optimal sailing routes in the Arctic, a navigation model will be used in combination with AIS and satellite-based ice data. Finally, the sea ice and navigation data combined with ice forecasts will be demonstrated to an ice-going vessel during a field trial. The demonstration will also include use of an ice forecasting model developed at NERSC. The results of the field trial will be assessed by users, and recommendation for future implementation of a monitoring and forecasting system will be worked out. Key user groups for the system are shipping companies, oil and gas companies, operational met-ocean services, coastal and ship traffic authorities, risk management and environmental organizations working in the Arctic.

Through workshops and meetings with users, the project results will be demonstrated and disseminated. The dissemination will target several user groups:

(1) Shipping companies operating in the Arctic. In recent years there has been an increasing number of ships sailing through the Northern Sea Route in the summer months. Many of these ships need to avoid sea ice at all, while others can operate in thin ice, depending on ice class.

(2) Offshore industry involved in planning of Arctic drilling and exploration activities in ice-covered areas.

(3) Operational weather and ocean forecasting centres and marine research institutes delivering oceanographic and sea ice data products.

(4) Other stakeholders such as regulatory bodies (IMO, with the Polar Code), intergovernmental bodies (Arctic Council), risk management organisations (DNV GL), coastal authorities, and the Knowledge Center under Svalbard Integrated Observing System.

(5) Education of students and young scientists in sea ice navigation, which will be done at UNIS (Svalbard) in a special course "Arctic Shipping".

After SONARC is completed, the plan is to implement the monitoring and forecasting system as a service to users who need information about the sea ice conditions to support marine operations. A main dissemination activity is to publish the project results in the referee journals. The expected impact will be improved methodology for sea ice monitoring and forecasting services for Arctic shipping and marine operations. Furthermore, it is also expected that the project will provide new knowledge about Arctic sea ice conditions, which is needed to plan safe and cost-effective operations in the Arctic.

During the installation of a calibration line (gravity range 528 mGal) for relative gravimeters in the Zugspitze area from 2004 to 2005, a series of monthly calibration experiments were made. These experiments showed a temporal variability of the absolute gravity values. A large part of the observed variations can be assigned to mass changes related to water, snow and ice.

The identification of these variations led 2013 to a first experiment for the detection of mass changes, effected by hydrological events, with the main focus on the permafrost lens below the summit region of Zugspitze. In two campaigns, gravity measurements were made in a tunnel below the peak (Kammstollen), one in May before and one in July during ice melting, showing a gravity difference of 0.042 mGal. Based on this observation of significant temporal gravity variability, in 2014 a regular measurement with monthly measurements was initialized to detect annual and sub-annual mass changes, providing important information of water/ice saturation of the rock, a quantity which is difficult to measure by other methods.

A digital 3D density model of the Zugspitze area was implemented, based on which the mass effects of water/ice saturated and water/ice free cubes can be computed. By means of iterative numerical forward modelling, a realistic water/ice mass distribution could be derived, which nicely explain the variations in the gravity field measurements. The model shows also very good agreement with complementary data from an electric resistivity tomography, resulting in reliable estimates for the 3D water/ice distribution alongside of the Kammstollen.

In summary, it could be shown that relative gravity field observations are a feasible method for deriving water/ice saturation changes in rocks, which is complementary and has substantial added value compared to other geophysical methods.

One of the goals of phase 1 of the European Space Agency’s (ESA) climate change initiative (CCI) sea ice project (SICCI project) was to investigate the suitability of satellite altimetry for Antarctic sea ice thickness (SIT) retrieval. Different algorithms based on satellite laser altimetry using the Ice, Cloud, and land Elevation Satellite (ICESat) were inter-compared with each other and with similar satellite radar altimetry algorithms using observations of the Environmental Satellite (Envisat) Radar Altimeter 2 (RA2) and CryoSat-2 (CS-2).

Sea ice thickness can be computed from altimeter freeboard observations using Archimedes’ principle. Required input parameters are the freeboard, snow depth and densities of water, sea ice, and snow. The employed radar altimeters allow to obtain the sea ice freeboard, which is the elevation of the ice-snow interface relative to the sea surface as long as the snow cover on the sea ice is dry and as long as the radar return at the ice-snow interface is strong enough. There is evidence from measurements and theoretical studies that this is often not the case due to the complex structure of snow on Antarctic sea ice and the finite pulse length. In contrast laser altimeters, such as aboard ICESat, allow to obtain the total (sea ice + snow) freeboard. Here, snow physical properties play a much smaller role because the main reflection of the laser pulse always occurs at the snow surface. Common to both, laser and radar altimetry, is the problem to find an accurate data set for the snow depth on sea ice for the computation of the SIT from the freeboard.

This paper presents two aspects of SIT computed from freeboard measurements. In the first part different algorithms (see below) to compute SIT from ICESat freeboard data are employed, inter-compared with each other and with a publicly available ICESat SIT data set. The different algorithms are a) using empirical relationships between freeboard and thickness (no snow depth required), b) employing climatological snow depths, c) approximating the snow depth by a seasonally varying ice density, and d) - which is the SICCI algorithm for ICESat - using snow depth retrieved from Advanced Microwave Scanning Radiometer aboard EOS (AMSR-E). Results obtained with these algorithms will be shown and discussed. One of the results is that it seems mandatory to include snow depth information in order to obtain a realistic SIT distribution.

In the second part, SICCI SIT from ICESat is compared with SIT retrieved from RA-2 and CS-2 freeboard observations. Both RA-2 and CS-2 freeboard estimates are found to often represent the total freeboard rather than the sea ice freeboard. Due to this we decided to compute SIT from RA-2 and CS-2 in two different ways. One assumes that these radar altimeters are indeed sensing the sea ice freeboard. The other assumes that the radar freeboard is equivalent to the total freeboard. Like for ICESat SICCI SIT retrieval AMSR-E snow depths are used to calculate SIT from the radar altimeter freeboards. Main results of this inter-comparison to be shown and discussed are i) if the radar freeboard is taken as a measure of the total freeboard then radar altimetry SIT under-estimates ICESat SICCI SIT; ii) if the radar freeboard is assumed to represent sea ice freeboard then radar altimetry SIT over-estimates ICESat SICCI SIT; iii) taking the radar freeboard as total freeboard (sea ice freeboard) agrees better with ICESat SICCI SIT during early winter (late winter / spring); and iv) RA-2 and CS-2 SIT agree with each other within the uncertainties computed but CS-2 has a positive bias of about 0.1 m for April – September 2011.

The area of west Spitsbergen shelf is the principal region for the Atlantic water pass and it is very dynamic area which has been changing a lot in XX-XXI centuries. In this study, we analyze changes, which were happening in the region of West-Spitsbergen (WSC) and Coastal currents (CC) since 1950 to present time. For this analysis, we used water temperature data from Nordic Seas database of Arctic and Antarctic research institute. It was found that the temperature in the study region are exposed to quasi-periodic changes: temperature increase is observed with anaverage period of 11-12 years. With maximums observed in 1962, 1992, 2005-2007 and less pronounced in 1970’s and 1980’s. There have been found positive trends for timeserias of temperature for the both currents (1°C per 50 years for WSC and 2-2,5°C for CC). Whaler’s Bay polynya is situated in the area to the north of Spitsbergen. This is sensible heat polynya, which is formed mainly because of the West Spitsbergen Current that causes upwelling of warm Atlantic water on the northern shelf of archipelago. Significant changes are happening in the ice-cover of this region:since 2006 polynya stays open for the most time of the year. In this study, changes of polynya size was calculated for the different time scales: inter-annualvariability based on daily AMSR-E Aqua high-resolution ice-concentration data andlong-term variability based on Global sea ice concentration reprocessing dataset 1978-2015 (v1.2, 2015), EUMETSAT Ocean and Sea Ice Satellite Application Facility (Norwegian and Danish Meteorological Institutes) and AARI ice-charts. Polynya size is considered as open water area and was calculated as a sum of pixels on image with value of ice-concentration equal to 0 % for high-resolution data and from 0 to 15 % for low-resolution data. Time series of polynya size and AW temperature were considered together, strong correlation between AW temperature and polynya size was found (correlation coefficient up to 0.7). In some periods,mainly in autumn, rapid changes in area of open-water are observed which are not connected to AW temperature. In recent studies I.Onarheim et al. (2014), S.Falk Petersen et al. (2014) wind speed and direction is considered as another important driver in changes of ice-cover. In this study NCEP reanalysis data is used for wind analysis. We use vectoral-algebraic approach to analyze wind conditions in the study region. Vectoral-algebraic approach is the approach for the analysis of vector series where a Euclidean vector is used as a model, and the vector temporal processes are described by small set of scalar invariant characteristics, derived from the dispersion and the spectral tensors. As a result, we analyze together set of received invariants and ice conditions in the region.

Acknowledgements: The work on this study goes with support of the Nansen Scientific Society PhD grants.

 

During phase 1 of the European Space Agency’s (ESA) climate change initiative (CCI) sea ice project (SICCI project) a sea ice concentration (SIC) data product was produced by employing a hybrid SIC retrieval algorithm comprising the Bristol and the Comiso-Bootstrap algorithm in frequency mode. SIC was computed from brightness temperatures (TB) measured at 19.4 GHz [18.7 GHz] and 37.0 GHz [36.5 GHz] by the space-borne microwave radiometer Special Sensor Microwave / Imager (SSM/I) [Advanced Microwave Scanning Radiometer aboard EOS (AMSR-E)] in both polar hemispheres. The product has daily temporal and 25 km x 25 km grid resolution and is available for the period 1992-2008 (SSM/I) and 2002-2011 (AMSR-E) from, e.g., http://icdc.zmaw.de.

Each data file contains a limited (to the range 0% … 100%) and an unlimited (see below) SIC, SIC retrieval uncertainty, SIC smearing uncertainty from the gridding process, and SIC total uncertainty. A flag layer allows to identify where SIC may be less reliable. The unlimited SIC contains the full range of SIC values retrieved. The natural variability of the measured TBs around the typical TBs at 0% and 100% SIC (the so-called tie points) causes SIC to spread around these two SIC values; consequently SIC can be negative or above 100%. In order to fully evaluate SICCI SIC this natural variability needs to be taken into account. In contrast to most other SIC retrieval algorithms the SICCI algorithm does not filter spurious sea ice over open water with a weather filter because by doing so often substantial portions of the sea ice cover along the ice edge are discarded.

This paper focuses on the evaluation of SICCI SIC for SSM/I. SICCI SIC precision was computed as one standard deviation in SIC around the reference SIC 0% and 100%. For 0% reference SIC we selected open water areas from sea ice charts. For 100% reference SIC we identified regions of convergent sea ice conditions in high ice concentration areas by means of sea ice motion analysis from consecutive Synthetic Aperture Radar imagery during winter. Precision estimates for SIC=0% are 2.2% ± 0.3% (Arctic) and 1.9% ± 0.2% (Antarctic). Those for SIC=100% are 3.0% and 3.3% for Arctic and Antarctic, respectively.

SICCI SIC was inter-compared with high-resolution optical satellite imagery obtained by LandSat during late winter / spring in the Eurasian Sector of the Arctic. For this purpose the surface albedo was computed. High-resolution LandSat SIC was computed using albedo thresholds for open water, thin and thick ice, and co-located with the SICCI SIC. LandSat SIC was averaged over each co-located SICCI SIC grid cell. SICCI and LandSat SIC differ on average by -0.5% ± 3.2% when taking into account SIC under-estimation over thin sea ice.

SICCI SIC data for the Antarctic were inter-compared with independent satellite SIC data and contemporary visual ship-based observations following the Antarctic Sea Ice Processes and Climate (ASPeCt) protocol. All SIC data sets were co-located and averaged along the ship track for days with more than 3 ASPeCt SIC observations. By employing statistical measures SICCI SIC ranks superior compared to SIC from Eumetsat OSI-SAF and NASA-Team algorithms.

Analyses of the seasonal and inter-annual variability of the SIC retrieval uncertainty were carried out separately for open water – defined as -15% < SIC < +15% - and sea ice: SIC > 90%. SIC retrieval uncertainties are stable over the entire SSM/I and also the combined SSM/I – AMSR-E periods. SIC uncertainties over sea ice follow a seasonal cycle which reflects the seasonally changing surface sea ice properties. Average retrieval uncertainties agree within 0.8% (0.3%) with the precision over open water (closed sea ice).

The concept of using the change in the wavelength of ocean waves on entering a frazil-pancake icefield, as measured by spectral analysis of  SAR images, as a way of mapping the thickness of the ice, was originated several year ago by a seminal paper [1]  and fully developed in the following years by two of the present authors [2, 3] In our current work in the EU SPICES project (www.nersc.no/nb/project/spices) we plan to use the results of theory and observations so far in order to develop a processing scheme for routinely deriving ice thicknesses in frazil-pancake regions of the Arctic and Antarctic and hence assess their mass and heat balance. To do this we require more ground truth than had henceforth been available, the only direct comparison between satellite and shipborne work dating from 2004 [3].

 In autumn 2015 (Sep 30 - Nov 10) the University of Alaska research ship "Sikuliaq" carried a team funded by the Office of Naval Research “Sea State Project” to study marginal ice zone processes in the Beaufort Sea, particularly wave-ice interaction processes. One such study involved a line of wave buoys laid out along a pre-declared line, which could thus be covered by simultaneous Cosmo-SkyMed images; in particular, the image dated from October 11 covered an area where frazil-pancake ice was the dominant ice type along the entire length of the survey line out to open water. The ground truth facilities deployed included: i) directional wave buoys placed in the water between pancakes; ii) "Swift" directional wave buoys placed outside of the ice edge; and iii) measurements of pancake ice thickness by recovery of cakes, and of frazil ice mass per unit area using a collector (a "frazilometer") giving a sample that is allowed to melt out. We report on the comparison between the wave fields measured by the buoys and the wave spectra derived from the SAR; and on the comparison between the ice thicknesses measured in situ and the thickness inferred from SAR wave number analysis with the application of a viscous theory.

 

REFERENCES

[1] Wadhams & Holt, “Waves in Frazil and Pancake Ice and Their Detection in Seasat Synthetic Aperture Radar Imagery”, J. Geophys. Res, Vol. 96, No. C5, Pages 8835-8852, 1991.

 [2] Wadhams et al., “The use of SAR to measure ocean wave dispersion in frazil-pancake icefields”, J.  Phys. Ocean.,  Vol. 32  Issue: 6, 1721-1746,   DOI:10.1175/1520-0485(2002)03.

[3] Wadhams et al., “SAR imaging of wave dispersion in Antarctic pancake ice and its use in measuring ice thickness”,  Geophys. Res. Lett.,  Vol. 31  Issue: 15,   DOI:10.1029/2004GL020340.

 

 

 

Snowline is an important indicator of mountain snow cover and reflects the changing climatic behavior in response to the recent warming. The climate of the Romanian Carpathians became warmer particularly in winter, spring and summer, exhibiting an increasing frequency of warm extremes and a obvious decrease of freezing days. The observed changes in the timing of snowmelt in response to the occurrence of milder winters, could explain most of the decline of snow cover duration in the areas below 2,000 m. In this study, we use the altitude of transient snowline (TSL) derived from satellite imagery as a proxies for evidencing the recent (2000-2015) climate change signals during the snow melting season (February-June). For this purpose, we use the multi-temporal Moderate-resolution Imaging Spectrometer (MODIS) reflectance products (MYD10 and MOD10 daily and 8-day composite) and a high-resolution Digital Elevation Model (DEM) (30 m). The observed vertical variations of TSL are discussed in relation to the changes in freezing height (FH) across the Romanian Carpathians, derived from MYD11A1, MYD11A2, MOD11A1 and MOD11A2 daily and 8-day composite products, available at a spatial resolution of 1 km. Python batch scripts using Esri ArcPy were developed and applied to download, subset, reproject and mask each MODIS product. The analyses were focused on producing and using daily and 8-day composites time series from both Terra and Aqua MODIS products for a period of about 12 years, starting from 2002 up to present day. The variability of snow cover persistence was investigated at both monthly and seasonal time steps, allowing to identify the spatial and temporal variations of TSL and FH, as well as the changes in the timing of snow areal depletion across the region. The paper is revealing the inter-seasonal changes in TSL and FH and snow cover behavior in response to the observed changes in temperature and precipitation regime in the areas above 800 m of the Romanian Carpathians. TSL is located with up to 600 m above FH during the accumulation season and with more than 600 m above it during the melting season. FH could statistically explain up to 86% of vertical TSL variation.

In this review, we present the shortcomings of Copernicus and other satellite data that we use to produce the ice charts for the German operational ice service. Moreover, we make suggestions on how these data can be used more effectively in the future.

In the Baltic, an ice chart depicts information about ice concentration and ice thickness, as well as ice ridges, rafting and ice pressure. Since many of the ice charts are produced on a daily basis, the data to produce the ice charts are needed not only in near-real time, but also with a good temporal coverage. Moreover, an adequate horizontal resolution in conjunction with large area coverage is needed, as passages and traffic channels can be quite narrow. 

The most important data to determine the ice concentration from satellite data are daylight and clear-sky independent SAR data. Although there are swath widths of 250 – 400 km and horizontal resolutions of about 20 – 100 m, there is no daily coverage for all regions using just one satellite. Overall, an ice analyst is still needed for the ice-chart production, as todays available automatic sea-ice-concentration algorithms are not good enough. Other satellite data, as passive microwave and multispectral data are also used, but due to their shortcomings are only of secondary importance.

Cryosat and SMOS are used to determine the ice thickness, but the data do still have large error bounds, are not available in near-real time, do have a much too coarse resolution (SMOS), or the  spatial and temporal coverage is not good enough (Cryosat). Microwave data are used to distinguish first-year ice and multiyear ice, but this discrimination is very coarse and of no use in the Baltic. In some cases, multispectral data can be used, but the necessary conditions are not encountered very often. Hence, at the moment the main source for the determination of the thickness of the ice are ice observers.

In the future, a combination of satellite data with previous ice charts and numerical models could provide better sea-ice-thickness data. Further parameters like ridging, rafting, and forms of ice are much more difficult to determine—both today and in the near future. A solution could be the usage of very-high-resolution images, but their spatial coverage is inadequate. Moreover, an approach similar to the one discussed for ice thickness is not suitable too, as almost all models do not include ridging, rafting, and forms of ice. Hence, we suggest to use both in-situ data and highly-resolved satellite data to find a parametrization for determining these parameters. Although we expect the accuracy of the results to be too low to import them directly into our ice-chart-production algorithms, their probabilistic-information content will help us to draw better and more detailed ice charts.

Classification of different habitats is a complex process since no two places in nature have identical environmental conditions or composition of plant species. One approach to create a classification system for vegetational patterns is the use of so-called nature types, which take into account the gradual variations along ecological gradients when building a hierarchy of biodiversity. Environmental gradients make use of abiotic factors like altitude, temperature, soil composition, etc.  Also more complex gradients can be formed by combining more than one factor, like i.e. pH and the contents of calcium and nitrogen in the soil. Practical identification and mapping of nature types requires cost effective methods, especially in remote areas. Here, remote sensing techniques can play an important role.

In alpine regions, the amount of snow during the seasonal cycle can vary greatly within a relatively small area of land due to local terrain conditions as well as differences in wind and precipitation. Since the snow cover limits the general growth season, knowledge about local snow conditions is crucial for determination of the nature type and distribution of plant species for a specific region in question.

In this work, two environmental gradients called reduced growing season due to snow-lie and snow cover stability are brought to focus, and a method for determining the date when the snow melts has been developed. A mountain region in southern Norway has been chosen as the study area, and time-series collected over several years of optical images from Landsat 7 and Landsat 8 have been aggregated in combination with Sentinel-1 SAR images for detection of snow.

After clouds and cloud shadows have been filtered from the optical images, the normalized difference snow index is computed by also including information from the thermal bands provided by the Landsat images. With a suitable threshold is applied to the index, snow-covered pixels are identified and the time instant is recorded. The majority of the available Landsat images are suffering from scattered clouds which limit the number of observations. Therefore, SAR-data is used as a supplement for positive detections of wet snow, thus adding more observations to the data set. With multiple observations of the same pixel over time, the probability for snow cover at a given time instant is estimated using a Bayesian generalized linear model. Estimation of the date for melted snow cover is then found by setting a probability threshold, and the uncertainty of the estimate can then be found by setting a confidence interval.

The proposed method is used on approximately 20 available observations to generate a gradual map of dates for melted snow cover. The results so far show that the proposed method is able to capture the spatial structure of the snow cover duration, and may therefore be suitable for large scale operation if more observations in the time series stack are acquired. However, in order to obtain a sufficiently high degree of accuracy, more than the current 20 observations are needed. Future work will include data from Sentinel-2.

The aim of study was to estimate melt from Gangotri glacier for a decade (2005-2014) using both Degree day factor method and remote sensing techniques. The Landsat (5 7 & 8) data was used and scenes were classified into three classes (snow, ice & rock) using ERDAS imagine software. Equilibrium line altitude was estimated and AAR (accumulation area ratio) value was calculated. Glacier melt was estimated using AAR technique. For DDF method, mean temperature, critical temperature and degree day factor was calculated and melt was estimated using formula. Results from both techniques were analysed and snow line variation for one decade was also studied.

The European Space Agency (ESA) has set up the Climate Change Initiative (CCI) program to provide reliable, long-term, satellite-based data products to successfully understand and manage climate change. Data products are provided for a set of key parameters, the Essential Climate Variables (ECV). Because of their import role in the climate system and their impact on present-day and future sea-level change, both the Antarctic Ice Sheet (AIS) and the Greenland Ice Sheet (GIS) have been identified as ECVs. Respondents of a user survey indicated that the ice sheet mass balance is one of the most important data products needed.

Mass balance products can be derived from GRACE (Gravity Recovery and Climate Experiment) satellite gravimetry data, which are directly sensitive to mass changes. Since 2002 GRACE has provided monthly snapshots of the Earth's gravity field with a resolution of about 200-500 km. Based on this data set two different Gravimetry Mass Balance (GMB) products are generated by the AIS_cci and the GIS_cci project: (a) time series of monthly mass changes for the entire ice sheet and for individual drainage basins, and (b) gridded mass changes covering the entire ice sheet. The gridded product depicts spatial patterns of mass changes at a formal resolution of about 50 km, although the effective resolution provided by GRACE is somewhat lower.

Here we present the first release of the ESA CCI GMB products, which are accessible through data portals hosted by the AIS_cci and the GIS_cci project. The algorithms used for the product generation have been selected within an open round robin experiment and are optimized to account for the complex GRACE error structures, to advance the limited spatial resolution and to separated signals super-imposed to mass changes of the cryosphere. Both the resulting basin averaged and the gridded products are assessed regarding their signal content and error characteristics. Moreover, up-to-date mass balance estimates are presented for both ice sheets.

Various studies have used satellite altimeter data to detect and measure the change in height of the surface of the Antarctic ice sheet and related the derived volume change to trends in the mass budget of the ice sheet.  In this presentation, along-track analysis of ICESat and CryoSat-2 data is applied to an investigation of the spatio-temporal variability of snow accumulation rates with a focus on the region between the inland plateau and the coastal margin. 

This ice sheet slope region spans as much as several hundreds of kilometres distance from the coast.  The ambient surface conditions are typically dominated by katabatic winds, and the episodic incursion of storms.  These storms are the primary source of much of the net snow accumulation.  The wind plays a major role in erosion of the surface, production of surface microrelief that can have a vertical magnitude up to metres, and spatial redistribution of snow.  This roughness, which exhibits significant temporal variability, influences the satellite altimeter returns, and can introduce "noise" into the observations.  The reworked surface is progressively buried by further accumulation events and so is incorporated into the deeper layers which influence the penetration of a radar signal into the snow/firn layer.

Along-track analysis the the multi-campaign data from the ICESat mission, shows the the trend in surface elevation in this region, but also exhibits significant noise.  By appropriate filtering of the values derived at each observation, a clear signal is uncovered which shows a bias in heights that is consistent with each campaign.  This elevation bias results from above- or below average increments of snow accumulation occurring between consecutive campaigns.  These campaigns were typically approximately at 6-month intervals.  The biases were spatially coherent over distances of several hundred kilometres north-south, and up to many hundreds of kilometres east-west.

Thus the signal from the satellite altimeters has contributions from processes operating at a range of spatial and temproal scales.  The analysis is extended in both space and time using the CryoSat-2 mission data, abeit with the added complication of the varying depth penetration of tha microwave signal.

Australian glaciologists have been providing information on sea ice conditions about Antarctica to ships supporting Australian Antarctic expeditions over many years.  In early years this activity drew on very limited information sources, primarily maps showing the spatial extent and concentration of sea ice derived from SSM/I instruments.  The level of support changed dramatically with the implementation of ESA's on-line rolling archive of data acquired by ENVISAT and the associated near-real-time processing systems, particularly of the ASAR data.  This data stream was complimented with visible wavelength images from MERIS, and the wide-swath visible and infra-red wavelength images from MODIS on NASA's Terra and Aqua satellites when near-real-time processing became available for those systems.

The regular and frequent delivery of images to the ships greatly assisted in planning routes to stations on the Antarctic coast as well as adjusting those routes to account for changes in the ice conditions.  The extent, movement, and deformation of the sea ice is continuously evolving in response to the influences of ocean currents, the passage of storms and the shape of the coastline and the distribution of obstacles to the movement of sea ice such as grounded icebergs.  So the timeliness of images is paramount to achieving a successful passage to a destination.

The general philosophy behind this activity is to provide as much relevant information as promptly as possible to ship's Masters and officers and expedition personnel to enable them to choose effective and efficient routing through the ice, and to be able to make strategic decisions on the timing, routing and order of destinations to optimise the use of limited time for access to the continent by ships.  In some cases the the mere presence of any ice will preclude an operation, and so identifying the often brief occasions when an area is free of ice, or alternatively when land-fast ice breaks out, is critical to a achieving an outcome.  It is both an enabling and a facilitating activity.

A selection of case-studies illustrates how the system operates, examples of the range of conditions that are encountered, and the outcome of providing the data.  In late 2015, the average period between Sentinel-1A acquisitions was about 6 days.  Increased acquisition frequency and a shorter time to product delivery is anticipated.

The sea ice information activity has expanded to a sea ice service which also provides support to other nation's expedition ships, and now operates under the (Australian) Antarctic Gateway Partnership within the University of Tasmania, in Hobart, Australia.

Both the Antarctic Ice Sheet (AIS) and the Greenland Ice Sheet (GIS) have been identified as key parameters, so called Essential Climate Variables (ECV), in the climate system. Within the framework of the Climate Change Initiative (CCI) of the European Space Agency (ESA), reliable long-term satellite-based data products are generated for selected ECVs. Since ice sheet mass balance is an ECV parameter of highest interest, both the AIS_cci and the GIS_cci project will provide mass balance products based on satellite gravimetry data: (a) time series of monthly mass changes for individual drainage basins, and (b) gridded mass changes covering the entire ice sheet.

Gravimetry Mass Balance (GMB) products are derived from data acquired by the GRACE (Gravity Recovery and Climate Experiment) mission. Although GRACE data have the advantage of being directly sensitive to mass changes, their limited spatial resolution and complex error structures place particular demands on the applied processing strategy. To choose the most suitable algorithm which minimizes the impact of GRACE errors and signal leakage errors on GMB products, an open Round Robin experiment was set up. Participants were asked to provide GMB products according to the specifications of the official products by the ESA CCI using their preferred processing strategy and GRACE release. In addition, the same algorithms should be applied to a series of synthetic datasets, which realistically mimic mass variations in various subsystems of the Earth (e.g. cryoshpere, ocean, continental hydrology). By comparing the derived synthetic mass changes with the a priori known 'synthetic truth', leakage errors can be quantified.

Here we inter-compare the Round Robin results from six individual contributions, relying on different processing strategies, including regional integration approaches, mass inversion strategies, and a forward modeling approach. Time series of basin averaged and gridded products are compared with respect to their specific noise level. The minimization of leakage errors is assessed from the synthetic results. Finally, for selected drainage basins the GMB time series are compared to independent mass balance products based on satellite altimetry and firn densification information from a regional climate model. This inter-comparison has aided the algorithm definition for the operational ECV production.

Based upon high-resolution thermal-infrared Moderate-Resolution Imaging Spectroradiometer (MODIS) satellite imagery in combination with ERA-Interim atmospheric reanalysis data, we derived long-term polynya parameters such as polynya area, thin-ice thickness distribution and ice-production rates from daily cloud-cover corrected thin-ice thickness composites. Our study is based on a thirteen year investigation period (2002–2014) for the austral winter (1 April to 30 September) in the Antarctic Southern Weddell Sea. The focus lies on coastal polynyas which are important hot spots for new-ice formation, bottom-water formation and heat/moisture release into the atmosphere. MODIS has the capability to resolve even very narrow coastal polynyas. Its major disadvantage is the sensor limitation due to cloud cover. We make use of a newly developed and adapted spatial feature reconstruction scheme to account for cloud-covered areas. We find the sea-ice areas in front of Ronne and Brunt Ice Shelf to be the most active with an annual average polynya area of 3018 ± 1298 and 3516 ± 1420 km² as well as an accumulated volume ice production of 31 ± 13 and 31 ± 12 km³, respectively. For the remaining four regions, estimates amount to 421 ± 294 km² and 4 ± 3 km³ (Antarctic Peninsula), 1148 ± 432 km² and 12 ± 5 km³ (Iceberg A23A), 901 ± 703 km² and 10 ± 8 km³ (Filchner Ice Shelf) as well as 499 ± 277 km² and 5 ± 2 km³ (Coats Land). Our findings are discussed in comparison to recent studies based on coupled sea-ice/ocean models and passive-microwave satellite imagery, each investigating different parts of the Southern Weddell Sea.

We will present our methods and results of combining specially processed* MISR (Multi-angle Imaging for SpectroRadiometer) surface spectral albedos and MODIS (MODerate resolution Imaging Spectroradiometer) sea-ice data over the Arctic region in order to produce a multi-temporal 1km resolution albedo product. We have taken advantage of pixel identification (cloud, water, land, sea-ice, snow) applied to MODIS and the multi-angular capability of MISR to create single orbital, daily and weekly MISR derived albedo products.

The albedo shown here is the BHR (Bi-Hemispherical Reflectance), also called “white sky albedo”, for four spectral bands (blue, green, red and near infra red), and two broadband (shortwave and visible). The resulting maps are degraded with a spatial resolution of 1km x 1km in a polar-stereoscopic map re-projection.

In this presentation, we will attempt to evaluate the variation of albedo during the same year time period over the last decade over the Arctic region. For this purpose, we have gathered all available MISR and MODIS products covering the Arctic region, and then applied the following processes: projection, masking (to eliminate clouds), conversion to broadband. The first processing stage creates orbital products, which are then considered as input for daily and weekly products.

A quantitative study is done on the resulting albedo maps to evaluate the variation of the periodicity of albedo during the last decade, and to reveal the most affected regions from climate change. In this work we will show the capability of fusing MISR with MODIS data for the monitoring of sea-ice and snow albedo over the Arctic region.

The snow cover on sea ice receives more and more attention in recent sea ice studies, because its physical properties dominate many processes at the surface and inside the sea ice. In particular the temporal and spatial distribution of snow depth is of crucial importance for the energy and mass budgets of sea ice as well as for the interaction with the atmosphere and for the oceanic fresh-water budget. Snow depth is also a crucial input parameter for sea ice thickness retrieval algorithms that are based on satellite altimetry data as on the past, current and planned ESA missions ERS-1/2, Envisat, CryoSat2 and Sentinel-3. Also, retrievals of thin sea ice thickness from low frequency passive microwave observations, e.g, from SMOS, are influenced by the snow depth. Recent ice volume time series of Arctic sea ice are based on a monthly climatology, which is not able to take annual changes of the snow depth and properties into account. For Antarctic sea ice, no similar climatology is available and large-scale estimates are missing. Hence, the need for more and more consistent observational data sets of snow depth on sea ice is frequently highlighted.

Here, we present time series measurements of snow depths from autonomous platforms on Antarctic and Arctic sea ice. From these we are able to characterize seasonal patterns for different sea ice regions. The results show common large-scale features, but also hint to the great importance of local effects and potential influences of dynamic sea ice processes on snow accumulation. In addition, in the framework of the ESA Climate Change Initiative project on sea ice (SICCI-2), we compare these findings to spatial variability and the seasonal cycle from the latest snow depth product based on AMSR2 satellite observations. Although single point comparisons are subject to scaling issues and small-scale variability, our results reveal significant differences between both observational methods. The comparisons allows new insight into snow cover variability, which may be used for future improvements of snow depth data products on sea ice.

The GlobICE Project, completed in 2012, was a part of ESA's Data User Element (DUE) of the Earth Observation Envelope Programme. The purpose of GlobICE was to derive information on sea ice motion and deformation that will improve understanding of the role of the Arctic in global climate in support of CliC and World Climate Research Programme objectives. The input to the project was Earth Observation data mainly from ESAʼs ENVISAT ASAR archives and supplemented by Radar Altimeter data.  The main products from the GlobICE processor are Eularian and Lagrangian ice motions, ice deformation, including open water fraction, and mass fluxes at key gateways.  

The demise of ENVISAT cut off the input data for the GlobIce processor, but the launch of Sentinel-1a has presented the opportunity of an alternative source.   This paper will demonstrate how the Sentinel-1 SAR data can be used by the GlobIce system.   We will show resulting output products and assess the validation work performed so far.  Finally we will assess the expected impact of adding Sentinel-1b input to the system.

Radar backscatter from a periglacial landscape is influenced by snow cover and its physical properties such as liquid water content, soil surface moisture, and frozen or unfrozen state of the soil surface. We present results from two sites in northern Norway and on Svalbard where we have studied the annual variability in radar backscatter using Sentinel-1 data. The periglacial landscapes are located in the Arctic (northern Norway), with discontinuous permafrost and the High Arctic (Svalbard) with continuous permafrost. Both locations are highly influenced by frequent low-pressure activity due to their maritime locations, resulting in – for their latitude – mild climate and winter rain-on-snow events. These rain events lead to a fluctuation between dry and wet snow, as well as soil surface freezing and thawing.

We show that time-series of Sentinel-1 data can be used to acquire detailed, site-specific information about the onset of soil surface freezing, the onset of the snow cover, snow cover build-up, snow wetting and finally snow melting and soil surface thawing. This information can be valuable for upscaling of permafrost models and in general for the monitoring of arctic areas where in situ monitoring is scarce.

For the two sites we have studied data for the first year of Sentinel-1 data (Oct 2014-October 2015). The data set from the site on Svalbard (Kapp Linné’) consists of 536 Sentinel-1 EW mode scenes, whereas the data set from Norway (Nordnes) consists of 133 scenes. The difference in temporal resolution makes it possible to resolve rapid transitions between soil surface freezing and thawing at Kapp Linne much better than at Nordnes. However, due to the spatial resolution, we also show that snow cover dynamics can be better resolved with IW mode (Nordnes) than with EW mode (Kapp Linne’).

 

Based on a model for the annual variability in backscatter we have developed a map product that can state whether a pixel consists of dry snow, wet snow or frozen/thawed soil. The model hinges on temporal smoothing of the data set, combined with thresholds (based on use of in situ data) for the transitions between different freeze/thaw states. The information provided by the products is very useful for other studies, e.g. ecological and geophysical studies of the landscapes. The products have been compared against in situ measurements of air and soil temperature, soil moisture and snow depth. We have also compared the dataset with Landsat-8 imagery and will snow that snow melting patterns are very similar at this spatial scale.

Human activity in the Arctic is expanding. The last few years have seen a decrease in ice cover in the region, promising easier access to its waters, while the attractiveness of the Arctic is growing on account of its resources (not only oil but also gas, ores and fish), shorter shipping routes and tourism value. Increasing human activity in the Arctic has impacts on safety, security and sustainability, with consequences that can reverberate outside the region. It is therefore important that human activity in the Arctic, that takes place for a large part at sea, is monitored. Knowledge derived from this monitoring is needed by policy makers and other stakeholders in order to make well-informed decisions concerning the region.

This paper explores ways to monitor maritime activity and its changes in the Arctic, based primarily on data from satellite sensors.

AIS (Automatic Identification System, a ship self-reporting system) has become a very powerful tool for monitoring ship traffic. Satellite AIS data collected with the Norwegian AISSat-1 and AISSat-2 already represent more than five years of maritime traffic data over the Arctic. The paper shows annual and seasonal variations in ship traffic, discuss trends in maritime activities, as well as present examples of situational awareness.

Still, the AIS only exposes the reporting fraction of the ship traffic. Satellite SAR can be used to assess the presence of the non-reporting ships. Sentinel-1 is especially attractive as it is intensively monitoring the Arctic. Its main operating modes are useful for ship detection: the EW mode can monitor a wider area, whereas the IW mode can detect smaller ships thanks to its higher resolution. The 20 meter IW resolution limits the detection of ships much smaller than that. But the larger challenge is the presence of sea ice, which gives rise to false alarms in the ship detection process. One avenue that is being investigated to deal with that, is to ingest externally provided sea-ice maps to mask out ice covered areas. The ERDDAP framework has been found to be an efficient tool to automatically access ice-cover predictions from the Real Time Ocean Forecast System numerical model from NOAA. Still, such data are an approximation of the actual situation, and algorithms to discriminate sea ice from open sea and ships in the context of ship detection in Sentinel-1 dual polarisation images are being explored.

On the scale of the entire Arctic, satellite AIS data are used to show how the ship traffic follows the seasonal retreat and advance of the ice. On more local scales, for example monitoring the Yenisei Gulf on the north coast of Siberia, it is seen that the ship traffic (seen with satellite AIS) uses the same opening in the ice (seen with Sentinel-1) for the entire winter season. On the Canadian side of the Arctic, RADARSAT-2 images reveal the shipping activities that serve Arctic communities such as Resolute and Iqaluit. The paper discusses how these images can be used to plan a passage and make a risk assessment.

The continuing collection of Sentinel-1 images, and those from other satellite SARs, over the Arctic has the potential to provide much information on the development of human activities in that region. However, to efficiently and effectively exploit the data, improvements in algorithms for automatic analysis, in particular automatic ship detection in waters with sea ice, need to be implemented. 

Arctic river deltas are highly dynamic environments at the interface of land to ocean. Arctic deltas are underlain by permafrost deposits, which are highly vulnerable to a warming climate. The amount of soil carbon stored in these deposits and thus potentially vulnerable to mobilization is poorly known and based on few data only. In this study we investigate the soil carbon pools of two small-scale Arctic river deltas entering the Beaufort Sea on the Alaska North Slope, the Ikpikpuk and the Fish Creek river deltas. Our approach couples soil carbon information with remotely sensed data to estimate the total carbon stock for these environments. Both river deltas are located within the continuous permafrost zone and are characterized by typical fluvial-deltaic features and processes, such as river channels and islands, floodplains and mudflats, sand dunes, as well as episodic flooding, erosion, and depositions. In addition, permafrost processes are an important factor for thaw, erosion, transport, and accumulation dynamics within these deltas. As a result, features specific to permafrost-dominated deltas are widespread such as thermokarst lakes, drained thaw lake basins and ice wedge polygonal tundra. Under future climate warming projections, Arctic river deltas will be threatened due to thawing permafrost (including melting and settling of ice-rich deposits) and a rising sea level in combination with coastal erosion. To better estimate how much soil carbon may be vulnerable under these projected changes and might be released as greenhouse gases, it is necessary to study the total soil carbon storage in Arctic river deltas. Previous studies estimated a carbon pool (below three meter depth) of 91±52 Pg in all Arctic deltas (Hugelius et al., 2014; Biogeosciences). However, this data is based on very few soil surveys and only coarse upscaling methods. Sampling points from small deltas have not been included so far.

This study presents the first carbon storage estimation for two small deltas which cover an area of about 100 and 80 km², respectively. Nine different soil cores between 54 and 215 cm depth, including both, non-permanently and permanently frozen deposits, were collected in April 2014 and July 2015, and were analyzed in the laboratory for total organic carbon (TOC), total carbon (TC), total nitrogen (TN), stable isotopes (δ¹³C), grain size, and deposit age (¹⁴C). The soil C parameters were upscaled to delta-wide landscape level based on landcover classifications derived from Landsat and Spot images in combination with high-resolution digital terrain models (DTM) from airborne LIDAR and IfSAR datasets. The upscaling of the total carbon storage was based on different approaches including the correlation of near surface soil carbon storage with various remotely sensed landcover indices. These indices, such as the Tasseled Cap or NDVI for the year 2014 were derived from linear trend analyses of Landsat data taking into account the full Landsat archive from 1985-2014. For comparison, a supervised classification (maximum likelihood) with Landsat 8 and Spot 5 images was established based on training areas derived from field information from two field trips, very high resolution aerial and satellite images, and high resolution surface elevation information.  The carbon content was finally upscaled based on mean carbon values for the different land cover classes.

The total organic carbon storage for the two deltas ranges between 1.5 and 2 megatonnes (Mt) of carbon each for the first meter of soil (excluding all water areas), depending on the upscaling method and dataset used. The results compare favorably (comparing the mean carbon storage values per square meter) with what has been previously estimated for other, larger Arctic river deltas.

This study shows that remote sensing is a suitable tool to upscale soil carbon values in remote and hardly accessible Arctic river deltas where only few soil data is available. We are further working on extending our approach to other Arctic permafrost-influenced river deltas, such as the large Lena river delta, Siberia, where we and other colleagues collected a substantial amount of soil carbon and landcover ground truth data.  

The ESA “Greenland_ice_sheet_cci” project  is currently making past and present space measurements of Greenland ice sheet changes available for use by scientists, stakeholders and the general public. The data are part of a large set of ECV’s (Essential Climate Variables) made available by the ESA Climate Initiative, as a contribution to the global Climate Observing System.

The ECV data produced for the Greenlandice sheet include detailed grids of elevation changes and ice flow velocities, as well as line data of grounding lines and calving front locations for major outlet glaciers. The “ice_sheets_cci” goal is to generate a consistent, validated, long-term and timely set of ECV’s, a.o. to improve the impact of satellite data on climate research and coupled ice sheet/climate models.  Special focus is on use of data from ESA missions such as ERS, Envisat and the new Sentinel missions, but in the 2^nd phase of the project, just initiated, mass balance data from the GRACE mission will also be included.

In the presentation the current CCI results are highlighted, including Greenland-wide elevation change results across 23 years of radar altimetrry from ERS, Envisat and CryoSat, ice velocities for the coastal regions and major outlet glaciers, new Greenland-wide ice velocities from the ESA Sentinel-1 mission, and 14 years of mass changes from GRACE.

The ECV data confirm a consistent overall picture of accelerating mass loss of the Greenlandice sheet, a mass loss which has more the doubled in the last decade. Current rates of changes well above 250 GT/year, corresponding to 0.7 mm/year global sea level rise, and originating especially from mass loss associated with major outlet glacier systems. In the future it is proposed to add additional ECV parameters to the products, including a.o. high-resolution mass change products from GRACE/CryoSat/S-3, and ocean sea level rise regional “finger print” data from the associated closure of the overall global sea level budget, in cooperation with Glaciers, Antarctica and Sea Level CCI projects.

Glacier surface albedo is crucial to determine the amount of energy absorbed by the snow/ice surfaces across the year. Especially in summer, when large parts of the glaciers are snow-free, the albedo of the heterogeneous bare-ice surfaces substantially impacts glacier melt rates. A general darkening of snow and ice surfaces due to light-absorbing impurities has been observed in various regions worldwide.
However, a ready-to-use albedo product for glacier surfaces is still lacking, although it is of great relevance to model glacier mass balance accurately. While in-situ measurements from weather stations provide valuable point data, indicating local temporal variations in albedo, airborne- and/or satellite-based sensors offer the possibility to attribute glacier surface albedo to larger areas with reasonable spatial (<30 m) but still limited temporal resolution.

In this study, we apply an approach combining in-situ observations as well as airborne and satellite data to characterize glacier surfaces, in particular the distribution of ice surface materials and their spectral albedo. The APEX (Airborne Prism Experiment) imaging spectrometer was used to acquire reflectance measurements over several glaciers in the western Swiss Alps in August 2015 with high spatial (2 m) and spectral (284 bands, 0.4-2.4 μm) resolution, serving as excellent dataset to evaluate data products derived with the same algorithms from spaceborne Landsat 8 and Sentinel-2 data available with a small offset in time.

We outline a methodology to derive spectral albedo products of glacier surfaces and compare obtained products derived from Landsat 8 and Sentinel-2 data. We use APEX and in-situ measurements of several Swiss glaciers to evaluate the precision and discuss uncertainties of derived products. Furthermore, we discuss the importance of such data products for glaciological applications, in particular for mass balance modelling. Our results contribute to a better understanding of bare-ice albedo of glacier surfaces. Further, they allow judging possibilities for upcoming data products in glaciology based on Landsat 8 and Sentinel-2 data. We conclude on the relevance of these new missions to extend the yet limited temporal resolution to map dynamic glaciological processes such as the darkening phenomenon of glacier surfaces, and to advance and facilitate modelling approaches to study future glacier evolution.

In the Argentinian Dry Andes the annual snow melt is the main source of superficial water and aquifer recharge. In this study we analyze the link between the seasonal and interannual variations of the discharge measured upstream the first dams on four rivers (Mendoza, Tunuyán, Diamante, Atuel) of the Argentinian Cuyo region (in the Federal Province of Mendoza) and those of the snow bed extent as mapped by optical remote sensing (MODIS MOD10A2 product) on a weekly basis in the 2001-2014 period, at the scale of each watershed.

Seasonal and interannual variations in the discharge appear directly related to those of the snow bed surface in each watershed. In particular we observe that the high-water period (September-April) discharge is directly related to the snow extent at the beginning of the snowmelt period, i.e. in September and October, as revealed by an about 0.8 correlation. Moreover the decreasing trend in the winter snow bed extent from 2001 to 2014 clearly explains the observed decreasing trend of the annual water discharge.

Agriculture and human activities in these oases mostly depend on the rivers discharge during the high-water period which in our results clearly depends on the snow extent. Our results indicate that it is possible to use remote sensing to forecast the average discharge in the September – April period (high water season) from MOD10A2 images with an average uncertainty of 15%. As MOD10A2 data are freely available ten days after acquisition, it is possible anticipating in early October the risk of water shortage in the following high-water period.

The Arctic is highly sensitive to changes in the global climate. One of the most prominent examples in this context is the shrinking sea ice cover over the past decades. The impact of the change in ice coverage on the global climate requires a thorough understanding of ice dynamics, as drifting sea ice transports salt and heat and in this way influences ocean dynamics. Sea ice motion is mainly driven by wind, ocean currents and internal ice stress. Convergent ice motion causes features such as ridges and rubble fields, which change the momentum exchange between atmosphere, ice and ocean. Openings in the ice due to divergent motion increase the exchange of heat and matter between ocean and atmosphere. On time scales of days to weeks, linear kinematic features, such as leads and ridges, evolve on spatial scales ranging from meters to tens and hundreds of kilometers. These features also emerge in numerical sea ice models when the resolution of the simulations is increased to a few kilometers. While plausible, their realism in the simulations is yet unclear and requires a detailed evaluation, with the help of (in our case satellite-based) observations.

We use Arctic-wide MITgcm simulations for 2006 at a spatial resolution of approximately 4km for a regional comparison with microwave satellite observations, e.g. from Synthetic Aperture Radar (SAR). We derive sea ice displacement from a sequence of satellite images by measuring the offset between matching patterns in different images. Discontinuities in the resulting velocity field indicate regions of instantaneous deformation that occur at some point in the time interval between the acquisition of the two SAR images used for displacement retrieval. The obtained quantities of deformation by divergence, shear and vorticity are scale-dependent. As a consequence, they depend on the spatial resolution of the SAR images and differ from the quantities calculated by the model. Hence, the comparison between model simulations and results of retrievals from remote sensing data is not straightforward. Therefore, we pay special attention to the spatial and temporal scales of the observed / modelled processes and introduce appropriate statistical tools.

By evaluating the kinematic features in the results of high-resolution sea ice models based on the microwave remote sensing data we expect to be able to assess and improve the ice rheology in these sea ice models.

The GlobPermafrost project develops, validates and implements information products to support the research communities and related international organisations like IPA and CliC in their work on understanding permafrost better by integration of EO data.

Permafrost cannot be directly detected from space, but many surface features of permafrost terrains and typical periglacial landforms are observable with a variety of EO sensors ranging from very high to medium resolution in various wavelengths. Prototype cases will cover different aspects of permafrost by integrating in situ measurements of subsurface permafrost properties (active layer depth, active layer and permafrost temperatures, organic layer thickness, liquid water content in the active layer and permafrost), surface properties (vegetation cover, snow depth) and modelling to provide a better understanding of permafrost today. The techniques will extend point source process and permafrost monitoring to a broader spatial domain, to support permafrost distribution modelling and mapping techniques implemented in a GIS framework and will complement active layer and thermal observing networks.

Initial user requirements have been gathered at the DUE-IPA-GTNP-CliC workshop in Frascati in February 2014, which have been further consolidated within the Permafrost community during 2014 in request of the WMO Polar Space Task Group. A subset of these requirements will be demonstrated within GlobPermafrost and assessed by user organisations:

Circumpolar permafrost extend

Permafrost dedicated land cover class prototype

Local cold spots

Regional transects for “hot spot” identification

Mountain permafrost areas

The N-ICE2015 cruise, led by the Norwegian Polar Institute (NPI), was a drift experiment with the research vessel R/V Lance from January to June 2015. The ship started the drift at 83°14.45' N, 21°31.41' E in the region north of Svalbard. When she drifted free the ship was reinserted into the ice at similar latitude and the drift was repeated.  In total 4 ice stations where installed during the 6 months; studying the complex ocean-sea ice-atmosphere system with an interdisciplinary approach.

During the N-ICE2015 cruise extensive ice thickness and snow depth measurements were performed during both winter and spring conditions. After the end of the N-ICE2015 cruise, the sea ice drifted south-west into the Fram Strait where the same sea ice pack was measured again in summer conditions during the NPI Fram Strait cruise in August and September 2015. To track the ice drift and to compliment the both field campaigns we use the ice and snow mass balance data from the autonomous buoys at the start of the N-ICE2015 cruise.  

Total ice and snow thickness was measured with ground-based and airborne electromagnetic instruments like EM31, GEM, and EM-bird; snow depth was measured with a magnaprobe. The comparison of the local datasets to regional scale measurements with the airborne EM-bird and the ice and snow mass balance buoys confirm that the local, smaller scale measurements on the ice stations are representative for the area. The ice mass balance data set can also be set into a temporal context by comparison to data from previous cruises in the same area.

Snow and ice thickness measurements were performed on repeated transects to quantify the ice growth or decay as well as the snow accumulation and melt rate. Additionally, we collected large numbers of independent values on surveys to determine the general ice thickness distribution. For calibrating electromagnetic surveys, direct measurements were made regularly in drillholes. Simultaneously, also freeboard was measured, which is an important parameter for altimeter-derived ice thickness datasets like CyoSat-2.

In terms of mass balance, average snow depths of 32 cm on first year ice, and 52 cm on multiyear ice were measured in January, the mean snow depth on all ice types even increased by the end of March to 49 cm. Compared with older data (Warren et al., 1999), the snow thickness was above average. The mean total ice and snow thickness in winter conditions was 192 cm.

During winter we found an unusually small ice growth rate of about 15 cm in 2 months, due to above-average snow depths and extraordinary storm events that came along with mild temperatures. In contrast thereto, we also were able to study new ice formation and thin ice on newly formed leads.

In summer conditions an extremely rapid melt rate, mainly driven by a warm Atlantic water inflow in the marginal ice zone, was observed during two ice stations with highest values of 20 cm per 24 hours.

Since in-situ measurements of the Arctic sea ice thickness and its snow cover are fairly rare, the data set presented here can be used for validation of diverse remote sensing applications. Additionally, this data set is an essential parameter for understanding the ocean-ice-atmosphere interactions, for calculating energy fluxes, and biogeochemical processes. 

Close monitoring of the tributary glaciers' mass balance is important in order to estimate future water discharge availability. Because of the vast area glaciers cover and the remoteness of these regions, a cost effective remote sensing observation is required to assess elevation changes. TanDEM-X, the first bistatic SAR mission in space, provides a valuable tool to investigate elevation change and mass balance of glaciers with unprecedented resolution and accuracy.

In this study we focus on Inylchek glacier in Kyrgyzstan, which is the largest glacier of the Tien Shan mountain range with a length of over 60 km. We analysed its height change by differencing digital elevation models (DEMs) from the C-SRTM mission and the TanDEM-X mission for time periods between February 2000, February 2012, March 2013 and November 2013.

Our results show for the last decade a remarkable mass loss distributed over the entire ablation and parts of the accumulation area. The interannual comparison of the TanDEM-X DEMs reveal an increase in mass loss as compared to the 2000-20013 annual mean: the Southern Inylchek branch lost 0.68±0.89 m w.e. between 2012-2013 compared to a mean mass loss of 0.23±0.34 m w.e. per year between 2000-2013. Similarly,  the Northern Inylchek lost 0.63±0.89 m w.e. between 2012-2013 compared to a mean mass loss of 0.42±0.34 m w.e. per year between 2000-2013. Highest decline occurs over the debris-covered lower ablation areas, which is especially prominent for Northern Inylchek, where we observe average elevation loss rates of ca. 3 m per year in contrast to 1 m elevation loss per year for Southern Inylchek.  

Both glacier branches show slight positive elevation change rates with a maximum of approximately 1 m per year only above 4900 m a.s.l. which  does not suffice to balance out the negative elevation changes rates of the larger glacier regions below. We therefore anticipate in the future a further thinning of the glacier branches. Besides the positive elevation change rates in the high altitudes, we observe a significant glacier thickening between 2012 and 2013 in the middle part of the Northern Inylchek’s ablation area (3650 m a.s.l.). This suggests a recent surging of that glacier region, which might be a follow-up to the surge of the lower Northern Inylchek ablation area that occured between 1990-1999.

As an extension of our work we also generate a surface velocity map of the entire Inylchek area by applying the amplitude tracking method on ESA’s Sentinel-1 radar data. With its wide swath and short revisit time the Sentinel-1 is an ideal sensor for continuously retrieving the kinematics of a glacier as large as Inylchek.

The SPOT5 (Take5) experiment conducted by CNES and ESA from April to September 2015 has also successfully acquired images for several sites in the Arctic and Subarctic. Non-glaciated or only partially glaciated terrestrial Arctic sites of the SPOT5 (Take5) experiment with successful image acquisition are (Alaska) Teshekpuk Lake, (Canada) Baffin Island, (Russia) Central Yamal, (Russia) Igarka, (Russia) Central Lena Delta, and (Russia) Kytalyk. The SPOT5 (Take5) Finland Sodankyla site has been specifically successful with relatively cloud free conditions however is not underlain by permafrost. (Norway, Svalbard) Kronebreen had very cloudy conditions. Most of the Arctic SPOT5(Take5) monitoring sites include the WMO/Polar Space Task Group (PSTG) ‘cold spots’. These are sites of interest of terrestrial arctic research with research stations, multi-measurement fields and long-term commitment for ecosystems and permafrost studies (community white paper in response to the WMO PSTG for Permafrost related Remote Sensing: Requirements for Monitoring of Permafrost in Polar Regions, Bartsch et al. 2014).

What are the potential future SENTINEL-2 applications for Northern permafrost landscapes? We will show examples of derived environmental information for the two ‘cold spots’ i) central Lena Delta (Siberia) (Samoylov permafrost observatory ) and ii) central Yamal (Western Siberia), (Vaskiny Dachi permafrost observatory ) such as hydrodynamics in lakes, snow in valley systems, vegetation phenology.

The SPOT5(Take5) level2 data will be exploited using vegetation indices such as NDVI, red/green ratio and ‘relative absorption depth’ that is indicative for the chlorophyll pigment absorption. Are these vegetation indices useful phenological indicators and therefore also classification tools for the tundra ecosystems? NDVI seems critical due to missing sensitivity for vegetation coloring and the frequent wet surface conditions of Arctic ecosystems.

First investigations already show the high potential to use the spectral information in the visible wavelength range and to develop phenologically-based classificators, e.g. analyses of early summer and late summer stages in central Yamal and central Lena Delta. Fig 1 shows SPOT5 derived Chl-a pigment absorption (relative absorption depth of red spectral band) in early and late summer 2015 in central Yamal (Western Siberia) and central Lena delta.

Reliable information on snow cover across the Northern Hemisphere and Arctic and sub-Arctic regions is needed for climate monitoring, for understanding the Arctic climate system, and for the evaluation of the role of snow cover and its feedback in climate models. In addition to being of significant interest for climatological investigations, reliable information on snow cover is of high value for the purpose of hydrological forecasting (related e.g. to floods and fresh water supply) and numerical weather prediction. Monitoring snow cover across the Northern Hemisphere is complicated by large differences on regional and local conditions as well as large gaps and biases in surface observing networks. Therefore satellite-based observations provide the only spatially and temporally extensive means for monitoring the snow cover information on the hemispherical scale.

The European Space Agency (ESA) Data User Element (DUE) GlobSnow project produced two hemispherical records of snow parameters intended for climate research purposes. The datasets contain satellite-retrieved information on snow extent (SE) and snow water equivalent (SWE) extending 17 and 35 years respectively. The dataset on snow extent is based on optical data of Envisat AATSR and ERS-2 ATSR-2 sensors covering Northern Hemisphere between years 1995 to 2012 [1]. The record on snow water equivalent is produced using a combination of passive microwave radiometer and ground-based weather station data, spanning years 1979 to 2014 [2].

The GlobSnow SWE record [2] utilizes a data-assimilation based approach combining space-borne passive radiometer data (SMMR, SSM/I and SSMIS) with data from ground-based synoptic weather stations. The satellite sensors utilized provide data at K- and Ka-bands (19 GHz and 37 GHz respectively) at a spatial resolution of approximately 25 km. The SWE record is produced on a daily, weekly and monthly basis. SWE information is provided for terrestrial non-mountainous regions of Northern Hemisphere, excluding glaciers and Greenland.

The GlobSnow SE processing system applies optical measurements in the visual-to-thermal part of the electromagnetic spectrum acquired by the ERS-2 sensor ATSR-2 and the Envisat sensor AATSR. The snow cover information is retrieved by a semi-empirical reflectance model-based SCAmod-algorithm [1]. Clouds are detected by a cloud-cover retrieval algorithm (SCDA, developed for GlobSnow purposes) and masked out. The product is provided in a latitude-longitude grid with a 0.01 degree spatial resolution (approximately 1 km) covering the Northern Hemisphere from 25°N to 84°N, which corresponds to the seasonally snow covered land areas of the Northern Hemisphere. With the sudden death of Envisat in April 2012 the GlobSnow SE product could not be produced. The data record will be extended with the upcoming Sentinel-3 satellite providing improved coverage and better spectral capabilities for snow monitoring.

The uncertainty associated with currently existing hemispherical datasets on both Snow Extent and Snow Water Equivalent are large. There are significant differences even between the well-established data products as demonstrated in the current ESA SnowPEx project, investigating and inter-comparing the available SE and SWE products. As reported in [3] the climatologies of total Northern Hemisphere snow water mass (SWM) vary among different observed and modelled datasets by as much as 50%. This underlines the current gap in our capability to correctly characterize the variability in the seasonal pack over hemispherical scales. Additional investigations within the ESA SnowPEx efforts indicate differences of similar magnitude between different remote sensed SWE datasets as well as large differences between SE products.

Analyses comparing satellite-retrieved SWE data with CMIP5 climate model simulations over hemispherical and multi-decadal scales indicate an even larger spread between the various CMIP5 models and those seen between satellite-based SWE datasets. Combining these findings with the reported dramatic changes observed in the spring time decrease of Northern Hemisphere snow cover, [4] which has been even more rapid than that observed for the late summer Arctic sea ice, highlights the need for improving our efforts for hemispherical scale snow cover monitoring.

A brief background on the current state-of-the-art as well as the justifications for further developments will be described.  Accordingly, a roadmap to improve our understanding and capabilities in monitoring long term changes in terrestrial snow cover (for both SE and SWE) will be presented.

REFERENCES:

[1] Metsämäki, S., Pulliainen, J., Salminen, M., Luojus, K., Wiesmann, A., Solberg, R., Böttcher, K., Hiltunen, M. and Ripper, E., "Introduction to GlobSnow Snow Extent products with considerations for accuracy assessment", Remote Sensing of Environment, Vol. 156, January 2015, pp. 96-108. DOI: 10.1016/J.RSE.2014.09.018

[2] Takala, M., Luojus, K., Pulliainen, J., Derksen, C., Lemmetyinen, J., Kärnä, J.-P, Koskinen, J., Bojkov, B., 2011, “Estimating northern hemisphere snow water equivalent for climate research through assimilation of space-borne radiometer data and ground-based measurements”, Remote Sensing of Environment, Vol. 115, Issue 12, 15 December 2011, Pages 3517-3529, ISSN 0034-4257. DOI: 10.1016/J.RSE.2011.08.014

[3] L. R. Mudryk, C. Derksen, P. J. Kushner, and R. Brown, 2015: Characterization of Northern Hemisphere Snow Water Equivalent Datasets, 1981–2010. J. Climate, 28, 8037–8051. DOI: 10.1175/JCLI-D-15-0229.1

[4] Derksen, C., and R. Brown, 2012, Spring snow cover extent reductions in the 2008-2012 period exceeding climate model projections. Geophys. Res. Lett., 39. DOI:10.1029/2012GL053387

Iceberg detection is important because of the expected growth in ship traffic and offshore operations in the Arctic. Icebergs are also important for the mass flux from marine terminating glaciers to the ocean. Studies of iceberg detection in the Franz Josef Land region have been performed by use of optical images in combination with Synthetic Aperture Radar images. Alternating Polarization (AP) images (HH- and VV-pol) from ASAR, Terra SAR-X and RADARSAT (Scan-SAR wide and Scan-SAR narrow) images were tested against Landsat and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) images for identification of icebergs of size between 50 and 400 m. The three types of SAR images had fairly similar capability to detect these icebergs, with the HH-pol from ASAR showing the most reliable results. The ScanSAR Narrow image failed to observe the smallest icebergs of size around 50 m. Terra SAR-X image is more capable in detection than the ASAR (AP) image may be due higher resolution of the Terra SAR. The HH polarization in the ASAR image is better than the HV polarization. The icebergs were located in fastice with relatively homogeneous backscatter which was favourable for iceberg detection. The studies showed that combined use of optical and SAR images for iceberg detection gives better results compared to previous studies where SAR and optical images were used separately.

Also Analysis carried out for iceberg distributions in Barents based on remote sensing data (optical and SAR) near the boundaries of the outlet glaciers in Franz Josef Land. Glaciers with good coverage of high-resolution SAR and optical images have been selected. The average number of icebergs per unit area has been calculated for these regions. These estimates have been compared with the historical maps, presented in Atlas by Abramov, which are based on air reconnaissance data and ship observations. The changes in iceberg distribution in the areas under study has been assessed and analyzed. Now, data from Multispectral Instrument on Sentinel-2 will be a useful tool in combination with SAR data from Sentinel-1 to study icebergs in the Arctic regions.

To implement ESA's Climate Change Initiative, data products are needed to monitor essential climate variables, such as the state of the Greenland Ice Sheet. Surface elevation change data from altimetry is one of a suite of parameters used to characterise ice sheets, and can provide spatially regular, consistently calibrated, long-term time series. This data is necessary for monitoring such globally important, rapidly changing regions and facilitating scientific enquiry into the links between ice sheets and climate change.

We present a comparison of surface elevation change datasets over Northern Greenland derived from Envisat, using repeat-track, modelled surface and crossover algorithms, and characterise them in terms that quantify their scientific value, such as geographical coverage, temporal extent and accuracy. This analysis is conducted over both inland ice sheets and towards the Greenland coast, where rapid change is taking place. It uses data from November 2002 to October 2010. We also compare the satellite data to ground truth from NASA's Operation IceBridge mission.

The Antarctic is one of the rare regions on Earth that are not influenced by mankind so far except through indirect climate effects. Nevertheless, activities either of scientific or touristic manner are becoming more common. The logistics for ships sailing through the Drake Passage and around the Antarctic Peninsula are considered as most important. In order to support these activities and to protect this unique environment, the German Remote Sensing Data Center (DFD) of the German Aerospace Center (DLR) in collaboration with the Maritime Safety Research Departments at DLR Neustrelitz/Bremen is developing a near-real-time (NRT) service in support of monitoring icebergs and sea ice for the Antarctic Peninsula. DLR’s German Antarctic Receiving Station GARS O’Higgins located at the Northern tip of the Antarctic Peninsula will serve as the data receiving and processing platform.

In support of maritime operations in the vicinity of the Antarctic Peninsula, already developed algorithms (by the Maritime Safety Research Departments at DLR Neustrelitz/Bremen) for the detection of icebergs, wind fields, waves, and the classification of sea ice will be used. These have to be adapted to Antarctic regimes as these are different to the Arctic. High-resolution Synthetic Aperture Radar (SAR) data has proven its capability to be most useful as it is neither influenced by clouds nor dependent on daylight. Currently, TerraSAR-X data is used. Some algorithms are already able to use Sentinel data and can be adapted to future satellite missions. In order to be able to deliver products in near-real time, the processing chain will be installed at GARS O’Higgins station so that data can be recorded and down-linked within a short time frame. After the reception, the data will be processed and post-processed on-site. The value added products would be available within 15-30 minutes after the reception. In order to gather experience with regard to requirements of the customers, pilot studies are already carried out supporting some vessels on their way to Antarctica.

Permafrost variations in terms of hydrological budget changes in Siberia (Russia) are one of the most challenging contributions for climate change and the eco-system of the living planet Earth. Therefore, monitoring of the change pattern in this partly frozen area is a key factor for understanding the underlying processes and for predicting possible consequences. There are many types of satellite sensors and systems for monitoring the geometrical and physical properties of the Earth’s surface such as gravimetric satellites like the twin-GRACE mission for observing Earth’s gravity field variations, active radar altimetry satellites (e.g. Jason-2) to observe lake and river level changes as well as satellite imagery to estimate lake and river extent variations.

 

In this study, the volume variations of water stored in big lakes and reservoirs are estimated based on satellite altimetry and satellite imagery techniques.  Those variations are an indicator for the complex permafrost thawing process in Siberia. Related mass variations obtained from combined satellite altimetry and imagery is compared to the GRACE results.

The Norwegian Satellite Earth Observation Database for Marine and Polar Research (NORMAP) has created a data repository for the Nordic Seas and Arctic Ocean from Earth Observation data from polar orbiting satellites serving research and application in marine, polar and climate sciences. The NORMAP infrastructure (http://normap.nersc.no/) facilitates easy and seamless access to remote sensing products for the scientific community who may not be expert satellite users. The development project started in 2010 and is currently working to secure sustainability beyond the development period that terminates in September 2016. The primary region of interest for NORMAP is north of 55ºN and the product portfolio is not static. Currently the system contains remote sensing products for sea ice parameters, surface wind information, sea surface temperature, absolute dynamic topography, and ocean colour. This product portfolio is continuously evolving.

NORMAP is developing a distributed data management system that integrates physically separated data repositories. These data repositories are filled with existing and new products generated from satellite remote sensing data. To support the scientists working with remote sensing products, a Python toolbox, NANSAT, has been developed (https://github.com/nansencenter/nansat/wiki). In this presentation we will demonstrate the interactive capability offered by the NORMAP repository and NANSAT functions to read and export geospatial data, to set scientific meaning to the data and to implement basic operations for data integration, harmonization and visualization including product overlays.

With the upcoming French-Chinese CFOSAT mission (carrying the SWIM real-aperture radar (RAR) system at Ku-band pointing at incidence angles between 0 and 10°) and the French-US SWOT mission (with Ka-band Radar Interferometer – KaRIN instrument with look angles limited to less than 4.5° providing a 120-km wide swath), there is a need to further understand the radar backscattering at near-nadir for these 2 wavelengths. 

Meanwhile, the “Core” satellite of the Global Precipitation Mission (GPM) carrying the first space-borne Ku/Ka-band Dual-frequency Precipitation Radar (DPR) has been launched in Feb. 2014. The DPR system provides three dimensional measurements of precipitation structure over 125 km (Ka-band) and 245 km (Ku-band) swaths at a horizontal spatial resolution of 5 km. However, the radar backscattering from the Earth surface can be measured when no rain occurs, with incidence angle ranging from ± 9° and ± 18° for the Ka-band and Ku-band, respectively. With the capacity to acquire data up to ± 65° North and South, the KuPR and KaPR systems are suitable to analyze the radar backscattering over sea ice.

In this study, the following analysis will be shown:

-  We first analyze GPM data over Antarctica. Seasonal time series over specific locations are analyzed, angular fall-off being specifically studied. Different scattering mechanisms from the freeze-up to the melting period are identified.  Monthly sea ice products from the KaPR or KuPR data stacks are generated. At a first order, they can be associated to a proxy of sea ice concentration. A comparison with NSIDC sea ice concentration products is carried out.

 - In order to further understand the near-nadir backscattering at Ku and Ka-bands, an assessment with respect to sea ice type is performed over the Hudson Bay and Labrador Sea. Ice charts from the Canadian Ice Service are collocated with GPM tracks.

Snow on sea ice is a major source of uncertainty in both ice thickness and ice concentration retrieval from satellite data. The potential of the dual-frequency Ku and Ka-band near-nadir data is analyzed and completed with Electromagnetic modeling. 

Since its launch in April 2010 Cryosat has been collecting valuable sea ice data over the Arctic region.

Over the same period ESA’s CryoVEx and NASA IceBridge validation campaigns have been collecting a unique set of coincident airborne measurements in the Arctic.  The CryoVal-SI project has collated the campaign data and selected the CryoSat ground tracks which have the best coverage by coincident campaign data to produce a list of “Golden Days”.   The campaign data were processed and resampled to coincident CryoSat footprints to make them easier to utilise.  The resulting validation dataset provides an independent metric that can be used to objectively evaluate any experimental changes or refinements to the CryoSat data processing that a user may wish to test.

This valuable resource is in itself an output of the CryoVal-SI project which is made openly and freely available to the scientific community.   In this talk we will describe the composition of the validation dataset, summarising how it was processed and how to understand the content and format of the data.  We will also explain how to access the data and the supporting documentation.

A continuous monitoring of sea level is particularly necessary along the coastal zones as these areas are the most exposed to flooding and storm surges that dramatically affect local economy and society. The aim of this project is to study the annual sea level variability in the Bay of Bengal exploiting five years of CryoSat-2 data. The work is focused on the analysis of CryoSat’s performances over coastal areas with using 20Hz SAR data from SARvatore online processing toolbox. This research is also an opportunity for comparing SIRAL’s LRM and SAR modes, which have covered the same area for different time frames: LRM from 2010 to October 2012 and SAR from October 2012 to present. Finally, SARvatore is compared with independent datasets like the ESA CCI product, AVISO and RADS.

Ice drift has a big impact on both operational needs and global climate system. Systematic monitoring of sea ice motion is started more than 100 years ago, but it became really operational since satellite era began. Only remote sensing data allow us to obtain ice drift fields over the entire Arctic and sub-seas. Due to increasing of navigation in ice in the Arctic region, sea ice drift retrieval is becoming one of the crucial task. But there are many difficulties caused by features of satellite instruments like SAR. For example a speckle noise which take place in SAR imageries force to use advanced filtering technique to detect and describe ice features to track.

Ice drift retrieval from satellite data is not a new thing but development of high resolution ice drift retrieval methods is ongoing. Proposed method is based on the most famous member of Computer Vision family algorithms - SIFT (Scale-invariant feature transform). Some improvements were made to adopt it for SAR images from Sentinel 1a. Now it is possible to obtain ice drift product of 1 km resolution from wide swath image pair (with the spatial resolution of ~100 m/pixel). The main idea is to detect a number of feature points in the first image and then to detect the same points in the second one. Also advanced filtering scheme was applied as preprocessing step. Then the feature points descriptors is construct. We propose a more robust method for SIFT descriptors. The algorithm was tested on Radarsat-2, archive ASAR and Sentinel 1a images and it showed good robustness against rotation, gray level variance and illumination changes. The results looks promising and we aimed to run this algorithm in operational mode as part of the SONARC system (“Development of sea ice monitoring and forecasting system to support safe operations and navigation in Arctic Seas” project).

Observed and projected climate change in the Arctic increases the vulnerability of terrestrial ecosystems to disturbances. For example, significant increases in air temperatures especially in high latitudes will impact the stability of permafrost landscapes that cover 24% of the northern hemisphere and dominate large parts of the Arctic. Resulting potentially large-scale permafrost thaw and subsequent changes in geomorphology, hydrology, and vegetation as well as mobilization of previously frozen soil organic matter will provide regional to global scale feedbacks that require better quantification. The release of this carbon as methane and carbon dioxide, both active greenhouse gases, would result in a positive feedback loop with climate warming (“permafrost carbon feedback”).

Permafrost affected areas are distributed over different climatic and biogeographic zones, from coastal tundra lowlands to continental boreal or mountainous regions and are therefore exposed to different natural conditions. So far, the thermokarst lake cycle, including lake formation, expansion, and drainage is monitored only within limited areas. These localized studies revealed different magnitudes and trajectories of lake area changes ranging from a net gain to a net loss of lake surface area. However, a consistent monitoring approach and comparison of permafrost thaw processes across broader spatial scales has not been accomplished yet.

Within our study, we expand the monitoring of thermokarst lake dynamics over many different permafrost regions of diverging climatic and biogeographical conditions across North America and Siberia. We focus on various transects from high Arctic coastal tundra lowlands to continental boreal regions, covering a representative region of >1,000,000 km².

With a robust trend analysis of key surface indicators (multi-spectral indices) on data from the entire Landsat archive, we observe land surface transitions on decadal scales (1985-2015) at high spatial (30 m) and temporal resolution. The transferability and quick application of this pre-analysis allows the quantification and comparison of land surface transitions and disturbances within and among different study sites. A (semi-)supervised machine-learning system is fed with the calculated trend- and morphological parameters to classify thermokarst related landscape changes and -processes. The calculated results are compared to existing data sources to estimate the accuracy of the spatially up-scaled results.

Using such large datasets it becomes necessary to introduce process automation. Therefore, we developed a highly automated workflow, which takes care of the repeated processes without user interaction and can be quickly adapted to new study sites or regions.

New knowledge on the magnitude and extent of thaw processes allows for a thorough characterization of potential impacts of climate change on different types of permafrost regions. With an integration of data sources, such as soil carbon storage, the quantification of thermokarst lake-related carbon fluxes can be estimated with unprecedented accuracy.

Glaciers and icecaps are dynamic landforms in the sense that climate has a direct effect on their geometry. Hence their elevation changes over time, which is observable through remote sensing on the scale of decades or even within a season. Consequently, a bias might exist between the exact topography during acquisition and the expected topography, or an available digital elevation model (DEM), respectively. This bias propagates as a projection error into a product which was orthorectified using a DEM that is non-synchronous with image acquisition and results in a geolocation error. This effect is especially present in areas with large elevation changes over time or/and at large looking angles. In this study we estimate this bias and assess its consequences for current pushbroom sensors, at several sites.

From a technical point of view, this effect is especially of interest for the newly launched wide-angle Sentinel-2 MSI sensor. With our methodology geolocation errors can be corrected and by doing so, increase the compatibility between different observation systems, in particular when different DEMs are used for orthorectifying, and/or when the systems have different viewing angles and orbits. Thus the correction is of benefit for co-registration purposes between products from different sources, but in addition it is a direct elevation bias estimate.

The largest observable elevation differences are at glacier termini, as well as in steep topography, where DEMs tend to be outdated or contain errors. For this study, the elevation changes on the glacier are of particular interest. However, the glacier moves inbetween acquisitions, thus geolocation errors are superimposed onto the glacial movement. Fortunately, our methodology is able to couple the estimation of velocity and elevation change. This results in an opportunity to assess glacier dynamics in space and time from data of different sources, or from the same sensor at different orbits.

The combination of this methodology and the wide looking angle of the Sentinel-2 MSI sensor provide a more precise look into current dynamic surface processes. However, the methodology also thrives on older remote sensing systems, opening the door to more precise exploitation of longer time series. This will hopefully improve quantification of the dynamic component of glaciers and there adjustment to the current climate.

Arctic lakes and their ice regimes are strong indicators of climate change at high latitudes. Changes in the ice phenology of Arctic lakes have major implications for the physical and biogeochemical processes, and the local climate as they are strongly dependent on the presence of ice. In a permafrost-rich region such as the Lena Delta in Arctic Siberia, shallow lakes are also associated with carbon dioxide and methane release consequent to permafrost thaw underneath the lakes. The presence of clouds, prolonged darkness or low sun elevation at northern latitudes limits the use of optical sensors during key months for lake ice phenology. With day and  night observations, in all weather conditions, at spatial resolutions raging from a few meters to 150 m, synthetic aperture radar (SAR) sensors provide year-round acquisitions, critical for lake ice regimes monitoring.

For the purpose of this study, optical data was essential to validate and compliment the SAR observations whenever available. The study shows employs the monitoring the evolution of radar return, also referred to as radar backscatter intensity or sigma-nought (σ°), for detecting ice melt onset and end of the ice season, as well as floating or grounded ice fraction (ice frozen to lakebed). Following image clustering and σ° analysis, results were examined against meteorological station observations and model simulations of ice thickness and break-up dates obtained with the Canadian Lake Ice Model (CLIMo).

Recent studies have shown that in other Arctic regions fewer lakes are freezing to bed during the winter (i.e. North Slope of Alaska) or are experiencing earlier spring melt (i.e. The Canadian High Arctic). Situated in the continuous permafrost zone and at the north Siberian land-ocean boundary, the lake-rich Lena River Delta area, is likely a very sensitive region to recent changes in climate conditions at high latitudes. Longer-term ice regimes of lakes in the Lena Delta have not yet been investigated and changes that these thermokarst lakes have undergone during recent years and the current state of ice conditions remain unknown. This study presents unexpected results from the analysis of a time series combination of C-band SAR (RADARSAT-1/2 ScanSAR Wide Swath, ASAR Wide Swath, Sentinel-1A Extra Wide Swath) and optical (Landsat 7 ETM+) satellite acquisitions, between 1997-2015. Results show the fraction of lakes that freeze to bed in late winter, as well as the timing of melt onset and end of break-up for thermokarst lakes for a sub-region in the western Lena Delta (73°25’ N, 124°31’ E), Siberia.

Sentinel-2 (S2) features a number of characteristics that will improve mapping and monitoring of glaciers, permafrost and related hazards, among which the large swath width of 290km, the spatial resolution of 10-20m, and the repeat cycle of at least 10 days (high towards the poles). In this study we perform a number of general tests on image radiometry and geometry as relevant to the glaciological image analysis. Based on commissioning-phase data, we find a geolocation accuracy of one pixel (at 10m) or better and an co-registration accuracy between repeat scenes of around 1/3 pixel. Both error magnitudes are well acceptable for most glaciological applications. Cross-track offsets in orthoprojected L1C data due to vertical errors in the DEM used have, however, to be observed. In particular at glacier tongues, DEMs will typically be outdated. For some examples in the Swiss Alps we found, for instance, lateral offsets in S2 images of 30-40 m over glacier tongues. Geolocation biases in S2-derived products would for instance affect glacier outlines, especially when compared to other data such as Landsat, but can be avoided to a large extent for glacier velocity measurements by relying on repeat data from the relative same orbit.

Through a number of case studies, we demonstrate and evaluate the capability of S2 for glaciological applications: Automatic multispectral glacier mapping based on S2 bands 4 (red) and 11 (SWIR) turns out to be very successful, among others due to the improved resolution compared to Landsat data. This improved resolution together with the high radiometric fidelity is also important for detecting and assessing glacier lakes and their changes over time. From S2 data it becomes possible to track velocities of smaller glaciers and even over seasonal scales, as we demonstrate for the European Alps, the Caucasus and Greenland. This opens up for the possibility of obtaining both summer and annual velocities from the same sensor. As a novel application, we also demonstrate that repeat S2 data can be used for inventorying the speed of rockglaciers (creeping mountain permafrost) on regional scales.

We combine satellite derived estimates of ice-shelf thinning rates and ice shelf flow velocities of Antarctica to arrive a new and improved values for ice-shelf melt rates.  Instead of calculating flux divergent directly from measurements of surface velocities we start by assimilating the velocities into an ice flow model. We then calculate flux divergence from modelled velocities. This gives us a much cleaner signal and a better spatial coverage. Estimates of ice-shelf melt rates for all Antarctic ice shelves are presented, and we discuss the possibility of generating time series of melt rates using this method.

The Norwegian Ice Service provide ice charts of the European part of the Arctic every weekday. The charts are produced from a manually interpretation of satellite data where SAR (Synthetic Aperture Radar) data plays a central role because of its high spatial resolution. After the launch of Sentinel-1A the number of available SAR data have significant increased making it difficult to utilize all the data in a manually process. This in combination with a user demand for a more frequent update of the ice conditions have made it important to focus the development on utilizing the high resolution Sentinel-1 data in an automatic sea ice concentration analysis.

A Bayesian SAR classification algorithm for Sentinel-1 data is developed with dynamic updates of the class statistics. The algorithm uses Sentinel-1 data in extra wide mode dual polarization (HH/HV) to separate ice and water in the full 40x40 meter spatial resolution. Thermal noise in the cross-polarized data have been estimated and removed from HV in the pre-processing. From the classification of ice/water the concentration is estimated in a 1 km resolution with concentration classes similar tot the ice charts. Results from the automatic classification will be presented.

Snow cover is a significant component of the hydrosphere in northern Scandinavia and provides seasonal flows that are important for local fluvial ecosystems and for energy production. Snow melt related floods are a threat to life, property and infrastructure; negative societal impacts of snowfall and snowmelt are projected to increase as a response to climate change. Snowpack properties also affect the ability of mammals to forage in winter with consequences for wildlife and managed reindeer populations. Snow cover has aquatic and terrestrial ecological importance and is a fundamental component of the climate system. Snowpack monitoring therefore fulfils societal, environmental and climatological requirements for information. Here we investigate methods to improve snowpack monitoring in northern forests using a time series of 11 polarimetric X-band synthetic aperture radar data acquired by the Tandem-X mission. Two sites are investigate, one in near Sulitjelma in northern Norway (67° N) and one at Hemavan, Sweden (65° N). Both sites drain into watersheds regulated for hydropower.
Scatterer phase centre heights were determined using polarimetric-interferometric digital elevation model differencing and phase differencing was performed. Co-polar phase differences have been related to snow depth in previous studies (Leinss et al., 2014). Spatio-temporal patterns in phase differences and scatterer phase centre heights were mapped, including across the treeline.  Polarimetric H/A/alpha decompositions were calculated to investigate scattering properties. Additionally measures such a polarisation ratio and polarimetric coherence were related to landcover, topography and snow cover. The results emphasise that snow depth mapping in PolSAR data is a non-trivial task and that further work is needed to provide physical explanations for patterns identified. Nevertheless, the results show that SAR responses can be related to snowpack parameters and that PolSAR techniques offer promising opportunities to advance spaceborne snow cover monitoring.

The remote sensing with an use of the SAR data for monitoring of snow cover and ice of the polar regions is consolidated with an one of the main techniques used in the glaciology. In this context, the modeling of the backscatter SAR data showed important to the extraction of information about the electromagnetic radiation behavior in the track of the microwaves along the snow package and ice cover, allowing a components division with the snow particle size, density, range, and others aspects. Thus, the generation of backscatter modeling of SAR data snow cover and the constitution of this modeling, is possible apply an inverse model, allowing extract the physical information about SAR data. The starting point of this study was to Radiation Transfer Model (RTM). The RTM takes into account the interactions between consistent waves in function of the distribution and the position of the target particles that interact with this wave showing is appropriate for modeling involving snow. In these approaches for modeling of the backscatter processes and interaction of microwaves with the snow and ice, a recent advance are the methods of backscatter integration methods of the volume proposed by, having show great results for the surfaces with differential rugosity and large dimensions. The work proposes an analytical modeling for interaction between a beam of microwaves in the X band vs.  snowpack for dry snow areas. To this end, statistical analysis were performed with SAR-X backscattering data vs. reference data from snowpits focusing the interaction between the microwave beam. Numerical methods were employed for solution of differential equations that make up this issue. The model was proposed regarding a study area including the Antarctic western portion, recognized as the Union Glacier. The backscattering modeling used was based under the assumption of RTM, considering as main variables the depth of accumulated snow, the surface roughness (air-snow interface and snow-ice), the size of snow crystals (grain size), the density profile of the accumulated snow and snow characteristics of the layers forming the surface snowpack (thickness, shape of the interface between layers variation between dielectric layers, among others). After that, reversal statistical model of backscatter was performed to estimate stratigraphic data allowing the local stratigraphy of estimated variables SAR backscatter from COSMO-SkyMed. To validate the proposed model, the same input data were considered for all experiments performed. These data were made up of stratigraphic data and snow temperature data in a profile 2m depth and SAR-X COSMO-SkyMed data (acquisition mode Stripmap / Himage with spatial resolution 3x3 m) acquired in Union Glacier snowpits and remote sensing SAR data during summer 2011-2012. The results showed average density of the snow pack surface from backscatter data SAR-X with R² ≥ 86%. The main contribution of this work is the resulting modeling for SAR-X backscattering for dry snow masses, which was proved to be statistically consistent with the ground truth data. Even with limited reference data, this result indicates the soundness of the proposed approach, allowing the spatial distribution of stratigraphic variables in dry snow areas from SAR-X data.

Permafrost is stated as an essential climate variable by the World Meteorological Organization and is an important physical landscape component of high-latitude environments.  The variability of the permafrost ecosystem parameters soil moisture (SM) as well as freeze-thaw (FT) has a strong impact on rapid permafrost degradation, on surface energy and water fluxes as well as on biogeochemical processes. Thus information about the mentioned parameters in high temporal and spatial resolution is important for the understanding of processes in permafrost landscapes. Synthetic aperture radar (SAR) operates independently of cloud coverage and polar night and today’s SAR satellite systems provide imagery with high temporal and spatial resolution. Existing operational satellite SAR data products of SM and FT are available only in coarse-scale resolution.

We are investigating high-spatial resolution SAR of TerraSAR-X (TSX), and in future ALOS-2, Sentinel-1, as well as optical very high resolution satellite imagery in combination with in-situ experimental monitoring data to investigate the spatiotemporal variability of permafrost disturbances, SM and FT on the watershed scale.

Our study site for rapid permafrost degradation is an actively eroding ice- and organic-rich permafrost riverbank from the so called Ice-Complex within the central Lena Delta, Siberia. Our studies on SM and FT focus on a small scale watershed on Herschel Island along the western Yukon Coast, Canada and can potentially be transferred to the Ice-Complex permafrost landscape in the Lena Delta.

Automated micro-stations with near to surface soil moisture and temperature sensors were installed in the Lena Delta (since 2013) and on Herschel Island (since 2015). Field work on Herschel Island and the Lena Delta included handheld soil moisture measurements as well as extensive soil sampling. In spring 2015 we conducted a GPS survey in the Lena Delta along the test site and installed a time-lapse camera as well as wooden poles with 50cm distance perpendicular to a rapidly eroding cliff top sequence. Time-lapse images were acquired from late June to late August.

We used TSX backscatter time-series from the years 2012, 2013, 2014 and 2015 to analyze rapidly eroding cliff tops along the riverbank within the central Lena Delta. Pre-processing was performed using the Next ESA SAR toolbox (NEST) and included radiometric calibration and conversion to backscatter coefficient sigma nought, multilooking and an ellipsoid corrected geocoding. We then used a threshold approach to visualize the transition line between undisturbed tundra surface and actively eroding cliff prior to mapping. Very high resolution orthorectified optical satellite images acquired in August 2010 and August 2014 were used as validation datasets for the TSX-derived results.

The TSX extracted annual retreat rates are in the same range as the ones from the optical reference dataset. The intra-annual TSX-derived cliff top retreat lines from 2014 at the test site showed rates of 2 to 3 m per month. The time-lapse field data at the same place showed similar results in summer 2015.

The TSX backscatter time-series show a high potential for the monitoring of rapid permafrost degradation with high spatial and temporal resolution. The results are valuable for the understanding of intra-seasonal permafrost degradation dynamics.

Future work on Herschel Island and the Lena Delta will focus on SM and FT dynamics on the watershed scale. ALOS-2, Sentinel-1 and TSX datasets are planned to be used and cross-validated with the field datasets.

The presented project is embedded in the German Helmholtz Alliance Earth System Dynamics (EDA) network and builds on existing datasets from the FP7 within the PAGE21 project. TSX-datasets were kindly provided by the Department Land Surface from the German Aerospace Agency (DLR).

The Ice Sailing European Championships are organized mostly in Sweden, Estonia and Poland. Polish team is among the world leaders, with numerous World and Europe titles. Every winter season there are at least several competitions in Poland. With climate changes, the new problems emerged.  Winter seasons become more warm and the days when the ice cover is present on the lake are more rare. This makes more and more difficult to predict whether the ice at the time of competitions will remain on the lake. The organizers are obligated to announce the final place of event from few to several days before a competition. For example, the place of The Great Western Challenge European edition have to be declared for 2 days before the event.  Thus, this is crucial to observe the ice cover on the lakes. Nowadays, the most popular method is the ground measurements. However, the use of satellite SAR data can facilitate observation of the lake surface area. Thanks to the high independence to weather conditions, SAR images can be used in winter seasons for such purpose.

In the study the five Sentinel-1 dual-pol images were used. The studied area was the south part of the Mazury lake district in Poland. In order to perform ice classification, the few method were tested. The first of method was Image Segmentation Using Watershed with Intensity-based Region Merging. The second and third method are based on the H/A/Alpha Decomposition. The second method uses the Bayes Classification and the last one are based on Wishart Classification. In order to increase the accuracy of classification, the ground-based data were used. The results show that SAR data can be very helpful in organizing of the ice sailing competitions. In the future, the other classification methods will be tested. The final results will potentially be applied by the ice sailing competition organizers.

Advances in polar altimetry have allowed studies of sea surface height (SSH) and ocean circulation to be made across the Arctic. Cryosat in particular has provided data of excellent quality and coverage, with lead retracking techniques allowing the retrieval of surface elevation data throughout the ice pack.

However the quality of some ocean products still lags behind those available for lower latitudes. For data related to dynamic topography and surface currents, contraints in geoid data and the lack of drifters inhibits the resolution that may be achieved compared to products for non-polar oceans.

We investigate potential improvments to surface current resolution at high latitudes using available regional data. By integrating data such as airborne gravity surveys and in-situ current velocities from moorings and ice-tethered profilers (ITPs) into our satellite product, we attempt to resolve more and smaller current features and gain greater insight into Arctic circulation.

Operational multi-sensor/multi-temporal fractional snow cover mapping has been available for about ten years for Scandinavia based on algorithms developed by NR and Norut through a service delivered by Kongsberg Satellite Services (KSAT). The product was originally based on fusion of data from Terra MODIS and ENVISAT ASAR. Now, a new generation of algorithms are developed by fusing Sentinel-1 SAR and Sentinel-3 OLCI/SLSTR data. The presentation shows the first results of single- sensor and fusion algorithms applied on Sentinel-1 and Sentinel-3 data (with Suomi NPP VIIRS as a backup).

The main weaknesses of current fractional snow cover (FSC) retrieval from optical data are accurate retrieval in forests and steep mountain terrain. Trees make it difficult to observe the full ground surface and create cast shadows darkening the snow surface. In steep terrain there are cast shadows only allowing indirect illumination of the ground surface. We develop local pixel-level models to compensate for both effects. The models utilise data from both SLSTR and OLCI sensors.

Radar backscatter from snow covered ground is a sum of contributions from the snow surface, the snow volume and the ground. Wet snow can be detected since the radar backscatter drop significantly. The performance of the wet snow detection algorithm is affected by the land-cover type and to some degree the snow wetness. To improve the accuracy of the algorithm we perform a sensitivity study for various land-cover types (forest, bogs, farm land and mountainous areas) and test whether adapted thresholds for individual land-cover types can give improvements. Another task is adaptation of the SAR wet snow detection algorithm to Sentinel-1 data and the transformation of the wet snow product into snow cover fraction (geocoding, land use maps, DEM) adapted to Sentinel-1 spatial resolution and data format. The improved spatial resolution of Sentinel-1 relative to ENVISAT ASAR and Radarsat-2 results in more accurate FSC retrieval.

Optical sensors are limited by cloud cover and daylight, whereas C-band SAR sensors may only map the snow as long as it is wet. As the two sensor types are measuring different physical phenomena, it is not straightforward to obtain a homogeneous and consistent snow map.  A new approach, based on state modelling and assimilation (data fusion at the electromagnetic level) using a hidden Markov model (HMM) has already proved success by combining data from other sensors. We have developed a novel approach to do a similar fusion based on Sentinel-1 and Sentinel-3 data that produces inter-sensor and time series consistency.

The new algorithms and FSC products are validated using high-resolution and very-high-resolution data based on Sentinel-2, Landsat 8 and VHR satellites like World View.

Why is the sea ice in the arctic melting faster than as predicted by models?

 

The interaction of water waves with sea ice has various implications involving interdisciplinary fields of research. Its range of influence extends from heat fluxes and water mixing at the ocean surface to land erosion, etc.. This leads to the issue of potential feedback processes between waves and ice, e.g. the break-up of ice by waves leads to a larger water/air interface. This, in turn, may be connected with enhanced absorption of solar radiation and less ice formation. All this is of particular relevance in the context of a changing wave and ice climate.

Here we present measurements from satellite data showing the interaction of waves and sea ice in the marginal ice zone  as compared to models. These results are compared to current state-of-the-art implementation of spectral wave prediction models. Overall, good agreement is observed, and

limitations of the remote sensing algorithm and the wave model are highlighted.

 

The need for wide-spread, up-to-date sea state observations in the emerging ice-free Arctic will further increase as the region will open up to marine operations.

 

As a result of climate change, sea ice conditions in the Arctic become increasingly attractive to maritime traffic in the summer months. However, shipping routes in the Arctic north remain affected through ice and icebergs. Timely variable ice conditions require continuous monitoring and where appropriate an adjustment of the route.

 

Generation of information on ice topography on a large scale is not possible with conventional space-based remote sensing technologies. Altimetry generates one-dimensional profiles (along the nadir point of the satellite) while repeat pass interferometry does not work under dynamic ice conditions (it can move driven by wind, currents and tides resulting in decorrelation of the interferometric SAR signal). Single pass bistatic SAR interferometry (InSAR), with and without polarimetric information, is a promising technique that is able to provide important information on ice deformation and topography with application to the build up of ice stresses and with the potential to identify possible bergs and deformation features of hazard to navigating through ice. The principle capability of bistatic TanDEM-X data to derive sea-ice topography has already been demonstrated in previous studies.

 

In this paper we present results of an experiment in which bistatic SAR data and ground information obtained during drifting buoys deployment and helicopter flights were jointly analyzed. The field campaign took place off Barrow, AK (USA) in March 2015 during which a set of drifting buoys were deployed as part of the International Arctic Buoy Programme (IABP). Coincident TanDEM-X image pairs were acquired. One goal of the experiment is the detection and measurement of flow bergs in sea-ice and of ice deformation features, like ridges, within fast ice areas.

The results show, that bistatic spaceborne SAR data is an excellent data source to measure ice deformation features with high spatial and vertical precision. Prominent ice deformation features are analyzed with available data from ground, air and space. Based on statistical analysis a vertical sensitivity of 10 cm or better can be assumed, allowing measuring ice topography with high accuracy. The paper further discusses favorable acquisition geometry conditions of bistatic TanDEM-X pairs and provides an outlook on its operational use in future.

Snow monitoring is essential for prediction of flooding due to rapid snowmelt, to provide snow avalanche risk forecasts and for water resource management – including hydropower production, agriculture, groundwater and drinking water. Snow wetness and snow liquid water are essential variables for monitoring the snow state and providing early warning of flood risk and snow avalanches during the melting season. The presentation shows the first results from the SnowBall project of single-sensor and fusion algorithms applied on Sentinel-1 and Sentinel-3 data (with Suomi NPP VIIRS as a backup) for frequent monitoring of the snow wetness during the melting seasons in Norway and Romania.

Sentinel-1 C-band SAR is sensitive to presence of wet snow. Wet snow can be detected since the radar backscatter drops significantly. However, with C-band SAR it is difficult to quantify how wet the snow is. Wet snow mapping into a set of five categories of wetness has been demonstrated in the past by NR using MODIS data. The combination of surface temperature and the temporal development of the effective snow grain size is used to infer approximately how wet the snow is. In the SnowBall project this approach is ported to the combined use of the Sentinel-3 OLCI and SLSTR sensors. The expected high radiometric quality of these sensors is furthermore expected to enable a more advanced algorithm giving approximate categories of snow liquid water (volume of liquid water per volume of snow) for the snow surface. Field measurements have been accomplished using spectroradiometer measurements and direct measurements of snow liquid water with a dielectric probe to develop a retrieval model. The retrieval model will be adapted to Sentinel-3 data and applied in the new algorithm.

Furthermore, to utilise the combined capability of Sentinel-1 and Sentinel-3 for more accurate retrieval and improved temporal coverage – given that optical sensors are limited by cloud cover – we develop a sensor-fusion approach. The algorithm applies a hidden Markov model (HMM) to simulate the snow wetness states the snow surface goes through, given the temporal observations of the surface conditions. The most likely current snow state is estimated, giving the current snow liquid water category. 

The snow products from SAR, optical and the multi-sensor approach are validated against cal/val sites providing frequent snow measurements in Romania and Norway, and additional field campaigns where a significant terrain relief is present providing corresponding significant gradients in snow wetness during the snowmelt season. Successful algorithms are implemented and demonstrated in a prototype system producing daily wet-snow maps of Romania and Norway. When the system is operationalised, the products will be used in operational hydrological models assisting flood prediction for issuing flood warnings. Similarly, the products will be used by the snow avalanche service issuing avalanche warnings.

Frazil and pancake ice (FPI) is becoming an important component of the Arctic cryosphere as a consequence of the dramatic decrease of sea ice extent and volume in the Arctic seas. In contrast, the role of FPI in energy exchanges between atmosphere and ocean is still poorly known. Thus, it would be advisable to develop methods for determining the extent and thickness of FPI. Remote sensing technology might contribute to achieve this task.

The aim of this study is to assess the capability of spaceborne SAR to provide information about FPI thickness. As no SAR proxy for FPI thickness exists, a way for its estimation is given by spectral analysis of SAR images, that is by exploitation of the dispersion and attenuation of waves which come from the open ocean and penetrate the FPI field [1].

We propose a novel SAR inversion scheme which improves the methodology described in [2] by taking advantage of the recent results obtained from the analysis carried on the unique FPI wave data set collected in the Weddell Sea [3], [4]. Such data drew a clear relationship between wave attenuation rates, α, and ice thickness. In this context, “ice thickness” is the “equivalent solid ice thickness” H=f_pV_ph_p + f_fV_fh_f, where h_p (h_f) is the thickness of pancakes (frazil); V_p (V_f) is the volume fraction in pancakes (frazil); and f_p (f_f) is the fractional area coverage of pancakes (frazil) [3], [4], [5]. A simple expression for α in terms of wave period in the range 7 s – 14 s and ice thickness was thus derived for application in research and operational models. Wave attenuation rates can be extended to the whole range of SAR wavenumbers by using a two-layer viscous model [6].­­

The SAR inversion procedure is based on knowledge of the open sea wind generated waves spectrum. It is essential that such waves come from the open ocean before travelling inside the FPI field. The mapping between ocean wave spectrum and SAR image spectrum is provided by the closed nonlinear integral transformation by Hasselmann and Hasselmann [7] and later developments which account for the SAR image cross-spectrum [8]. The wind wave spectrum is thus changed according to the rheological properties and to the thickness of the ice layer crossed in order to achieve the best fit between the observed SAR image spectrum and the simulated one. Finally, swell waves are estimated directly from the observed SAR image spectrum regardless their provenience.

The method is being applied on ERS SAR images acquired in the winter/spring 1997 in the Greenland Sea in coincidence with the in situ measurements taken during the oceanographic campaign carried on as part of the “ESOP-2” project (EU MAST-III programme). Pancakes were sampled in the Odden area, which represents the largest region of the Arctic oceans where vast fields of FPI have been observed so far. In addition, pancake and frazil thicknesses outputs from a salt flux model suitable for the Odden area are also used for comparison over extended areas [9].

 

References

[1] Wadhams P., V. A. Squire, J. A. Ewing and R. W. Pascal (1986), The effect of the marginal ice zone on the directional wave spectrum of the ocean, J. Phys. Oceanogr., 16, 358-376.

[2] Wadhams P., F. Parmiggiani, G. De Carolis, D. Desiderio and M. J. Doble (2004), SAR imaging of wave dispersion in Antarctic pancake ice and its use in measuring ice thickness. Geophys. Res. Lett., 31, L15305, doi:10.1029/ 2004GL020340.

[3] Doble M. J., M. D. Coon and P. Wadhams (2003), Pancake ice formation in the Weddell Sea, J. Geophys. Res., 108, C7, 3209, doi: 10.1029/2002JC001373.

[4] Doble M. J., G. De Carolis, M. H. Meylan, J.-R. Bidlot and P. Wadhams (2015), Relating wave attenuation to pancake ice thickness, using field measurements and model results, Geophys. Res. Lett., 42, doi: 10.1002/2015GL063628.

[5] Doble M. J. (2009), Simulating pancake and frazil ice growth in the Weddell Sea: A process model from freezing to consolidation, J. Geophys. Res., 114, C09003, doi:10.1029/2008JC004935.

[6] De Carolis G. and D. Desiderio (2002), Dispersion and attenuation of gravity waves in ice: A two-layer viscous fluid model with experimental data validation, Phys. Lett. A, 305, 399–412.

[7] Hasselmann K. and S. Hasselmann (1991), On the nonlinear mapping of an ocean wave spectrum into a synthetic aperture radar image spectrum and its inversion, J. Geophys. Res., vol. 96, no. C6, pp. 10713-10729, 1991.

[8] Engen G. and H. Johnsen (1995), SAR-ocean wave inversion using image cross spectra, IEEE Trans. Geosci. Remote Sensing, vol. 33, pp. 1047–1056, 1995.

[9] Toudal Pedersen L. and M. D. Coon (2004), A sea ice model for the marginal ice zone with an application to the Greenland Sea, J. Geophys. Res., 109, C03008, doi:10.1029/2003JC001827.

Increased temperatures over Greenland and other ice masses will likely lead to increased glacial melting, greater calving rates, and consequently to a rise in iceberg populations. However, it is unknown if there will be associated changes to the pattern of iceberg production and their trajectories. Majority of icebergs in the Northern seas calve from glaciers on the western coast of Greenland and are carried by the Greenland current into Baffin bay. Some become grounded and deteriorate, others drift along the western coast eventually reaching Grand Banks and the trans-Atlantic shipping lanes. Given these risks and uncertainties, monitoring and understanding the sources of the icebergs and ice-islands is essential for safe operations in these waters. Moreover, iceberg size, distribution and changing patterns are required within wide range of scientific contexts including impact of freshwater input to the ocean circulation and marine biology.

Number of methods have been developed to detect icebergs on operational basis often using Synthetic Aperture Radar (SAR) imagery such as RADARSAT-2 and in the past Envisat-ASAR. In 2014 new open data source has become available with the launch of ESA’s Sentinel-1a SAR mission. This study aims to investigate suitability of Sentinel-1 data for iceberg detection and tracking especially in regard to iceberg size. The 10 m spatial resolution of the Sentinel-1 interferometric wide swath mode is especially suitable for smaller Arctic icebergs that are difficult to detect with lower resolution imagery. Moreover, with the expected launch of the Sentinel-1b the revisit period will be reduced to close to one day, making the Sentinel-1 constellation data unparalleled data source for daily monitoring in the high seas.

The Constant False Alarm Rate (CFAR) method is commonly used for target detection with set false alarm rate under unknown sea clutter levels.  The Python algorithm developed in this study is based on the adapted CFAR run in iteration to avoid missing targets. The algorithm was tested under various sea state conditions and compared to results of existing operational detection algorithms and aerial in-situ observations provided by the International Ice Patrol. Both extra wide swath (EW) and interferometric wide swath (IW) mode products in both polarizations (HH, HV) were analysed, showing that icebergs down to size of 20m can be successfully detected. The feasibility of iceberg tracking and recognition between consequent images was tested to retract iceberg paths to the calving site and improve understanding of the iceberg movement in the Baffin Bay. Next possible step would be the use drift and decay models to address the gaps between image acquisitions. 

The final aim of the study was to ensure compatibility for eventual implementation of the developed automated iceberg detection algorithm within the ESA Polar Thematic Exploitation Platform (P-TEP). Examples of possible future use within the platform are generation of daily iceberg locations to provide iceberg warning but also to study changing patterns in iceberg size and distribution. 

As a major component of the landscape, lakes play an important role in the Arctic. The ecological and thermal processes in these lakes, including gas emissions (CH4, CO2) are strongly dependent on the seasonal freeze-thaw cycle (i.e. the lake ice phenology). The phenology of the lake ice is in large part controlled by climatic variables, in particular, air temperature has been determined to be the major controlling factor of break-up and freeze-up timing. Lake ice phenology is therefore considered as a robust indicator of both long term changes and short term variability in regional climate.

The aim of this research was to develop an automated method to derive the timing of start (BUS) and end (BUE) of the break-up period from optical remotely sensed data and quantify the changes occurring in lake ice break-up timing since the start of the millennium. The BUS and BUE dates were derived for over 13 000 lakes larger than 1km² in five study areas distributed evenly over the Arctic to capture the variation in the regional climate condition across the Arctic. 13 years of time series of daily surface reflectance data at 250m spatial resolution derived from the Moderate Resolution Imaging spectroradiometer (MODIS) was used to extract the lake ice break-up timing. The dates for the end of break-up (BUE) were validated against in-situ observations of Northern European Lakes.

The estimated BUE dates were strongly correlated with the in-situ observations (R2 0.65, RMSE 6.16). Statistical analysis performed on the results showed predominant shift towards earlier BUS ranging from average -0.10 days/year for Northern Europe to -1.05 days/year for area around Lake Taymir in northern Russia. Similarly the BUE has shifted in all study areas on average by -0.14  to -0.72 days/year. Number of lakes has showed trends significant on min 95% confidence interval with average ranging from -0.84 to -1.40 days/year for BUS and -0.52 to -1.10 days/year for BUE. Derived ice phenology was also related to various climatic and non-climatic factors such as daily air temperature, precipitation, snow depth, wind speed, and lake size. Deeper understanding of how various factors affect the timing of ice phenological events will help to predict the effect of ongoing climate change on the Arctic.

Space-based radio echo sounding (RES) of the continental ice sheets can potentially offer full coverage, uniform data quality and sampling.  Ice sounding radars must operate at low frequencies in order to ensure low attenuation of the signal as it propagates down through the ice and back from base of the ice sheet.  Typical frequencies of airborne radar ice sounders are between 60 MHz and 150 MHz.  However, the lowest possible frequency for space-based radar ice sounders is 435 MHz.  In 2004 the International Telecommunication Union (ITU) radio regulations allocated a 6 MHz band at 435 MHz (P-band) enabling space-based Earth observation radar missions at a frequency that might be applicable for ice sounding. The payload of ESA's Earth Explorer 7 mission, Biomass, is a P-band radar.

At P-band the attenuation and scattering properties of the ice sheets are not as well known as they are at the lower frequencies commonly used from aircraft, but in 2005 ESA commissioned development of a P-band polarimetric airborne radar ice sounder (POLARIS) [1], and encouraging results were obtained with data acquired in Greenland.  In February 2011 POLARIS data were acquired in Antarctica as part of a close scientific collaboration between seven organizations in Europe and North and South America [2].  The primary objective of this IceGrav campaign was to measure gravity in Queen Maud Land, but a secondary objective was to acquire P-band sounder data, benefitting from the large coverage offered by the Basler DC3 aircraft used.

In this study the feasibility of space-based radar ice sounding is assessed.  A two-step approach is applied:

(1) Key ice sheet parameters are estimated from the airborne POLARIS data acquired in Antarctica.

(2) The performance of potential space-based ice sounding radars is simulated based on the estimated ice parameters and system parameters envisioned for a space-based radar. 

The first step is accomplished by establishing empirical models of the attenuation coefficients and backscatter coefficients for the surface, volume and base of glaciers, ice shelves, central ice sheets etc..  The models are used in combination with the POLARIS system parameters and the data acquisition geometry to iteratively simulate return waveforms, compare them with the measured waveforms, and adjust the ice parameters until the simulated waveforms and the measured waveforms match.  

This iterative approach is supplemented by a direct data analysis estimating the scattering patterns via the Doppler spectra of the POLARIS data.  The scattering patterns of the ice surfaces are relevant because the geometry of a space-based radar increases the risk that off-nadir surface clutter masks the nadir depth-signal of interest.

Currently the ice sheet model is being established and validated.  At the symposium measured and simulated satellite waveforms will be compared, and the feasibility of space-based ice sounding will be addressed.

 

[1] J. Dall, S.S. Kristensen, V. Krozer, C.C. Hernández, J. Vidkjær, A. Kusk, J. Balling, N. Skou, S.S. Søbjærg, E.L. Christensen, “ESA’s polarimetric airborne radar ice sounder (POLARIS): Design and first results”, IET Radar, Sonar & Navigation, Vol. 4, No. 3, pp. 488-496, June 2010.

[2] J. Dall, A. Kusk, S.S. Kristensen, U. Nielsen, R. Forsberg, C-C. Lin, N. Gebert, T. Casal, M. Davidson, D. Bekaert, C. Buck, “P-band radar ice sounding in Antarctica”, Proceedings of the IEEE 2012 International Geoscience and Remote Sensing Symposium, pp. 1561-1564, Munich, July 2012.

The climate of the southern Canadian Arctic Archipelago has warmed rapidly over the past few decades, and regional mass losses from the glaciers of Baffin Island have approximately doubled between 1963-2006 and 2003-2011 (Gardner et al., 2012). This has coincided with rapid warming of the near-surface snowpack on Penny Ice Cap by ~10°C between the mid-1990s and 2011 (Zdanowicz et al., 2012). In this study we update mass loss rates for Penny Ice Cap to 2015 with NASA IceBridge altimetry data (2005, 2013, 2014, 2015), supplemented by ICESat (2003-2009) and CryoSat-2 (2010-2015) data, and verified with differential GPS (dGPS) measurements of ablation stakes. We adjust elevation changes determined from the altimetry data by removing densification and glacier dynamic components to obtain a true mass loss. A time varying firn densification model, which reflects the impact of recent warming on the snowpack, is validated with in-situ dGPS measurements and used to correct elevation changes for firn-rich parts of the ice cap (above ~1400 m). dGPS measurements collected along three transects are used to correct the ice-rich areas for glacier dynamics. Envisat ASAR data and in-situ Ground Penetrating Radar measurements are used to map the location of firn-rich and ice-rich regions. The total volume loss and contribution to sea level rise calculated from raw elevation change and adjusted data sets are compared. Together, these measurements provide the most comprehensive record to date of mass losses across Penny Ice Cap and their spatial and temporal variability.

Monitoring the sea ice extent and thickness over the Arctic region is of crucial importance for marine operational applications and climate studies. When processed with dedicated algorithms, satellite altimetry missions can provide highly valuable estimates of these two quantities. In this study, we will focus on lead identification, SLA computation and freeboard estimation from which the ice thickness estimates are derived.

In the last years, two new missions have complemented the altimeter constellation : Cryosat-2  embarking an innovative Delay Doppler altimeter in Ku band and Saral/AltiKa operating in Low Resolution Mode in Ka band. Both missions have already demonstrated their ability to provide enhanced performances compared to the historical Ku band LRM missions with improved range precision and higher spatial resolution. We propose in this talk to illustrate the benefits that can be achieved through the use of these combined measurements over the Arctic sea ice.

The estimation of the parameters of interest (freeboard, sea height, ice extent, ice thickness, ice volume, ice mass ...)  in the Arctic is carried out in two steps. The first stage consists in identifying the type of surface that is beneath the satellite, discriminating precisely leads, water surfaces, polynya from ice floes. We will highlight the good consistency of the results derived from classification of the Pseudo-LRM echoes, of the multi looked Delay Doppler power echoes (that are obtained by averaging beam waveforms in the stack along the azimuth direction) and of distributions of power across the looks in the stack (Range Integrated Power).

The second step consists in estimating the height of each observed surface, from which the freeboard can be deduced by difference. Heights of the water surface and heights of the floes have been historically computed using empirical retrackers.  We propose here to analyze the results of a new estimation method based on a physical model accounting for different mean square slopes of the surface (from rough to very specular surfaces) and including instrumental characteristics (Point Target Response and antenna pattern). The same range estimation method will be applied on both LRM and SAR measurements and the results compared. Synergy between Ku (SAR/RDSAR) and Ka  (LRM) results will be evaluated and its exploitation discussed.

Nansen International Environmental and Remote Sensing Centre in Saint-Petersburg and Nansen Environmental and Remote Sensing Centre in Bergen for a long time have been working on development of automated sea ice classification methods for the synthetic aperture radar (SAR) images. These algorithms have been developed for ENVISAT and Radarsat-2 data. Since 2014, new satellite Sentinel-1 has started to work on the orbit. Adopted classification algorithm for Sentinel-1 based on the support vector machine (SVM) technique uses the Extra Wide (EW) swath dual-polarization mode data, i.e. HH (horizontally transmitted and horizontally received) and HV (horizontally transmitted, vertically received). EW mode acquires wide SAR image of 400 x 400 km with 40 m resolution resulting from several narrower SAR beams. The raw data in HH and HV polarizations are first corrected for angular dependence, thermal and random noise and smoothing the image with several filters. At the next step texture characteristics are computed using a Gray Level Co-occurence Matrix (GLCM). This algorithm has an additional step - the principal component analysis of textural features with further clustering results. The analysis and training of the algorithm by the sea ice expert had been carried out based on automatically obtained images with clusters.  For the ice/water retrieving algorithm SVM was trained on results of visual analysis of 20 Sentinel-1 images. The next step is performing the automated sea ice types classification. Using the developed technique the open water area is masked, and sea ice area can be classified into several sea ice types. Training Sentinel-1 SAR data were collected for ice classification into categories such as open water, broken ice on the ice \ water edge, leads, thin first year ice, deformed first year ice and multiyear ice. Main results are: - version of the algorithm, which gives satisfactory results for automatic separation of ice-water was developed and tested; - on the obtained maps with classification results large inhomogeneities such as leads and cracks can be distinguished; - test version of the ice types classification algorithm was developed. Despite the general high accuracy of results there are still some problems with classification of the rough water areas on some images due to the similar average backscatter and texture as young newly formed ice. Development and testing of a new classification algorithm is performed in the framework of NORRUS project conducted jointly by the Russian Foundation for Basic Researchand the Research Council of Norway: "Development of the system of monitoring and forecasting of sea ice to support the safe operation and navigation in the Arctic seas".

The last decades the ice sheets of Greenland and Antarctica have undergone changes in their volume due to alteration in the climate. In Greenland, rises in air temperature have led to an increase in ice sheet surface melting, and ice flow has fluctuated in response to changing ocean and atmospheric conditions. The mass loss from 1992 to 2011 was approximately 2700±930 Gt of ice in Greenland and 1350±1010 Gt of ice in the Antarctic continent and the average contribution of the two continents to the sea level rise from 1993 to 2010 was around 0.60 mm yr^-1. Thus, monitoring their changes is essential to assess their on-going impact upon society. Aiming to improve the understanding of the main instability processes occurring in the polar ice sheets, interferometric techniques using satellite synthetic aperture radar (SAR) images have being developed since the 1990s to detect displacements in the glacier’s surface. Most recently, Sentinel-1a was launched in 2014 and offers the opportunity for continued monitoring of the evolution of glacier flow. The 12 day repeat period of images provided by this satellite are helping us to understand the timescales over which glaciers evolve and enable better projections of future contribution of ice sheets to sea level rise to be made. This study provides analysis of Sentinel-1a data over Western Greenland and the Southern Antarctic Peninsula ice sheets. We estimated surface glacier velocities using an intensity tracking algorithm based on single-look complex (SLC) Sentinel-1a (S1a) Interferometric Wide swath (IW) mode data acquired from 2014 to the present. The high temporal resolution glacier's surface velocities maps generated, with 12 day repeat period images, have been compared to consolidated databases such as the National Snow and Ice Data Center.

As NASA and the French Space Agency (CNES) plan the Surface Water and Ocean Topography mission (SWOT), there has been much interest among surface water hydrologists in using remote sensing to monitor river discharge. During its three-year mission, SWOT will retrieve river heights with a temporal sampling of approximately weekly (better at higher latitudes). Meanwhile, flow area within a defined river reach (a "satellite gauging site") is also a direct indicator of discharge. River Watch Version 2 at the University of Colorado (http://floodobservatory.colorado.edu/DischargeAccess.html) uses microwave radiometry pre-processed at the European Joint Research Centre, from NASA’s TRMM, AMSR-E, and GPM, and the Japanese Space Agency’s AMSR-2, to monitor flow area changes along an expanding suite of gauging reaches, including those in the high latitudes. Data are updated daily. Below 50 degrees latitude, the daily river flow area signal extends continuously back to 1998; at higher latitudes, data begin with AMSR-E launch in mid-2002. A global hydrologic model (for five years, with daily time steps; WBM) is used to calibrate the flow areas to discharge information.

A warming Arctic region is experiencing intensification of the hydrological cycle, with increasing permafrost and glacial melt and possibly more precipitation resulting in higher river runoff. A significant increase of nearly 10% in annual river flux has been observed via ground stations in 13 major rivers throughout the Arctic over the last 30 years. However, direct measurements are sparse for 100’s of smaller-scale rivers, and measurements even for the large rivers are constrained by seasonal ice coverage, break-up and freeze-up dynamics, difficult access, and interruptions of the flow record. To overcome such, we develop remote sensing-based river discharge measurement techniques using a variety of satellite sensors and as adapted to these Arctic rivers. Thus, AMSR-E/AMSR-2 microwave radiometry can provide consistent determination of ice cover removal in the spring, and establishment of complete ice cover in the fall. At a site along the Pechora River, Russian Republic (http://floodobservatory.colorado.edu/SiteDisplays/100158.htm), the microwave radiometry-determined ice free season ranges from 166 to 208 days (2002-2014), and the peak discharge from 4700 m³/sec to 23200 m³/sec. Interestingly, the well-known floods of 2007-2008 in the high Arctic in this region did not occur during times of unusually long observed ice-free durations or high peak discharges at this site. Instead, satellite-determined total runoff volume, expressed here as depth over the 247665 km² drainage area, varied in the same period from 133 to 333 mm; values for 2007 and 2008 were 273 and 333 mm respectively, and 2005 was also unusually high at 333 mm. 2005, 2007, and 2008 experienced the three highest runoff volumes for the period of record. The high influx into the Arctic Ocean from at least this river during 2007-2008 was from sustained high summer discharge.

The opportunity to monitor in fully automated mode, via satellite sensors, the ice-free season along all major Arctic rivers and to directly measure the summer hydrographs provides a consistent method to quantify the meteorological causation of and the quantities of freshwater discharge into the Arctic Ocean. The implications with regard to land-fast coastal ice dynamics and inter-annual extent can thereby also be explored. Finally, to validate the microwave results for river ice cover and break-up, our team uses a variety of other sensors, including MSI on Sentinel-2 and MODIS on Aqua and Terra; all provide higher resolution information about the dynamics of ice removal, the spring flood, and the late summer decline into baseflow conditions and ice cover establishment. Regarding synthetic aperture radar (SAR) data, Sentinel-1 SAR data is used to determine river ice cover, independent of darkness and cloud cover, and during the summer provides critical validation information for site surface water extent.

In the cryosphere of the Central section of the Andes Cordillera, rock and covered glaciers [1] are both inland freshwater resources and good indicators of regional climate variability [2]–[4]. Those landforms, tongue or lobed shaped, are developed near at high mountain slopes within permafrost areas [5]. They are dynamic systems with varying activity, where rock fragments are mixed with ice in different proportions, consisting in a permanently frozen mixed ice-debris core and a top layer suffering seasonal thawing ("active layer") [6]–[8]. Theoretically, the variation in its topography would be motivated by the seasonal loss and recovery of its ice content, and slope-down gravitational action too. Verticals and lateral movements of the detrital surfaces indicate the evolution of the mass of underlying or intrinsic ice [9]. In practice, the identification and analysis of some characteristics of these cryoforms can be performed using remote sensing data. These tools complement and expand the glaciological and/or geomorphological field works.

In this paper, we present a study of rock and covered glaciers located in the San Juan Province, within the Central Andes of Argentina. The main objective of this work is to identify and characterize the cryoforms present in the area in order to obtain a more comprehensive idea of the dynamics of this type of glaciers and their impact in the Central Andean environment.

Differential SAR Interferometry (DInSAR) is an extensively and suitable tool to provide ground deformation measurements [10]. However, high deformation rates commonly produced within the glaciers make difficult to use it and satellites with short revisit time become mandatory in order to assess the displacements. Glaciers within the area of interest are mainly rock and covered glaciers characterized by a deformation rate between 0.5 to 1 m/y [11], what represents an advantage in favor of the applicability of DInSAR.

When large displacement are expected, an alternative is to employ a SAR amplitude signal based technique (i.e. Pixel Offset – PO) [12], [13]. The main drawback of PO is related to its accuracy, what is one order of magnitude worst at least than for DInSAR. Availability of high-resolution SAR systems is an improvement, bringing the precision of the obtained results near to that attainable with DInSAR [14].

We analyze the area of interest denoted in Figure 1, following a two steps procedure. First, we analyze a set of differential interferograms computed from two Cosmo-Skymed HImage SAR dataset (ascending and descending pass), allowing us to identify the active cryoforms and their activity condition. Second, we complementary evaluate the extent and velocity rate of those most active cryoforms by using PO technique. Both approach allow us to characterize the rock and covered glaciers in depth.


[1]    IANIGLA, “Inventario Nacional de Glaciares y Ambiente Periglacial: Fundamentos y Cronograma de Ejecución.” Instituto Argentino de Nivología, Glaciología y Ciencias Ambientales, 2010.
[2]    A. Rabatel, H. Castebrunet, V. Favier, L. Nicholson, and C. Kinnard, “Glacier changes in the Pascua-Lama region, Chilean Andes (29° S): recent mass balance and 50 yr surface area variations,” The Cryosphere, vol. 5, no. 4, pp. 1029–1041, Nov. 2011.
[3]    G. F. Azócar and A. Brenning, “Hydrological and geomorphological significance of rock glaciers in the dry Andes, Chile (27°-33°S),” Permafr. Periglac. Process., vol. 21, no. 1, pp. 42–53, Jan. 2010.
[4]    A. Brenning, “Geomorphological, hydrological and climatic significance of rock glaciers in the Andes of Central Chile (33-35°S),” Permafr. Periglac. Process., vol. 16, no. 3, pp. 231–240, Jul. 2005.
[5]    J. P. Milana and A. Güell, “Diferencias mecánicas e hídricas del permafrost en glaciares de rocas glaciánicos y criogénicos, obtenidas de datos sísmicos en El Tapado, Chile,” Rev. Asoc. Geológica Argent., vol. 63, pp. 310 – 325, 2008.
[6]    F. A. Croce and J. P. Milana, “Internal structure and behaviour of a rock glacier in the Arid Andes of Argentina,” Permafr. Periglac. Process., vol. 13, no. 4, pp. 289–299, Oct. 2002.
[7]    F. Croce and J. P. Milana, “Electrical Tomography applied to image the 3D extent of the permafrost of three different Rock Glaciers of the Arid Andes of Argentina,” in Geophysical Research Abstracts, 2006, vol. 8, p. 03026.
[8]    A. Kääb and M. Weber, “Development of transverse ridges on rock glaciers: field measurements and laboratory experiments,” Permafr. Periglac. Process., vol. 15, no. 4, pp. 379–391, Oct. 2004.
[9]    L. Liu, C. I. Millar, R. D. Westfall, and H. A. Zebker, “Surface motion of active rock glaciers in the Sierra Nevada, California, USA: inventory and a case study using InSAR,” The Cryosphere, vol. 7, no. 4, pp. 1109–1119, Jul. 2013.
[10]    A. K. Gabriel, R. M. Goldstein, and H. A. Zebker, “Mapping small elevation changes over large areas - Differential radar interferometry,” J. Geophys. Res., pp. 9183–919, Jul. 1989.
[11]    C. Harris, “The nature and dynamics of mountain permafrost: introduction,” Permafr. Periglac. Process., vol. 15, no. 3, pp. 189–189, Jul. 2004.
[12]    T. Strozzi, A. Luckman, T. Murray, U. Wegmuller, and C. L. Werner, “Glacier motion estimation using SAR offset-tracking procedures,” IEEE Trans. Geosci. Remote Sens., vol. 40, no. 11, pp. 2384–2391, Nov. 2002.
[13]    F. Casu, A. Manconi, A. Pepe, and R. Lanari, “Deformation Time-Series Generation in Areas Characterized by Large Displacement Dynamics: The SAR Amplitude Pixel-Offset SBAS Technique,” IEEE Trans. Geosci. Remote Sens., vol. 49, no. 7, pp. 2752 –2763, Jul. 2011.
[14]    N. Riveros, L. Euillades, P. Euillades, S. Moreiras, and S. Balbarani, “Offset tracking procedure applied to high resolution SAR data on Viedma Glacier, Patagonian Andes, Argentina,” Adv. Geosci., vol. 35, pp. 7–13, Jun. 2013.

Surface elevation measurements from CryoSat-2 data were examined to determine their utility for measuring ice sheet grounding line locations and ice thickness in Antarctica. The boundary between grounded and floating ice is an important glaciological parameter, because it delineates the lateral extent of an ice sheet and it marks the optimal location for computing ice discharge. We present a method for detecting the grounding line as the break in ice sheet surface slope, computed from CryoSat-2 elevation measurements using a plane-fitting solution. Furthermore we measure ice thickness at the grounding line using firn corrected CryoSat-2 data based on the theory of hydrostatic equilibrium.  We apply these techniques to map the break in surface slope and ice shelf thickness at the grounding line in four topographically diverse sectors of Antarctica - the Filchner-Ronne ice shelf, the Ekström ice shelf, the Amundsen Sea Sector, and the Larsen-C ice shelf - using CryoSat-2 observations acquired between July 2010 and May 2014. An inter-comparison of the CryoSat-2 break in surface slope with independent measurements of the hinge line position determined from quadruple-difference synthetic aperture radar interferometry (QDInSAR) shows good overall agreement, with a mean separation of 4.5 km. In the Amundsen Sea Sector, where in places over 35 km of hinge line retreat has occurred since 1992. The CryoSat-2 break in surface slope coincides with the most recent hinge line position, recorded in 2011. Ice shelf ice thickness measurements are validated with Radio Echo Sounding (RES) point data and show good overall agreement with BEDMAP 2 ice thickness data. The techniques we have developed are automatic, computationally-efficient, and can be repeated in the future given further data acquisitions offering a complimentary approach to existing techniques.

Polar regions and the cryosphere constitute  a unique, valuable, yet extremely sensitive component of the Earth system. Obtaining robust, routine information products for better scientific understanding and development of applications in these last remaining frontiers of our planet requires collective action. 

In the 50 years between the International Geophysical Year in 1957 and the 4th International Polar Year (IPY) in 2007-2009, the space era transformed our capability to study the cryosphere. As the emerging economic opportunities begin to dictate interests in these fragile regions it becomes imperative to assemble scientific knowledge with which to exercise responsible management and stewardship. Space Agencies who participated in IPY in a federated effort to collect an integrated set of unique snapshots of the polar regions have already contributed to an important data legacy. The next step is to harness this experience to build an integrated observing system and monitoring capability for the cryosphere in general. The vision for this system contains a space-borne  element that provides data product as essential input to  informed decision making at national and international level.

The Polar Space Task Group (PSTG), formed after the success of IPY operates under the auspices of the World Meteorological Organization (WMO) Executive Council Panel of Experts on Polar and High Mountain Observations Research and Services (EC-PHORS). It comprises nominated representatives of Space Agencies and seeks to coordinate the efforts of space-faring nations in the joint endeavour of providing scientific information about the polar regions and the cryosphere in support of science and applications. PSTG activities contribute to the Global Cryosphere Watch project, the Polar Activities of WMO, and to the strategic goals of the respective Member Agencies.

Based on its transition from the successful IPY-focused group, the IPY-STG, the PSTG has developed a series of coordinated actions and a strategic plan which supports monitoring the polar regions and cryosphere. Here we present unique examples of the benefits of PSTG interagency coordination, spanning the respective thematic areas covered by the group. These range from monitoring ice sheets, glaciers and ice caps; floating ice (sea, lake and river ice); permafrost, snow, and the atmosphere (in respect to improving polar prediction capability).

Through these efforts, and the commitment of the respective Member Space agencies, the  cryospheric science community has access to a legacy dataset comprising multiple space agency satellite data portfolios acquired by a large variety different instruments.

Satellite synthetic aperture radar (SAR) data are used by agencies tasked with discriminating sea ice types and defining ice conditions in support of safe voyages in ice-prone waters. SAR based sea ice information is also used to support scientific research and the activities of people in northern communities. During summer the reliability of SAR data is severely compromised by melt water impacts on SAR measured microwave backscatter. Challenges include developing a better understanding of summer sea ice physical properties and microwave backscatter at critical scales, and developing new SAR technologies which overcome melt related ambiguities and offer observational stability. The goal of this research is to advance the utility of C-band and L-band frequency SAR, for discriminating sea ice types and geophysical information during the the advanced melt period.  

In this paper, new results from a 2015 field and image acquisition campaign in the Victoria Strait section of the Canadian Arctic Archipelago are examined. During this campaign a suite of RADARSAT-2 (C-band) and ALOS-2 PALSAR-2 (L-band) SAR images comprising a mixture of landfast sea ice types were acquired during pre-melt (winter) and summer periods. Coincident sea ice thickness and ice surface roughness data were acquired using a suite of helicopter-borne sensors flown over the study site during the pre-melt period. Ice thickness and roughness data were used to identify contiguous zones of first-year ice (FYI) and multiyear ice (MYI) types, from which baseline polarimetric backscatter information was characterized. After the onset of advanced melt, demarcated by the occurrence of surface melt ponds on the sea ice, a set of coincident high-resolution (50 cm pixel) visible-near infrared images of FYI and MYI were acquired from the World-View 2 satellite. Optical scenes were partitioned into relative fractions of ice, open water, and melt pond. The pre-melt to advanced melt changes in polarimetric backscatter parameters according to ice type, as well as the contribution of sea ice melt pond fraction to variations in backscatter parameters, are detailed.

Soil moisture is an essential climate variable that affects weather and climate and is involved in water and energy cycles between land and atmosphere. It is important for understanding land surface processes and surface energy balance. Soil moisture is also an important variable for both climate and numerical weather prediction models. Soil freeze limits growing season and thus affects carbon balance, and it also has an effect on hydrologic cycle by preventing water to enter soil in frozen state.

 

The retrieval method presented is based on an inversion technique and applies a semiempirical backscattering model that describes the dependence of radar backscattering of forest as a function of stem volume, soil moisture or permittivity, vegetation canopy moisture, surface roughness and incidence angle. It gives an estimate of soil permittivity or relative soil moisture using active microwave measurements. Applying a Bayesian assimilation scheme, it is also possible to use other soil permittivity or soil moisture retrievals to regulate this estimate to combine for example low resolution passive observations with high resolution active observations for a synergistic retrieval. This way the higher variance in the active retrieval can be constricted with the passive retrieval when at the same time the spatial resolution of the product is improved compared to the passive-only retrieval. The retrieved soil permittivity or soil moisture estimate can also be used to detect soil freeze/thaw state by considering the soil to be frozen when the estimate is below a threshold value.

 

The method was tested using SAR (Synthetic Aperture Radar) measurements from ENVISAT ASAR instrument for the years 2010-2012 and from Sentinel-1 satellite for the year 2015 in Sodankylä area in Northern Finland. The synergistic method was tested combining the SAR measurements with a SMOS (Soil Moisture Ocean Salinity) radiometer based retrieval. The results were validated using in situ measurements from automatic soil state observation stations in Sodankylä calibration and validation (CAL-VAL) site, which is a reference site for several EO (Earth Observation) data products.

In this paper we present scientific investigations of both ice and snow based on high resolution Synthetic Aperture Radar (SAR) Tomography (TomoSAR).

The objective of these measurements was to characterize the vertical structure or Radar scattering from snowpack and ice through a three dimensional (3D) reconstruction, which is obtained by focusing in the 3D space multiple SAR acquisitions gathered from slightly different points of view. This approach has been largely considered in recent years for forestry applications as it entails a fundamental advantage over traditional (i.e.: 2D) SAR imaging, namely the possibility to see the vertical structure of the imaged volume, to be afterwards employed as a robust basis for validation and development of physical models.

The instrumentation was the Ground-Based Synthetic Aperture Radar (GB-SAR) developed by the SAPHIR team at IETR, University of Rennes I. Signal transmission and reception are controlled by a Vector Network Analyzer (VNA) able to operate from 3 GHz to 35 GHz (C,X,Ku,Ka bands). The VNA supports an antenna array, which can be manually shifted to increase the vertical aperture in repeat-pass mode. A synthetic aperture along the horizontal direction is obtained by moving the VNA and the antenna array along a 3 m rail. The data are focused in the 3D space via Time Domain Backprojection, resulting in a 3D reconstruction of the scattering elements within the snowpack.

Snowpack features were investigated in the frame of the ESA campaign AlpSAR, by illuminating a 70 cm snowpack in the valley of Leutasch, north of Innsbruck, at an altitude of about 1150 m and a 140 cm snowpack in the valley of Rotmoos, close to the Italian border, at an altitude of about 2300 m. Data over sea ice were gathered at Kattfjord, Tromsø, Norway, during a campaign by the University of Rennes 1 and the University of Tromsø.

For both snowpack and sea ice the images produced by the GBSAR revealed a multi-layered structure. In many cases the backscattered from the bottom layers was observed to dominate the one from the surface and near subsurface by over 20 dB. This result was found to be largely independent of the incidence angle. GBSAR images also turned out to provide sensitivity to propagation velocity within the snowpack, as revealed by the apparent depth variation with respect to the incidence angle.

This effect was exploited to retrieve wave velocity directly from the data, by implementing an iterative procedure for the retrieval of wave propagation velocity in different layers. The data were afterwards refocused by assuming the estimated wave velocity. The resulting vertical structures were found to be consistent with in-situ data obtained from coincident snow-pit measurements.

We will present an update on the development of an automated Microwave/Imager Sea Ice Classifier (MISIC) algorithm that integrates satellite imagery from optical instruments onboard polar orbiting platforms, such as the Moderate Resolution Imaging Spectroradiometer (MODIS) from the Terra and Aqua satellites and the Visible Infrared Imaging Radiometer Suite (VIIRS) onboard Suomi National Polar-orbiting Partnership (S-NPP), with data from passive microwave satellite sensors such as AMSR-2 onboard GCOM-W1. The intent is to develop an algorithm suitable for real-time processing that is capable of producing a high spatial resolution sea ice product that would provide spatially continuous (cloud gap free) characterization of the ice extent and the ice edge at maximum possible spatial resolution.

To meet these objectives, we have developed an algorithm that first extracts maximum information on the sea ice cover from imaging instruments VIIRS and MODIS, including regions covered by thin, semitransparent clouds. This ability to identify ice cover underneath thin clouds, which is usually masked out by traditional cloud detection algorithms, allows for expansion of the effective coverage of the sea ice maps and thus more accurate and detailed delineation of the ice edge. In the next step of our algorithm, the labels obtained from individual imager overpass observations are supplemented by the microwave measurements, resulting in 5 intermediate classes: sea ice, ice-free water, ice under thick clouds, water under thick clouds and an undetermined class where there is insufficient information. For example, grid cells under cloud near the ice edge, cannot be precisely determined by the lower resolution microwave ice concentration product. The confidently labeled grid cells serve as a training set for a Linear Discriminant Analysis (LDA) classifier in an experimentally derived feature space constructed from reflective, shortwave IR and thermal bands centered around 0.46nm, 0.64nm, 0.86nm, 1.24nm, 1.6nm, 11nm and 12nm from the imager instrument as well as polarization and gradient ratios of 18 and 36 GHz radiometer channels of the microwave instrument. The microwave data makes it possible to classify grid cells under thick clouds that cannot be classified using the imager input alone. Although this particular step of the algorithm relies on the classification involving microwave data and therefore results in a somewhat degraded effective spatial resolution of the ice extent, these grid cells have a special “microwave” tag and weighted less in the last step of the algorithm which combines the labels from the individual overpasses in an optimal way into daily ice extent product, resulting in the enhanced, high resolution, gap free ice extent product.    

We have performed a validation of our integrated ice maps using high-resolution Landsat and Sentinel-2 imagery as well as active SAR measurements from RadarSat. We have also developed a web-based monitoring system that allows potential users to compare the performance of our daily ice extent product with the several other independent operational daily products. 

The paper presents the generalization of the results of the author’s long-term studies on the substantiation of the design works of building the underground structures of a different purpose in the permafrost zone workings. The aim of the above studies is to prove the possibility and energy and economic expediency of the complex use of permafrost zone workings, operating or laying-up mining facilities (underground mines, pits) as well as new ones including those not associated with mining (storehouses, refrigerators, collectors).

The idea of the complex use of workings implies the inclusion of specially built or some abandoned workings of the double purpose in the general system of ventilation of an underground structure. That is, these workings can be used as heat exchange facilities (minimizing expenses for keeping the standard parameters of the microclimate) during structure operation according to its designated purpose, as well as the protection structure of the civil defense for working shifts and population during some emergency situations of a natural or technogenic character.

Complex theoretical and experimental studies have been performed including full-scale tests at operating mines of the North. The results obtained present a new conception of developing the underground space of the permafrost zone. The above conception is based on the control of the processes according to the criterion of energy saving and complex use of workings.

Particularly, theoretical principles of calculating the systems of the control of thermal conditions on the basis of heat exchange workings are developed. Three classes of the above systems are studied: general, recuperative and regenerative ones. Optimum mathematical models of heat exchange in workings are used that allows not only to predict the microclimate parameters but choose the optimum (from the energy and economical point of view) technical and technological parameters of the systems at different operating conditions.

The new methods of supporting the workings with the use of the sprayed-concrete heat-protection multifunctional load-carrying coverings are developed. The above methods have been scientifically substantiated and subjected to the experimental tests. The presence of such coverings in combination with the modulus principle of designing the underground structures allows to use the double purpose workings both during normal and emergency period of operation.

The methodical principles of the choice of the layout and arrangement of underground structures by a heat factor as well as optimum conditions of ventilation during building and operation, that provide the necessary safety level at minimum energy and material expenses for obtaining the standard microclimate parameters, are developed.

         

Sea-ice strongly influences global climate through the insulation of ocean from atmosphere, the ice-albedo feedback mechanisms, and the ocean circulation modification. Its extensions have been well observed since early 80’s by satellites. However, two major dimensions are still far imprecisely known for the global climate models: the variations of sea ice volumes and the sea level over Arctic ocean, both of them being identified as Essential Climate Variables (ECV) by ESA-CCI and GCOS.

Only the satellite altimeters are able to catch this 3rd dimension at a pan-Arctic level, and, among these altimeters, only Cryosat2 has the ability to fly up to the higher latitudes. Moreover, Cryosat2 is the first altimeter that uses the SAR technology, which will be also employed on Sentiel-3 and Sentinel-6.

The CryoSat2 altimetric mission has already demonstrated its capacity to estimate sea-ice freeboard and thickness. Nowadays, the new challenges are to better understand the uncertainties and to potentially improve the accuracy of these estimations.

Unlike previous LRM altimeters, in CryoSat2 the FBR echoes are downloaded and the Doppler beams and Multilook echos are processed off-line. This offers an unique opportunity to master the chain process and to adapt it according to the type of the observe surfaces. For instance, the Doppler beams ground position can be chosen; some filtering may be added in order to avoid the secondary lob effects; or even the peaky waveforms can be re-interpolated using the FBR data …

Several products, using different signal treatment models or parameters, have been developed: ESA baselines B and C, SAMOSA2, SAMOSA+, CPPv14. In this study we evaluate the impact of these choices and assess the importance of different parameterisation of physical radar models of retrackers in different sea ice conditions (including smooth first-year ice, rough multi-year ice with low/high snow depth, fast ice etc.).

The approach consists in use of CryoSat2 and Sentinel-1 missions over-passes for local scale analysis and in further extrapolation of the results for the whole pan-arctic domain.

This study should help to assess for the first Sentinel-3 measurements over sea ice.

Acknowledgment: This work is supported by the CNES through the TOSCA SICKAyS project.

In order to quantify the effects of absorbing material on snow and define its contribution to climate change, we have conducted a series of dedicated bidirectional reflectance measurements. Chimney soot, volcanic sand, and glaciogenic silt were deposited on snow in a controlled way [1, 2]. The bidirectional reflectance factors of these targets and untouched snow have been measured using the Finnish Geodetic Institute's field goniospectrometer FIGIFIGO, see, e.g., [3, 4] and references therein. It has been found that the contaminants darken the snow, and modify its appearance mostly as expected, with a clear directional signal and a modest spectral signal. A remarkable feature is the fact that any absorbing contaminant on snow enhances the metamorphism under strong sunlight. Immediately after deposition, the contaminated snow surface appears darker than the pure snow in all viewing directions, but the heated soot particles sink down into the snow within minutes. The nadir measurement remains darkest, but at larger zenith angles the surface of the soot-contaminated snow changes back to almost as white as clean snow. Thus, for an observer on the ground, the darkening by impurities can be completely invisible, overestimating the albedo, but a nadir looking satellite sees the darkest points, now underestimating the albedo.

After more time, also the nadir view brightens, and the remaining impurities may be biased towards more shadowed locations or less absorbing orientations by natural selection. This suggests that at noon the albedo should be lower than in the morning or afternoon. When sunlight stimulates more sinking than melting, albedo should be higher in the afternoon than in the morning, and vice versa when melting is dominating. Thus to estimate the effect on climate change caused by light-absorbing impurities deposited on snow, one need to take into account possible rapid sinking of the impurities inside the snow from its surface. When the snow melt rate gets faster than the sinking rate (under condition of warm outside temperatures), as was observed at the end of the experiment reported here, dark material starts accumulating into the surface. Light-absorbing impurities deposited on snow at warm temperatures initiate rapid melting and may cause dramatic changes on the snow surface.

 

[1] Svensson, J., Virkkula, A., Meinander, O., Kivekäs, N., Hannula, H.-R., Järvinen, O., Peltoniemi, J. I., Gritsevich, M., Heikkilä, A., Kontu, A., Hyvärinen, A.-P., Neitola, K., Brus, D., Dagsson-Waldhauserova, P., Anttila, K., Hakala, T., Kaartinen, H., Vehkamäki, M., de Leeuw, G., and Lihavainen, H. (2015): Soot on snow experiments: light-absorbing impurities effect on the natural snowpack. The Cryosphere Discuss. 9, 1227-1267, http://dx.doi.org/10.5194/tcd-9-1227-2015

[2] Peltoniemi J.I., Gritsevich M., Hakala T., Dagsson-Waldhauserová P., Arnalds Ó., Anttila K., Hannula H.-R., Kivekäs N., Lihavainen H., Meinander O., Svensson J., Virkkula A., de Leeuw G. (2015): Soot on snow experiment: bidirectional reflectance factor measurements of contaminated snow // The Cryosphere Discuss., 9, 3075-3111, http://dx.doi.org/10.5194/tcd-9-3075-2015

[3] Peltoniemi J.I., Hakala T., Suomalainen J., Honkavaara E., Markelin L., Gritsevich M., Eskelinen J., Jaanson P., Ikonen E. (2014): Technical notes: A detailed study for the provision of measurement uncertainty and traceability for goniospectrometers. Journal of Quantitative Spectroscopy & Radiative Transfer 146, 376-390, http://dx.doi.org/10.1016/j.jqsrt.2014.04.011

[4] Peltoniemi J.I., Gritsevich M., Puttonen E. (2015): Reflectance and polarization characteristics of various vegetation types. Springer Praxis Books, Light Scattering Reviews 9, pp. 257-294, http://dx.doi.org/10.1007/978-3-642-37985-7_7

Detection of disturbed snow (principally, avalanche and snowmobile tracks) has been discussed as a significant aspect within the search and rescue (SAR) community (Cooper 2005, Buhlet et al. 2009). The fundamental advantage of detecting the most recently disturbed snow is the potential reduction in time that can be gained by identifying the last known path of travel for the individual(s).

 

            With the objectives of SAR in mind, the aims of this study are two-fold: (1) provide a measure of the sensitivity of NIR-SWIR spectral imaging data to recently vs. non-recently disturbed snow; and (2) assess the expected quality of using applicable satellite data (Sentinel-2 and Landsat 8) as a platform to aid SAR investigations.

 

            To investigate the application of hyperspectral data to recently disturbed snow, a simple methodology has been developed that is based upon suspected spectral differences of the re-crystalized disturbed snow compared to the surrounding un-disturbed snow.  Specifically, in the water absorption regions (810-835nm, 920-945nm, 1110-1140nm, 1400nm-1600nm and the 1850nm to 2100nm windows) data from previous airborne hyperspectral and ground campaigns have provided preliminary results showing that 1) the phase transition of water as a function of temperature exhibits a shift in the water absorption trough within the water absorption regions and, 2) within the 1400-1600nm region, a significant and repeatable result in the imagery shows the difference between disturbed/ un-disturbed and recently disturbed snow.  It is hypothesised in this study, that these differences may be the result of the re-crystallization effects that occur due to the disturbance itself – pressure induced partial melting of the snow crystal structure. 

 

The current study will focus on the results of a new site that is located in Ottawa, Canada and will include large areas of un-disturbed snow as well as well-controlled recent and non-recently disturbed snow.  The disturbances will be caused by both personnel and vehicles.     Our primary task in this work is to assess the potential to use satellite imagers as SAR tools by upscaling the ground and airborne results to the level of spaceborne vehicles.  By planning our airborne and ground campaign to be temporally coincident with Landsat 8 and Sentinel-2 overflights, during the winter of 2016, we will have the necessary data to provide a real (non-computer generated simulation) assessment of using current state-of-the-art satellite data for SAR operations.

 

References

D.C. Cooper, Fundamentals of Search and Rescue, Jones and Bartlett, Sudbury, Massachusetts, 2005.

Y. Bühler, A. Hüni, M. Christen, R. Meister, T. Kellenberger, “Automated detection and mapping of avalanche deposits using airborne optical remote sensing data”, Cold Regions Science and Technology 57, pp. 99–106, 2009

 

 

The main objective of this work is to validate CryoSat-2 (CS2) SARIn performance over sea ice by use of airborne laser altimetry data obtained along a CS2 ground track in the Wingham Box during the CryoVEx 2012 campaign. A study by Armitage and Davidson [2014] has shown that the extra information from the CS2 SARIn mode increases the number of valid sea surface height estimates which are usually discarded in the SAR mode due to snagging of the radar signal. As the number of valid detected leads increases, the uncertainty of the freeboard heights decreases.

In this study, the snow freeboard heights estimated using data from the airborne laser scanner (ALS) are used to validate the sea ice freeboard obtained by processing CS2 SARIn level 1b waveforms. The possible reduction in the random freeboard uncertainty due to the inclusion of the phase information provided by the CS2 SARIn mode is investigated comparing two scenarios. In the first one, a SAR acquisition is emulated by discarding the phase information as well as by processing only purely specular waveforms – evaluated through the pulse peakiness (PP) parameter. In the second one, waveforms generating from off-nadir leads having a lower PP are processed and the phase information is used to correct for the associated range error.

It is observed that using the extra phase information, CS2 is able to detect leads located up to a distance of 2370 m from the satellite nadir. A reduction in the the total random freeboard uncertainty of ~40% is observed by taking advantage of the CS2 interferometric capabilities, which enable to include ~35% of the waveforms discarded in the SAR-like scenario.

REFERENCES

Thomas W. K. Armitage and Malcolm W. J. Davidson. Using the Interferometric Capabilities of the ESA CryoSat-2 Mission to Improve the Accuracy of Sea Ice Freeboard Retrievals. Ieee Transactions on Geoscience and Remote Sensing, 52(1):529–536, 2014. ISSN 15580644, 01962892. doi: 10.1109/tgrs.2013.2242082

Meighen ice cap (79N, 80W) is a small stagnant ice mass occupying ~60km² of eastern sector of Meighen Island, Nunavut, Canada. Despite its low elevation profile (maximum height reaches 270m a.s.l.), the Meighen Ice Cap has maintained a relatively healthy mass balance since measurements began the early 1960’s with 6 out of 10 years experiencing growth prior to 2005. The persistence of this low elevation ice cap has been attributed to its proximity to the Arctic Ocean from which frequent fog events suppress summer melt. Since 2005 however, the Meighen Ice Cap has experienced rapid shrinkage in both area and thickness due to a sharp increase in the magnitude and duration of summer warming. This presentation will integrate results from in-situ glacier measurements and airborne laser altimetry in order to validate and independently assess seasonal and inter-annual changes in mass of the Meighen Ice Cap as measured from the CryoSat-2 radar altimeter since its launch in 2010.  

Global multispectral microwave radiometer measurements have been available for several decades. However, most current sea ice concentration algorithms still only takes advantage of a very limited subset of the available channels. Here we present a method that allows utilization of all available channels as well as the combination of data from multiple sources such as microwave radiometry, scatterometry and numerical weather prediction.

 Optimal estimation is data assimilation without a numerical model for retrieving physical parameters from remote sensing using a multitude of available information. The methodology is observation driven and model innovation is limited to the translation between observation space and physical parameter space

 Over open water we use a semi-empirical radiative transfer model developed by Meissner & Wentz that estimates the multispectral AMSR brightness temperatures, i.e. horizontal and vertical polarization at channels between 6 and 89 GHz as a function of a limited set of physical parameters, i.e. atmospheric water vapor, cloud liquid water, wind speed, surface and air temperature. This type of model is ideal for optimal estimation applications because of its limited set of free variables. The atmosphere/open water model is adapted to simulate the atmosphere over sea ice and to work over intermediate ice concentrations by linear scaling of the surface emissivity and surface effective temperature.

The simulation of the surface brightness temperature of sea ice requires a separate forward model. Important physical parameters include snow layering, scattering in the snow and sea ice, effective temperature and ice concentration. Here we are testing and evaluating different models. Ice emissivity model development and validation is based on time series of data from Ice Mass Balance Buoys which is co-located with AMSR data and ERA Interim data.

A priori knowledge of each of the physical parameters is used to constrain the solution and improve the retrieval. We test two different a priori options: 1) climatology and 2) numerical weather prediction.

The retrievals are compared to the ESA CCI round robin reference dataset to verify improvements.

A prescribed co-variance matrix both for the a priori set of parameters and for the suite of AMSR brightness temperatures are used in addition to constrain the retrieval. These matrices are derived from an analysis of the ESA CCI round robin reference dataset. Over open water the reference data is a co-location of satellite SST, ERA Interim re-analysis data and observed brightness temperatures. Over ice the reference data is a co-location of ERA Interim re-analysis data, and observed AMSR microwave brightness temperatures.

Due to the nonlinearity of the radiative transfer equation we need an iterative approach to obtain the optimal estimate.

The paper will demonstrate results of retrievals of SST and Sea Ice Concentration.

The grounding line location (GLL) of glaciers and ice streams is an important parameter for mass balance calculations, ice sheet dynamics modeling and the understanding of ice - ocean interaction. The migration of the GLL can be a first indicator for ice thinning (at the surface or bottom).  Within ESA’s Antarctic Ice Sheet project which is carried out within the frame of the Climate Change Initiative (CCI) program, we use SAR data from various satellites to map the grounding line for major ice streams and outlet glaciers in Antarctica. The outcome of the project will be climate data records related to the Antarctic Ice Sheet essential climate variables created from long-term spaceborne observations.

InSAR is an excellent tool for delineating the grounding line with high accuracy. A minimum of three subsequent repeat pass acquisitions are combined to remove the constant ice velocity in the double difference interferogram in order to reveal the vertical deformation in the transition zone between the grounded and fully floating ice parts. Besides data availability (especially in higher latitudes) the phase decorrelation due to natural changes at the ice surface between subsequent satellite overflights is the main limiting factor for identifying the grounding line.

Our investigations focus on ice streams at the interior of the Antarctic Ice Sheet where we expect highly coherent interferometric TerraSAR-X data sets. Although less dynamic than the coastal areas, a significant amount of ice is transported into the largest ice shelves (Ross and Ronne-Filchner) of the continent. In addition preliminary results of GLL at latitudes below 78°S derived from Sentinel-1 data will be shown.

In order to obtain the GLL, the upper limit of flexure, the end of the fringe belt, must be mapped which is done manually in most cases. For the processing of a larger amount of data, possibly with an operational character, it is desirable to detect and derive this line automatically, especially for simple shapes and highly coherent interferograms. We present results of a prototype which is based on Laplace and morphological Laplace operators which will be applied on the unwrapped phase images of TerraSAR-X.

Remote sensing is a crucial component to gain insight in the worlds ice sheets and glaciers. Spaceborne Synthetic Aperture Radar data have proven to be a key resource to monitor the great ice sheets in Antarctica and Greenland. International efforts undertaken during the last International Polar Year resulted in the collection of vast amounts of data used to generate the first continent-wide ice velocity map of Antarctica, a series of full velocity maps of Greenland, and time series data in key regions. The Antarctic grounding line was also mapped at unprecedented accuracy using InSAR. The end of several SAR missions since 2010 has posed a significant challenge in the effort to provide ongoing data acquisitions.

In international collaboration through the Polar Space Task Group, space agencies coordinate their science acquisitions in Polar Regions. New generation missions show potential to not only fill the data gap, but to make the collection of ice sheet data part of the ongoing acquisition scenarios, therefore ensuring data continuity. Several future missions are in various stages of development, thus further adding to the suite of sensors potentially available to collect data in Polar Regions going forward.

With broad input from the larger ice sheet science community, we have worked closely with space agencies to define science requirements and to develop acquisition scenarios that maximize science value for ice sheets Sentinel-1 constellation of the European Copernicus Programme with its open data policy, plays a key role in providing regular data acquisitions in coastal Antarctica and on the Greenland ice sheet. RADARSAT-2, heavily utilized in acquiring data post IPY, is now focused on again covering interior regions of Antarctica with a dedicated left looking campaign. This major data acquisitions are complemented by ALOS-2 PALSAR-2 (L-Band) mapping select regions as part of its basic observation scenario as well as X- Band SAR missions TerraSAR-X, TanDEM-X and COSMO SkyMed collecting high-resolution data in key coastal regions. The coordinated acquisition strategy of each individual mission enables the big picture that leads to a lasting data record of Antarctica and Greenland. Challenges include sensor availability, acquisition conflict resolution, cost, and data availability. The latter is an issue because of different data ownership scenarios, distribution models, and legal frameworks for each SAR sensor.

Here we give an overview of SAR data repositories of ice sheets, highlight the current collaboration effort of SAR data providers through the PSTG SAR Coordination Working Group, summarize the input of the ice sheet science community to the Polar Space Task Group, and present the acquisition strategy that resulted from these efforts. Based on recent acquisitions, we also demonstrate how the current generation of SAR satellites will be used to provide Earth System Science Data for ice sheets and how they fit into existing data sets.

Icebergs account for a substantial part of the freshwater flux in the Southern Ocean, even as large as the basal melt rates of Antarctica ice shelves. Yet, the processes through wich they inject freshwater into the ocean -which are direct melting and fragmentation- are insufficiently informed , and the space and time scales involved are still poorly known.
Indeed, most small icebergs stems from the dislocation of larger ones. Since they don't necessarily drift in the same way (big icebergs are more sensitive to « deep » currents, while small ones are mostly moved by winds) they tend to inject freshwater following different spatio-temporal patterns.
To adress these issues, we analyze two icebergs databases  : a 20 year database of small to medium (<8 km²) icebergs distribution around Antarctica derived from the measurements of 9 altimeter missions (for instance Topex/Poseidon or ERS-1 altimeters) (Tournadre et al. 2015), and the tracks of large (>200 km²) icebergs from NIC and BYU databases as well as an estimate of their volume (Tournadre et al. 2015). We have then acces to a global ice volume and size distribution.

First, we monitored the decrease of the freeboard of the giant iceberg C-19a (with Envisat and Jason altimeters) in Southern Pacific along with the surrounding sea surface temperature (SST) (we considered water with positive SST) and estimated a mean basal rate of 40m/year, consistent with previous modelling studies. We also evaluated an average side melt rate of 3,5m/day, but this needs to be regarded as a rough estimate considering the altimeters' sampling and the fact that icebergs tend to spin often. Analyzing the two iceberg databases together, we found that this melting accounts for around 18% of the total mass loss of « big » icebergs while the remaining 80% is through breaking. Consistent with that melting, we observed strong SST anomalies (up to -1,5°C) in the wake of C-19a (as well as for other big tabular icebergs). The persistence and horizontal spreading of this « coldwater plume » (which could result from the melting of small icebergs born from fragmentation) will be investigated using microwave SST (level 2 AMSR-E) compared with winds and currents data.

We secondly studied the fragmentation of the icebergs. Big tabular icebergs (namely C-19a, which is originally 163 by 37 km) tend to drift with a plume of « smaller » icebergs (maximal length <2km) corresponding to large anomalies in ice volume in open sea, especially when these tabular icebergs go far north. Nevertheless, a non negligible number of small icebergs are located at least 500 km away from a contemporary « big » one, apart in the coastal areas and in the « iceberg alley » (upper part of the weddell gyre) where small icebergs either originate from big ones or merely follow the same trajectories. This this could be due to the fact that small icebergs calving from a large one can drift over vast distances and long periods, which hampers a direct correlation between the two. To test this hypothesis we performed a cross-correlation analysis between the ice volume of small and large icebergs : the correlations reach a maximum for time lags depending on the basins and latitude considered. Finally, the size distribution follows well a power law of slope -3/2 which is typical for brittle fragmentation.
Dependance of the fragmentation mechanisms with respect to environmental parameters such as sea state, sea and air temperature or winds could be discussed.

Seasonal snow and also multi-year snow on glaciers and ice sheets shows a specific anisotropy which is determined by external thermodynamic forcings. Snow fall, gravity, and a temperature gradients act on snow in the vertical direction whereas no such forces exist in the horizontal plane. The layered structure of snow is an obvious evidence that snow is not isotropic. Less obvious is an anisotropy of microscopic ice grains. Nevertheless, the anisotropy of the microstructure of snow has wide implications on the thermal conductivity, mechanical stability, and also the electromagnetic permittivity.

The anisotropic dielectric permittivity makes it possible to determine the anisotropy of snow with microwaves sensors in the field, from radar sensors mounted on towers, and even from space-borne sensors like active and passive microwave satellites.

I will present results of anisotropy measurements done by the SnowScat instrument in Sodankylä, Finland in comparison with space-borne results from TerraSAR-X. The currently longest time series of 4 years of anisotropy data provide a basis for further development of current snow models. Current snow models do not (yet) consider the anisotropy of snow despite its strong influence on the thermal conductivity of snow. To close this gap, I will suggest a phenomenological snow model which is able to simulate the measured  time series of the anisotropy. The model does not only confirm the radar measurements, but provides even a vertical resolution of the anisotropy distribution within the snow pack. The simulated anisotropy profiles are validated by in-situ measurements of vertical anisotropy profiles determined by computer tomography.

The possibility to measure the anisotropy, but also the understanding gained by the developed thermodynamic opens new applications for radar remote sensing. I will show examples of Sibiria, Greenland, and the Swiss alps where anisotropic propagation effects of polarimetric microwaves have been observed. The knowledge gained by understanding the anisotropy might explain polarimetric effects over Greenland which still remain unsolved.

The glaciers are a natural global resource and one of the principal climate change indicator at global and local scale, being influenced by temperature and snow precipitation changes. Starting from 1894, several investigations have been carried out to study the complex glaciers dynamic. Among the parameters used for glacier monitoring, the glaciers surface velocity is an important element, since it influences the events connected to glaciers changes (mass balance, hydro balance, glaciers stability)

In detail the glacier surface velocity is:

connected to the glaciers mass balance, since it is useful in order to compute the flow rate that reaches the ablation area

a stability indicator, since it measures rate at which a glacier is sliding

important for tracking material transportation and erosion phenomena

In particular, the surface glacier velocity can be measured using both in-situ survey and remote sensing geomatic techniques. The optical and SAR satellite imagery enable the continuous monitoring of wide areas of the Earth surface and provide information independent from logistic constraints. On the other hand, using the in-situ survey, it could be difficult to cover wide and not accessible areas.

SAR data with respect to optical one has several advantages; it is characterized by very precise satellite orbit that provides high resolution and accurate mapping capabilities; moreover SAR sensor has the remarkable advantage to collect images in any illumination and weather conditions. The leading idea is the development of a new software able to automatically retrieve glaciers surface velocity using SAR data in particular Sentinel-1 data. The key aspects that Sentinel-1 images present are the free access policy, the very short revisit time and the high amplitude resolution.

The innovative software will perform the following main functionalities:

imagery co-registration

glaciers surface velocity computation adopting a reliable algorithm specifically developed for SAR data

availability in real time of all the computed weekly, monthly and annual glacier velocity fields

In addition to the glaciers velocity analysis, an automatic update of the glaciers footprint could be performed using the Sentinel-2 optical data, exploiting the multispectral information which easily allow to classify the glaciers areas.

In order to verify the reliability of the proposed approach, a first experiment has been performed using Sentinel-1 imagery acquired over the Baltoro Glacier that is one of the world’s largest debris-covered glaciers (about 66 km long) located on the south side of the Karakoram Range.

In particular, a stack of 7 images acquired in the period from December 2014 to June 2015 has been used in order to investigate the potentialities of the Sentinel-1 SAR sensor to retrieve the glacier surface velocity; at present, only the half of the available image stack was considered, so that the images were approximately collected every month. The aim of this test was to measure the glacier surface velocity between each pair in order to produce the velocities fields along the investigated period. The necessary co-registration procedure between the images has been performed and the glacier area has been sampled using about 300 well-distributed points, where the velocity has been measured, for each image pair, using a template matching procedure

Subsequently, an outlier filtering procedure based on the signal to noise ratio values has been performed, in order to exclude from the analysis unreliable points. The uploaded figures reports an example of the velocity fields obtained with image pairs acquired between October and December 2014 and the reported values are similar to those obtained in previous studies. The results highlight that it is possible to have a continuous update of the glacier surface velocity field that is very useful to investigate the seasonal effects on the glacier fluid-dynamics.

Significant wave height (SWH) and surface wind speed (WS) products from the CryoSat-2 Synthetic Aperture Radar (SAR) Mode are validated against operational ECMWF atmospheric and wave model results in addition to available observations from buoys, platforms and other altimeters. The data used here is output from the SAMOSA ocean model processed in the ESRIN G-POD service open to the Community called SAR Versatile Altimetric Toolkit for Ocean Research & Exploitation (SARvatore). The data cover two geographic boxes: one in the northeast Atlantic Ocean extending from 32°N to 70°N and from 20°W to the prime meridian (NE Atlantic Box) and the other in eastern Pacific extending from 2.5°S to 25.5°S and from 160°W to 85°W (Pacific Box). The period extends from 6 September 2010 to 30 June 2014 for the NE Atlantic box and from 7 May 2012 to 30 June 2014 for the Pacific Box. The amount of data is limited by the CryoSat SAR mode acquisition capability over ocean but high enough to ensure robustness and significativeness of the results (Sentinel-3 will operate in SAR mode over the whole ocean). The results show that the quality of both SWH and WS products is very high. Detailed statistics will be reported.

The Pamir and Karakoram mountain ranges are one of the places on Earth with a high abundance of surge-type glaciers (Sevestre & Benn 2015; Copland et al. 2011; Kotlyakov et al. 2008). Thereby, analysis of multi-temporal satellite images and local field evidence revealed that a large number of glaciers in the central Karakoram are currently (last 15 years) surging (Rankl et al. 2014; Hewitt 2007) or have done so in the past and now surge again (Copland et al. 2011). It was speculated that the recent high surge activity in this region is also an expression of climatic changes (e.g. increased winter precipitation), which was named the ‘Karakoram Anomaly’. However, only some of the more recent surges have been analysed and described in detail (e.g. Hewitt 2007) and an overall and/or detailed analysis of individual surges going back to the 1960s is missing despite available satellite images.

For this study a large number of declassified satellite images from the Corona (1961, 1965, 1969, 1971) and Hexagon (1973, 1980) missions were analysed in combination with a more or less complete (near-annual) time series of Landsat images (1989 to 2015) to reveal the timing of individual surges and identify previous surges for 25 glaciers. The resulting fifty-year time-series revealed that most of the currently surging glaciers have also surged in the 1950s and 1960s and that the previous and current appearance of surges is in most cases very similar. The analysis also showed a large variety in the timing of the surges (from very fast and short-lived to very slow and long-lasting) with a partly complex interaction between surging tributaries and blocked main glaciers. Also the repeat cycles vary greatly, from one glacier that is surging every 20 to 25 years (Figure 1) to another one (just 8 km away on the opposite side of the valley) that is advancing since 50 years. In contrast to most other regions in the world, many of the surging glaciers in this specific region are comparably small, steep and debris free.

The large variety of glacier surges and surge-type glaciers that was found in this small region confirms initial observations forwarded in a previous study by Meier and Post (1969). It also stresses the principle difficulties in clearly discriminating surge-type glaciers from others.

REFERENCES

Copland, L., Sylvestre T., Bishop M.P., Shroder, J.F., Seong Y.B., Owen L.A., Bush A., & Kamp, U. 2011: Expanded and recently increased glacier surging in the Karakoram, Arct. Antarct. Alp. Res., 43, 503–516.

GlHewitt, K. 2007: Tributary glacier surges: an exceptional concentration at Panmah Glacier, Karakoram Himalaya, J. Glaciol., 53, 18-188.

Meier, M.F. & Post, A. 1969: What are glacier surges?, Can. J. Earth Sci., 6, 807-817.

Rankl, M., Kienholz C. & Braun M. 2014: Glacier changes in the Karakoram region mapped by multimission satellite imagery, Cryosphere, 8, 977-989.

Sevestre, H. & Benn D.I. 2015: Climatic and geometric controls on the global distribution of surge-type glaciers: implications for a unifying model of surging. J. Glaciol, 61 (228), 646-662.

Kotlyakov, V.M., Osipova G.B. & Tsvetkov D.G. 2008: Monitoring surging glaciers of the Pamirs, central Asia, from space, Ann. Glaciol., 48, 125-134.

Special Session: Earth Observation and Cryosphere Science meeting

Mountain glaciers are of great importance for the water balance on the Tibetan Plateau. Melting water from snow and ice has a high hydrological importance for the large rivers originating on the Tibetan Plateau. Rising temperature cause rapid loss of glacier mass on a global scale. Recent studies showed a negative mass balance for most of the Tibetan Plateau and slightly positive mass balance in its NW part in the last two decades. To predict water availability on the Tibetan Plateau and the downstream regions, it is important to quantify the changes of the cryosphere. As glaciers in this region are remotely located and difficult to access, remote sensing data play a paramount role in the estimation of glacier mass balance and hydrological modelling.

In our study we focus on the mass balance of glaciers on the Tibetan Plateau over the last decades. For this purpose we analyze TanDEM-X data from the scientific phase for three test sites distributed over the Tibetan Plateau with different accumulation regime. The X-band radar system due to its low penetration abilities suits well for glacier surface analysis. We suppose that the interferometric baseline has an important role for glaciological applications in high altitude context. Therefore we evaluate TanDEM-X acquisitions from the complete scientific phase with different baselines to understand what the optimal parameters for glaciological applications are.

Since October 2014 the TanDEM-X SAR satellite constellation operates in an experimental mode, the TanDEM-X science phase, which is split into a long and a short baseline phase. In our study we derive volumetric changes of glaciers by comparing current bistatic TanDEM-X Science Phase data with older TanDEM-X data. Alternatively, we use optical stereoscopic data, SRTM-X data from February 2000 or elevation data from ICESat (2003 – 2009) for the comparison.

To generate recent elevation data, single pass bistatic TanDEM-X data are interferometrically processed to gain high resolution DEMs. Bistatic single pass approaches hereby fit best to avoid temporal decorrelation effects of the glaciers on the interferometric products. To validate the TanDEM-X DEMs, ICESat elevation data on bare rock surroundings are used.

A further objective of our work is to assess the potential of the TerraSAR science phase data for an estimation of snow pack thickness over the accumulation areas of glaciers. We process a time series of TerraSAR-X acquisitions over a study site in SE of the Tibetan Plateau. The thickness estimates will be compared with the occurrence of solid precipitation from reanalyzed climate data sets.

The Arctic sea-ice cover has experienced large changes. Since the 80’s is observed an increase of summer melt time period, a decrease in ice extent and ice thickness. Resent observations showed an increase in drift speed, a growth of strain and deformation of ice fields. The ice motion and strain produce leads and polynyas that have an crucial importance for heat, water and CO2 transfer between ocean and atmosphere, for regional surface albedo, winter ice production, transportation and arctic animals survival.

SARAL/AltiKa altimetric mission launched in 2013 has demonstrated a good capacity for detecting the leads and the coastal polynyas as well as high potential for observation of spatial and temporal dynamic of water openings. The method is based on analysis of along-track radar waveforms with collocated high-resolution Landsat images in order to localise ice/water transitions. We show that comparing to ENVISAT RA-2 altimeter, the pulse peakiness parameter can not be used for leads detection due to saturation of AltiKa signal over leads. We have found that maximal peak power (Pmax) of returned radar echo is a good discriminative parameter and defined the threshold for lead detection. The monthly lead fraction maps 0.5°x0.5° resolution are produced and available on LEGOS CTOH. For three arctic regions - the Baufort Sea, the Kara Sea and the Laptev Sea, we compare the altimetric estimates of lead fraction with thermal infrared MODIS leads retrievals. Both products agree well in temporal dynamic and clearly show the spatial differences. However, AltiKa lead fraction estimates are lower than those of MODIS. The main reason of the discrepancy between these products arises from different spatial resolution of sensors. While MODIS starts to detect the leads larger than 1 km in width, AltiKa sees the openings in the range of ~200 m - 2-3 km. So, the combination of the high-resolution altimetric estimates with medium-resolution thermal infra-red lead fraction products could enhance the capability of remote sensing to monitor sea ice fracturing.

Understanding the causes and magnitudes of changes in the cryosphere remains a priority for Earth science research. Over the past decade, NASA’s and ESA’s  Earth-observing satellites have documented a decrease in both the areal extent and thickness of Arctic sea ice, and an ongoing loss of grounded ice from Greenland and Antarctic ice sheets. Understanding the pace and mechanisms of these changes requires long-term observations of ice-sheet mass, sea-ice thickness, and sea-ice extent.

NASA’s Ice, Cloud, and land Elevation Satellite (ICESat) mission, which operated from 2003 to 2009, pioneered the use of laser altimeters in space to study the elevation and freeboard of the Earth’s ice sheets, glaciers, and sea ice. Since ICESat stopped collecting data in October 2009, the IceBridge and CryoSat-2 missions continue these important observations. Unlike ICESat, which had a single laser beam with a 70-m footprint and took measurements at 150 m along-track intervals, ICESat-2 has three pairs of beams, each pair separated by about 3 km across-track with a pair spacing of 90 m. The spot size is 14 m with an along-track sampling interval of 0.7 m. This measurement concept is a result of the lessons learned from ICESat.  The multi-beam approach is critical for estimating cross-track slope around the margins of Greenland and Antarctica enabling the calculation of elevation change on a seasonal basis.  For sea ice, the dense spatial sampling (eliminating along-track gaps) and  the small footprint size are especially useful for sea surface height measurements in the often narrow leads needed for sea ice freeboard and ice thickness measurements. Additionally the three pairs of beams provide significantly better spatial coverage. The talk with illustrate the concept of ICESat-2 and its potential for polar research.

Data from the snow product algorithm from the Advanced Microwave Scanning Radiometer – 2 (AMSR2) aboard the Global Change Observation Mission – Water are used to monitor snow accumulation through the 2014-15 and 2015-2016 winter seasons. AMSR2 is designed as a follow-on from the successful Advanced Microwave Scanning Radiometer – EOS that ceased formal operations in 2011. Coarse spatial resolution passive microwave measurements are sensitive to volume scattering from snow as a function of dry snow accumulation but have limitations in their capabilities for operation activities not least because of the generalization of snow accumulation by coarse resolution observations, but also because they are unable to retrieve snow accumulation during the snow melt season, when the scattering properties of snow are nullified. The European Space Agency’s Sentinel-1 mission carries a C-band synthetic aperture radar (SAR) or C-SAR system that is sensitive to wet snow and can detect regions of wet snow at sub-pixel resolutions in the passive microwave data. Furthermore, the Sentinel-2 instrument package caries the Multi-spectral Instrument (MSI) that is capable of observing the Earth at even higher spatial resolutions at visible-infrared wavelengths to map snow cover extent under cloud-free conditions. In this study we combine Sentinel-1 and Sentinel-2 observations of snow to illustrate how a multi-sensor approach can help explain uncertainties in the AMSR2 product. Several test case studies are used to illustrate how the Sentinel instruments are helping improve the AMSR2 estimates of snow accumulation.

Here we present a detailed analysis of atmospheric form drag over Arctic sea ice using high resolution, three-dimensional surface elevation data from the NASA Operation IceBridge Airborne Topographic Mapper (ATM), and sea ice roughness estimates from CryoSat-2.

Surface features in the sea ice cover are detected using IceBridge data and a novel feature-picking algorithm. We derive information regarding the height, spacing and orientation of unique surface features from 2009-2014 across first-year and multiyear ice regimes in the western Arctic Ocean. CryoSat-2 surface roughness estimates are used to extrapolate these results across the entire Arctic.

The ice topography estimates are used to explicitly calculate atmospheric form drag coefficients; utilizing existing form drag parameterizations. The atmospheric form drag demonstrates strong regional variability, linked to variability in the height and spacing of sea ice surface features.

These results are also being used to calibrate a recent form drag parameterization scheme included in the sea ice model CICE, to improve the representation of form drag over Arctic sea ice in global climate models.

During the winter of 2008/2009 an unprecedented avalanche cycle occurred basically over the entire Slovakian mountain range. The cycle peaked in the period from 25^th to 31^st of March resulting in more than 200 avalanches in the area of the Tatra national park (738 km²). The whole spectrum of avalanche sizes were observed, ranging from small ones to very large ones (some of which with return periods of up to 100 years). Several huts, bridges, two automatic weather stations and 1 km² of forest were destroyed. Field observations and GPS measurements were done by the team of the Slovakian Avalanche Prevention Centre (APC). The aim was to document and map the event for the purpose of an avalanche cadaster update. Due to the extent of the cycle, the remoteness of most of the area and the risk of secondary avalanches, it was not possible to map many of the avalanches. Therefore, most of the area remained unmapped and much of the valuable information on the spatial extent of the avalanches got lost during the following melting season. Very High Resolution (VHR) satellite imagery was fast recognized as potentially being an important source of information to map avalanches inaccessible by field campaigns. WorldView-1 imagery from April 2^nd, 2009, covering parts of the High Tatra Mountains, was acquired by APC in order to detect and map avalanches in an automatic manner.

We present the results of an object oriented mapping algorithm developed in eCognition in order to automatically identify and map avalanche deposits. In order to quantitatively assess the performance of the algorithm, we selected thirteen test areas in which all avalanches were visually identified and manually digitized by an avalanche expert. In these areas, the avalanches automatically mapped by the algorithm were quantitatively compared with the avalanche expert mapping.

Over the past two decades rapid changes have been observed on the Antarctic ice sheet including ice thinning, ice speed up and inland grounding line retreat. This change has been most pronounced in the Amundsen Sea Sector of West Antarctica where over a 40% increase in ice speed has been observed on Pine Island Glacier and surface elevation change of up to 9 meters per year is now occurring on Smith Glacier. Marine based sectors of the Antarctic continent such as the Amundsen Sea sector are especially vulnerable because the ice and bedrock geometry promote accelerated grounding line retreat, raising concerns for future sea level rise. In the Amundsen Sea sector it has been shown that over 30 km of inland grounding line retreat and ice mass loss was caused by warm Circumpolar Deep Water melting the ice shelf base. However, it is still not clear whether these changes are the result of increased temperatures in previously cold ocean water, or the migration of a warmer ocean current towards the Antarctic coastline. Estimates of global ice mass losses are now entirely reliant on satellite Earth observation datasets such as ERS-1/2, TerraSAR-X, CryoSat-2, Sentinel-1, as they provide the spatial and temporal sampling and multi measurement capability necessary to monitor and determine the processes driving imbalance. Here we exploit a suite of techniques including quadruple difference interferometry and offset tracking, and satellite datasets such as Synthetic Aperture Radar and altimetry data acquired over the last 25 years, to quantify the magnitude and spatial and temporal pattern of ice thinning, speedup and grounding line retreat in West Antarctica.

Understanding snow accumulation and its spatial and temporal variability has for a long time been an almost unattainable goal due to the complex variability in time and space. In order to increase the accuracy in the linkage between image acquisition/modeling and snow physics, the key is tailored collection of field data to make high quality validation dataset representing the required resolution.

 

Snow depth measurements via satellite are more or less non-existing, however the quality of the snow water equivalent (SWE) products and models estimating SWE improves such as ESAs GlobSnow for example. Here we propose three steps towards generating qualitative validation datasets for remote sensing purposes with adequate resolutions.

 

First: We use snow depth information from a range of sources (GPR, probing, automatic weather stations) and implement a model that provides an accurate calculation of snow bulk density that can be used to derive the SWE parameterization based solely on snow depth. This in order to validate existing products. 

 

Second: The project will also develop new algorithms that combine multiple sensors exploiting the higher spatial resolution of synthetic aperture radars. Current operational systems are poorly adapted to upland environments, despite the importance of these regions in for water storage in the snowpack.

 

Third: By gathering data from a range of providers who all have different reasons for observing snow depth we aim to generate datasets with extensive cover; spatially and temporally by coordinating projects and data usage. Synergies will generate easy data access, links for cooperation and a secure storage for snow observation.  

 

By fulfilling these three steps we can provide more accurate and current validation datasets in different resolutions for a range of satellite imagery and satellite-derived products. 

Thermokarst lakes are abundant and highly dynamic landscape features of permafrost lowland regions in western Alaska and provide important ecosystem services as habitats, hydrological feature, biogeochemical hotspots, and for surface energy budgets. Permafrost in this ca. 300,000 km² region follows approximately a North to South gradient of spatial continuity from continuous to sporadic permafrost zones, which also affects lakes and their dynamics on various temporal and spatial scales. Climate change in western Alaska has resulted in a significant warming of air and ground temperatures over the last decades and is projected to continue on that trajectory. To characterize the vulnerability of lakes as well as permafrost to climate change in this region, we assessed historic lake changes in major lake districts of western Alaska for the period ca. 1950 to ca. 2015 using various remote sensing approaches within a set of several independently funded studies. In particular, we were interested in the dynamics of lake growth and drainage in relation to permafrost degradation. Our method focused on the analysis of image time series built from the 30-60m resolution Landsat record for the 1970-2015 period. The observation period was further extended by unaltered historic USGS topographic maps that contain hydrology features and are based on aerial photography from ca. 1950. Our remote sensing studies were complemented by permafrost and lake hydrology field studies as well as aerial flights to validate remotely sensed lake drainage events. Additional validation of lake change was conducted locally with high resolution imagery from Spot-5, aerial photographs, and the DigitalGlobe constellation of satellites. Here, we synthesize the core results from these studies.

The data was processed in three main categories. First we extracted water bodies from recent (2013-2015) Landsat-8 Observing Land Imager (OLI) images of the entire region using simple pixel threshold methods in ENVI^TM and compared these with waterbodies digitally extracted with ArcGIS^TM tools from unaltered historic (ca. 1950) USGS topographic map data to identify hotspots of lake change for the entire 65 year period. Second, we processed Landsat data covering major lake districts in the region from three time periods using an object-based segmentation and classification method specifically designed for lake extraction in eCognition^TM. Third, we applied a robust trend analysis developed with open source software and established image pre-processing algorithms to the entire Landsat-record for several large subregions to derive Tasseled Cap, NDVI, and NDWI land cover indices which are useful for studying annual trends in lake changes.

Our findings suggest that a significant portion of lakes in this region has drained over the last decades and that in particular large lakes are vulnerable to disappearance. Initial analyses of relationships of lake drainages with permafrost distribution in the region suggest positive correlations between lake loss and permafrost degradation in much of the region. Our findings highlight that permafrost and lake-rich landscapes in Alaska are already changing rapidly and permanently in a warming world.

This set of studies was supported by funding from NASA Carbon Cycle Sciences, NSF Arctic System Sciences, an European Research Council Starting Grant, and the Western Alaska Landscape Conservation Cooperative. Our study of lake dynamics in a thaw vulnerable permafrost landscape affected by climate change highlights the need for continuation of the Landsat mission as well as the increase of observation density with the new ESA Sentinel-2 mission.

High-resolution information on the spatial distribution of snow accumulation on the West Antarctic Ice Sheet (WAIS) is scarce. Satellite based studies are among the most appropriate ways to estimate accurately these mass balance parameter. In order to evaluate the potential of COSMO-SkyMed X-band Synthetic Aperture Radar (SAR) to identify spatial patterns of snowpack characteristics related to snow accumulation (e.g., snow stratigraphy, snow density, crystal size and morphology), a remote sensing survey was carried out covering Union Glacier (79°46’S, 83°16’W) during the austral summer 2011/2012, simultaneously with ground-truthing measurements on the glacier surface. Acquired COSMO-SkyMed data consisted of four Spotlight-2, seven Stripmap Himage, two Stripmap PingPong, and two ScanSAR Wide Region images, summing a total area of 13400 km² covering Union Glacier and surrounding areas. On the field, seven snow pits were dug in Union Glacier areas with different backscattering patterns previously observed and selected in COSMO-SkyMed images acquired in July 2011. Snow stratigraphic data were then used to test five empirical and five physical based models of snow backscattering at X-band. In general, physically based models showed better results than that obtained by empirical models, and a model representing the interaction between the SAR beam and the snow pack was adapted for the Union Glacier area, representing the physical process involved. Furthermore, statistical modeling was used to understand the statistical relation between SAR-X band backscattering and snowpack stratigraphic data, resulting in a set of three empirical equations relating the backscattering with different stratigraphic parameters of the snow pack. A maximum penetration depth for the X-band SAR beam of 109 cm was also estimated for dry snow areas. Relation between both satellite and ground-recorded data evidenced that X-band backscattering on the Union Glacier surface is affected by surface roughness and snowpack characteristics, giving important information on the spatial patterns of snow accumulation related parameters (i.e., snow stratigraphy and snow density). Additional activities are scheduled to investigate the relationship between small-scale snow stratigraphy variations as detected by GPR surveys and the signal measured by X-band SAR sensors. These data together will give us a better insight in the accumulation process in regional scale, as it will support extracting stratigraphic information through X-band SAR images as well.

A remote sensing procedure to map permafrost probability in Alta Valtellina (North Italy), and that could be extended to larger areas is here presented. The main assumption is that rock glaciers (abundant in the study area), can be considered a proxy for permafrost presence. Therefore, the rock glacier outlines were drawn againstthe high resolution Bing basemaps in QGIS. Only the active ones were selected by visual inspection. Then, a Landsat ETM+ image of September 2001 was processed to mask glacier ice, snow, lakes, shadows, clouds, and vegetation, using different techniques (i.e. band ratio, thresholding, supervised classifications), to retain bare soil and rocks.

To produce a permafrost probaiblity map,the pixels of each Landsat band were separated into two groups: the first representing the rock glacier sample; the second, the other unmasked pixels to be classified. The statistical analysis showed significant difference between the relfectance values of the two groups in the Near Infrared band (band 4). The pivotal statistics of band 4 pixels from the rock glacier locations were then used to perform a first round classification on the remaining pixels and to retain those which showed a certain degree of similarity to the rock glacier set.

Then, hourly global radiation values were modelled using GRASS GIS, and summed up over the summer season. The assumption is that rock glaciers (and hence permafrost), can persist only where the solar input is sufficiently low during thaw season. The modelled values were validated by comparison with long term field observations from three automatic weather stations in the study area (Passo Foscagno, Oga-San Colombano, and La Vallaccia, ARPA Lombardy network). The values over the rock glacier sites were lower than elsewhere on average as expected. The 75th percentile of solar radiation values in the rock glacier sample was set as the upper threshold to discard all the unclassified pixels with cumulative global radiation value above.

Eventually, only those pixels included in both the two previous classifications, and which were located above the minimum rock glacier elevation (>2323 m a.s.l.), were retained in the final permafrost probability map.

The result was then compared with other existing permafrost maps in the area derived with PERMACLIM and PermaNET models.

The new elastic-plastic anisotropic (EAP) rheology that explicitly accounts for the sub-continuum anisotropy of the sea ice cover has been implemented into the latest version of the Los Alamos sea ice model CICE. The EAP rheology is widely used in the climate modeling scientific community (i.e. CPOM stand alone, RASM high resolution regional ice-ocean model, MetOffice fully coupled model). Early results from sensitivity studies (Tsamados et al, 2013) have shown the potential for an improved representation of the observed main sea ice characteristics with a substantial change of the spatial distribution of ice thickness and ice drift relative to model runs with the reference visco-plastic (VP) rheology.

The model contains one new prognostic variable, the local structure tensor, which quantifies the degree of anisotropy of the sea ice, and two parameters that set the time scale of the evolution of this tensor. Observations from high resolution satellite SAR imagery as well as numerical simulation results from a discrete element model (DEM, see Wilchinsky, 2010) have shown that these individual floes can organize under external wind and thermal forcing to form an emergent isotropic sea ice state (via thermodynamic healing, thermal cracking) or an anisotropic sea ice state (via Coulombic failure lines due to shear rupture). In this work we use for the first time in the context of sea ice research a mathematical metric, the Tensorial Minkowski functionals (Schroeder-Turk, 2010), to measure quantitatively the degree of anisotropy and alignment of the sea ice at different scales. We apply the methodology on the GlobICE Envisat satellite deformation product (www.globice.info), on a prototype modified version of GlobICE applied on Sentinel-1 Synthetic Aperture Radar (SAR) imagery and on the DEM ice floe aggregates.

By comparing these independent measurements of the sea ice anisotropy as well as its temporal evolution against the EAP model we are able to constrain the uncertain model parameters and functions in the EAP model.

 

Icebergs are a main threat for marine traffic at high latitudes. This study tackles the issue of iceberg monitoring using the new acquisition strategy of Sentinel-1 SAR intrument at high latitudes in the Southern Hemisphere: observing waves with wave mode imagettes in the rest of the open ocean. We investigate the possilibity to monitor iceberg positions at lowest latitudes taking advantage of this acquisition mode, which consists in 20x20km imagettes every 100km with 5m spatial resolution. Despite its scare space and time sampling, this mode provides numerous imagettes that can be automatically analyzed using a CFAR-based algorithm to detect iceberg ranging between 25 to 100m, sometimes much more.

In the current study, we present the algorithm applied to Single Look Complex (SLC) wave mode imagettes to monitor iceberg location on a monthly basis. Those measurements are compared to altimetry observations despite their different orbits and ranges of detectable icebergs. Ultimately, the premises of a multi-sensor long term iceberg monitoring service in the Southern Ocean is introduced

Ice velocity is fundamental to capture ice dynamics along the coast of the large ice sheets in Greenland and Antarctica. Synthetic aperture radar from ERS, Envisat, ALOS, Radarsat and now Sentinel-1a are useful to document ice velocity, changes over long periods and now changes over short/sub-seasonal periods. These data are in turn combined with airborne sensing of ice thickness with radar sounders to calculate the ice discharge into the ocean. The results are then compared with mass accumulation from snowfall minus surface melt on a basin per basin basis to determine the ice sheet mass balance. Here we will present a recent update of the mass fluxes that goes back the 1970s using Landsat until present with Sentinel-1a temporal time series, and corresponding state of mass balance of Greenland glaciers. A particular emphasis will be placed on the usefulness of Sentinel-1a data and why it is important to have denser time series of velocity measurements of these glaciers to better understand their interaction with climate forcing and their past, present and future contribution to sea level. This work was funded by research grants from NASA's Cryospheric Science Program.

Precise monitoring of the glaciers velocity plays an important role in characterization of glacier related hazards. In this research, Damavand Mountain as the highest one in the South-West Asia was studied in terms of the glacial activity rate. The main glaciers that were precisely investigated included Siyooleh, Northern, Sardagh, Simorgh, and Western.

In this study we have evaluated capability of both optical and radar imagery to derive glacier-surface velocities in the mountainous terrain. We have implemented the methods that are commonly employed to measure glacier surface movements including, cross correlation of optical and radar satellite images, SAR tracking techniques and multiple aperture InSAR (MAI). We also assessed time series glacier surface displacement using our modified method, i.e. Enhanced Small Baseline Subset (ESBAS). The ESBAS has been implemented in StaMPS software with modification of several aspects within the chain of the processing including: filtering prior to phase unwrapping, topographic correction within three-dimensional phase unwrapping, reducing the atmospheric noise, and removing the ramp caused by ionosphere turbulence and/or orbit errors.

Our results show an average surface velocity rate of 50 mm/yr around Damavand mountainous areas.

Recent global low resolution mass estimates from glaciers and ice caps show a significant mass deficit for many ice covered regions over the world. The uncertainty on the mass balance is still very high, in particular for calving glaciers, which are the main contributors to ice loss in certain climatic sensitive areas. The present work contributes to the development of a method for the estimation of the net mass balance component due to the ice export.

Calving is the mechanical loss of ice from glaciers and ice shelves, and can happen in freshwater or in tidal water. The calving flux represents the volume of ice that a glacier is loosing on the terminus in a determined period of time. It can be estimated using an empirical relationship between the frontal glacier width, mean glacier thickness and the calving rate. The calving rate represents the difference between the ice velocity at the glacier front and the change in time of the terminal position. The present work will show how these parameters needed for the estimating the calving flux will be retrieved.

 

The front position of the glacier terminus is detected with an automatic procedure based on variations of the backscattering amplitude of SAR data. This is further needed for the frontal width determination as well as for the displacement of the calving front. The surface ice velocity is obtained by means of feature tracking applied to TerraSAR-X repeat pass acquisitions. The critical part of the calving flux estimation is the ice thickness at the glacier front. Here two approaches are considered. When ground penetrating radar (GPR) data from the glacier-bed are available, is possible to calculate the ice thickness by combining with glacier surface elevation from a DEM (e.g. obtained from TanDEM-X bistatic data). On the other hand, when bathymetry of the pro-glacier lake or fjord was measured it can be used in complement with the frontal height derived from TanDEM-X to calculate ice thickness. Usually in situ depth sounding is older than the SAR data acquisitions. Therefore due to the change of the front position in time the bathymetry data must be adapted to the real glacier front.

  

The former documented calving flux measurements were made on visual iceberg counting and indirect ice volume estimation. In the late 90s the first efforts were made to get the calving flux through satellite images, unfortunately the results had a big error margin due to the use of low-resolution images. More recently a few well implemented investigations using remote sensing techniques successfully report calving on major outlet glaciers in south Patagonia, Greenland and Alaska, although represent isolated efforts. On the contrary our work will present a tool for systematic and repetitive measurements.

 

Our first test site is Upsala glacier located in the South Patagonia Icefield, a well documented freshwater glacier where also previous calving flux calculations are available. Moreover the data at the area allow us to validate our results through the hypothesis on linear relation between calving rate and water depth. 

In this study, InSAR Small Baseline Subset analysis is used to investigate the gradual elevation change due to summer thawing of active layer in tundra permafrost landscape of Barrow, Alaska. This site is one of the Global Terrestrial Network for Permafrost (GTN-P) sites. Being a GTN-P site helps us have access in Barrow to a lot of ancillary data acquired in the last decade, which will be helpful in interpretation of our Radar data analysis. We used a variety of SAR sensors including TerraSAR-X and ALOS images to assess elevation changes in summer seasons. The preliminary result obtained using TerraSAR-X time-series clearly delineates subsidence during the summer by identifying thousands of coherent pixels on the ground. The spatial pattern of InSAR elevation change reflects different thaw-related landscape features of permafrost. To better understand the process and the link between estimated subsidence and landscape features, we also used airborne hyperspectral data to derive characteristics like wetness and vegetation cover of the landscape.

Cryosphere is one of the most important factor of controlling the radiation budget of earth. The snow, a part of cryosphere, is present in porous form and mixed with many impurities like dust after falling to the surface. The freshly fallen snow have the highest albedo and reflect the maximum amount of sunlight incident upon its surface. By the time this snow is also contaminated due to dust and other impurities which causes the decrement of albedo from its surface. The reflectance of snow is very high in visible region of electromagnetic spectrum but decreasing towards the near infra-red. The snow albedo also depends on its grain size (Wiscombe and Warren, 1980; Warren and Wiscombe, 1980; Aoki et al., 2003; Negi and Kokhanovsky, 2011), so this may be a factor which can be essential to study climate change studies. After the deposition of fresh snow it changes it's properties like shape, size, density etc cause of metamorphism. This snow cover is also very important for various other fields such as water resources, climate change, hydropower projects etc. So the monitoring of the snow cover can help in many ways to the human and other living species on the planet. In this study snow grain size has been calculated for the Beas river basin in Himachal Pradesh, part of North Western Himalaya. Since grain size affects its albedo so this study is focused to find out the trend of albedo due to changing snow grain size. To calculate the grain size we have used hyperspectral dataset Hyperion. The snow grain size has been calculated and compared with the albedo of the area. Since the albedo of freshly fallen snow is very high in the visible region of spectrum and decreases with the age of snow but in near-infrared this decrement is in considerable figure (O’BrienandMunis, 1975; Warrenand Wiscombe,1980; Wiscombeand Warren, 1980; Singh, et al., 2010) 2010 ).

Wiscombe, W. and S. Warren (1980a). A model for the spectral albedo of snow. I: Pure snow. J. Atmos. Sci., 37:2712–2733. doi:10.1175/1520-0469(1980)037.

Wiscombe, W. and S. Warren (1980b). A model for the spectral albedo of snow. II: Snow containing atmospheric aerosols. J. Atmos. Sci., 37:2734–2745. doi: 10.1175/1520-0469(1980)037.

O’Brien, H.W. and R.H. Munis. 1975. Red and near-infrared spectral reflectance of snow. CRREL Res. Rep. 332.

Negi, H, S, and Kokhanovsky A. 2011. Retrieval of snow grain size and albedo of western Himalayan snow cover using satellite data, The Cryosphere

Singh, S. K., Kulkarni, A. V., Chaudhry, B. S. 2010. Hyperspectral analysis of snow reflectance to understand the effects of contamination and grain size. Annals of Glaciology

The modern periglacial relief of the Northeast Siberian lowlands formed as a result of thawing Late Pleistocene ice-rich Ice Complex (IC) deposits in Holocene. Watershed remnants of these deposits form uplands, called Yedoma. Climate warming at the end of the Pleistocene – beginning of the Holocene contributed to the activation of thermokarst, which has then become the leading relief-forming factor. As a result, the area covered by thermokarst lakes and basins (alases) and river valleys now exceed by far the Yedoma distribution. However, the IC deposits are characterized by an ubiquitous occurrence of intra-sedimentary ground ice and huge polygonal ice wedges that formed in the ground.  Often this ground ice sums up to very high ice  contents of 90 % per volume. Thus, under modern climate conditions, the territories of the Yedoma distribution are still vulnerable to thermo-mechanical disturbances, yet only little is known about their total volume.

            The aim of this research is the analysis of the qualitative type and quantitative distribution of modern relief-forming permafrost-related processes in the Kolyma lowland tundra. Particularly we focus on an inventory of thermokarst lake area changes and bogging processes using Landsat images. Geomorphometric and volume analyses of Yedoma remnants is based on a DEM derived from ALOS Prism stereo imagery. High resolution change detection of recent process intensity is based on GeoEye and Kompsat-2 time series.

            The Kolyma lowland tundra occupies about 45000 km² and is located within the continuous permafrost zone. Mapping of the Quaternary deposits using Landsat images shows that most part of study region is occupied by alas complex (72%), while IC occupies only 16 %. The overall limnicity of the study region due to thermokarst lakes is 13,5 %. The analysis of thermokarst lake area changes was done by visual mapping using Landsat images for the period from 1972 to 2015. The general trend is the complete or partial drainage of lakes. However, it concerns only a small portion of lakes since for the major part of the study region no substantial lake area dynamics have been revealed. Accordingly, the formation of new thermokarst lakes on Yedoma uplands was not detected. The majority of thermokarst lakes are located inside existing alases, suggesting either residual lakes or lakes of secondary genesis. The more widespread drainage of thermokarst lakes occurred in different areas and at various times, without a pronounced spatial pattern. In contrast, a slight but steady increase of thermokarst lakes area was detected in the lower most coastal area. Interannual dynamics are typical for lakes within the river valleys and not considered as permafrost-related dynamics.

            Our analyses revealed that the Yedoma relief can be differentiated  into various morphological types. The massive morphological type with an area of up to 50 km² is the most vulnerable to thawing under modern climate conditions. For the massive Yedoma type, sporadic small early-stage thermokarst lakes and bogging processes on the flat surface are typical. The DEM-based analysis of the key site around Cape Maly Chukochy contributes to recent efforts of an inventory for the whole Yedoma region in Siberia and Alaska. It allows to estimate the potential occurrence of permafrost thaw-related processes such as land surface subsidence and to understand how changing climate conditions may impact on the ice and soil organic carbon-rich Yedoma terrain. 

Now there are initial indications that ground ice in permafrost is thawing in response to rising temperatures in the Arctic. However, particularly in the light of the enormous area underlain by ice-rich continuous permafrost, still only few observations of permafrost-thaw related landscape dynamics exist. Thermokarst lake development, active layer detachments, widespread and irreversible land surface subsidence, and coastal erosion, all of these processes taken together present a sketch of current permafrost degradation with consequences for local hydrology, ecosystems, biogeochemical cycling, and sometimes communities. In this study, we focus not only on monitoring these processes, but also aim to develop a best practice strategy for remote sensing image data fusion, in order to combine 2D and 3D information from very high resolution (VHR) imagery of low acquisition frequency (GeoEye, WorldView, KOMPSAT, ALOS PRISM) with high temporal resolution data such as from the RapidEye satellite constellation. Local field measurements (meteorology, ground temperature, geodetic surveys) during several recent Russian-German Arctic expeditions complement our remote sensing studies and help differentiating factors causing relief and land cover changes. Our work aims at finding commonalities and differences of change or no change on yedoma uplands, slopes, and thaw depressions on the landscape scale using multi-temporal DEMs from historical aerial photographies, modern stereo imagery, and on-site geodetic surveys.

 This ground truth data provides the basis for calibration and ortho-correction of simultaneously acquired VHR data. More than 30 RapidEye scenes (level 1B) from 2014 to 2015 were adjusted using a bundle block adjustment procedure. Incoming raw data is constantly included only via common Tie Points, minimising processing efforts. Band metrics such as NDVI and PCA-analyses of multi-temporal image composites were used to differentiate seasonal from inter-annual processes and to identify signatures and trends typical for permafrost thaw related processes on the surface. Our high spatial resolution monitoring for the last decades shows that the current relief development in ice-rich permafrost enhances not only drainage of thermokarst lakes but also drainage of the entire terrain, which provides favourable preconditions for land surface lowering due to permafrost thaw subsidence. A condensation of observation intervals also showed that coastal erosion has accelerated during the last five years, which could be related to parallel sea ice reduction and arctic summer warming. This local understanding of processes will help to identify and quantify permafrost degradation over large remote Polar Regions with future Earth Observation missions. 

Projected future warming is expected to be particularly strong in high northern latitudes. Permafrost present in those areas contains high quantities of “frozen carbon” that could be released in the atmosphere during melt, leading to a strong positive climate feedback. This communication will present improved approaches to monitor continuous twice-daily ground temperature variations in summer (without snow) and in winter (under snow cover), using satellite-based thermal and microwave brightness temperature (Tb) measurements.

For the snow-free summer period, the upper layer Ground Temperature (LGT) can be estimated from thermal infrared Land-Surface Temperature (LST) measurements during cloud-free days. We improved the summer LST monitoring by combining 37 GHz passive microwave brightness temperatures (unaffected by clouds) with thermal infrared data. Calibrated on a pixel-based scale, this approach takes into account the land-cover variation. Using a diurnal LST variation model, we generated a new hourly LST dataset at 25 km scale throughout the ten-year analysis period (2000–2011). However, the climate-induced snow cover variation impacts the ground temperature variation as an insulating layer (depending on its thermal conductivity), which itself is a strong function of the climatic conditions. One must thus develop a method to continuously monitor LGT along the year.

We used satellite MODIS LST and passive microwave AMSR-E Tb data assimilated in a climate land surface scheme (Canadian CLASS model) driven by reanalysis meteorological forcing data and coupled with a radiative transfer model (HUT) in order to generate a twice-daily Tb corresponding to the simulated surface conditions (e.g. soil in summer and soil+snow in winter). The inputs of the land surface scheme and also the parameterization of the snowpack properties are adjusted to minimize the simulated/measured Tb biases. We show that the retrieved simulated ground-temperatures are improved by up to 2 to 5 K when using satellite data compared to the simulated LGT using the model (alone) without constraint. This product was locally evaluated at five experimental sites over Canadian – Alaska Arctic (part of the EU-PAGE21 project) against ground-based measurements over summer periods as well as in winter under the simulated modified snow cover.

The results demonstrate the usefulness of combined thermal and microwave satellite data to generate a yearly continuous LGT database linked to the permafrost evolution at a better resolution and accuracy compared to model reanalysis