Glaciers and Ice Caps 2

The use of SAR interferometry and offset-tracking between repeat-pass space-borne Synthetic Aperture Radar (SAR) images is well established in research and operational projects for the estimation of glacier motion. Knowledge on ice surface velocity of glaciers and ice cap contributes to a better understanding of a wide range of processes related to glacier dynamics, for example glacier mass flux, flow modes and flow instabilities (e.g. surges), subglacial processes (e.g. erosion), supra- and intra-glacial mass transport, and the development of glacier lakes and associated hazards. In addition, the comparison of the spatio-temporal variations of glacier velocities will improve understanding of climate change impacts.

The ESA Glacier_CCI project (http://www.esa-glaciers-cci.org) has the main objective to contribute to the efforts of creating a globally complete and detailed glacier inventory. However, as meltwater from glaciers and icecaps provide a substantial contribution to global sea-level rise, the project will also create two additional products in selected key regions, elevation changes and velocity fields. Within the ESA Glacier_CCI project we apply SAR interferometry and offset-tracking to historical and ongoing satellite SAR data for the monitoring of the ice surface displacement of glaciers and ice caps. Particular attention is paid in the project to the Canadian Arctic, Svalbard, the Russian Arctic, and the Pamir and Karakoram mountainous regions. For many glaciers and ice caps historical maps of the 1990's (ERS-1/2 and JERS-1) and 2000's (ALOS-1) were computed.Starting from 2015 velocity fields of glaciers and ice caps are routinely computed with a remarkable quality from Sentinel-1 data for the regions where suitable data pairs separated by a time interval of 12 or 24 days are acquired.

We will present selected results obtained with Sentinel-1 data and highlight significant changes of ice surface velocities evident from the comparison with older maps. A nearly complete ice velocity map of the Svalbard archipelago was computed with winter Sentinel-1 data of 2015 and regular monitoring on selected glaciers is ongoing since mid August 2015. The most evident signal is observed over Basin 3 on the Austfonna ice cap, which shows dramatic changes since 1995, but accelerations are also depicted on a few other glaciers across the archipelago. The comparison of two nearly complete ice surface velocity maps over the Canadian High Arctic for the periods 2007-2011 (PALSAR-1) and the winter of 2015 (Sentinel-1) confirms slightly deceleration in many cases and acceleration for a couple of cases. Over Novaya Zemlya only summer Sentinel-1 results are available so far, indicating in comparison to JERS-1 results of 1998 and PALSAR-1 results of 2008/2009 a general steady increase of frontal velocities along with a retreat of frontal positions. For the Karakoram and Pamir mountainous regions Sentinel-1 performs remarkably well and time series of velocity fields were computed for many large glaciers. A large variability of ice surface velocities is observed in space and time.

Ice velocity measurements produced from Sentinel-2 data are examined to determine how well Europe’s newest multispectral mission performs in various snow and ice environments. Ice velocity is a key observation that can be used to analyse the response of the Cryosphere to changes in climate. Ice velocity can be used in combination with auxiliary ice thickness and surface mass balance data, to quantify ice sheet mass balance but also changes in the ice dynamics that reflects changes in the boundary condition of a glacier linked to basal, frontal or surface processes. Ice velocity has contributed to demonstrate increased mass loss of the Greenland and Antarctic Ice Sheets [Shepherd et al., 2013], destabilization of ice caps [McMillan et al., 2014], slowdown of land terminating glacier of the Greenland ice Sheet [Tedstone et al., 2015] and of mountain glaciers [Heid and Kaab, 2012].

Along Synthetic Aperture Radar (SAR), Multispectral data are the most efficient space-based observation to derive ice velocity, tracking surface features such as crevasses and debris; the use of multispectral data for deriving surface velocity are limited however by the spatial resolution of the sensor which defines the lower bound of detectable surface velocity, as well as constrain relative to cloud coverage, saturation and shadows or to solar illumination (polar environments). Recent advances in processing of multispectral datasets for ice velocity have exploited complete archive to reduce the impact of clouds, saturation and shadows and to improve the accuracy of velocity measurements up to the theoretical limits [Dehecq, Gourmelen and Trouvé, 2015]. 

The recent launch of Sentinel-2 in June 2015 brings in a new era of Earth Observation, and the short 10-day repeat period and large scale coverage will be a useful resource for Cryosphere monitoring in the future. Here we describe the feature tracking methodology developped through an ESA Dragon project and used to produce ice velocity from a complete archive of multispectral imagery with examples applications in the Greenland Ice Sheet and the Himalayas from Landsat 5-8. We will present the first velocity results from Sentinel-2 data over ice sheets and glaciers and will assess the quality that Sentinel-2 achieves over ice with a 10 day repeat, in comparison to existing multispectral and SAR missions. We will compare 2015 ice velocity measurements from Sentinel-2 with data from Sentinel-1 and Landsat-8 and discuss synergy between these various missions.

References

Dehecq, A., N. Gourmelen, and E. Trouve (2015), Deriving large-scale glacier velocities from a complete satellite archive: Application to the Pamir–Karakoram–Himalaya, Remote Sens. Environ., 162, 55-66.

Heid, T. and A. Kaab (2012), Repeat optical satellite images reveal widespread and long term decrease in land-terminating glacier speeds, The Cryosphere, 6, 467-478.

McMillan, M. et al. (2014), Rapid dynamic activation of a marine-based Arctic ice cap, Geophys. Res. Lett., - 2014GL062255.

Shepherd, A. et al. (2012), A Reconciled Estimate of Ice-Sheet Mass Balance, Science, 338, 1183-1189.

Tedstone, A., Nienow, P., Gourmelen, N., Dehecq, A., Goldberg, D., Hanna, E., (2015) Decadal slowdown of a land-terminating sector of the Greenland Ice Sheet despite warming

Recently published results of glacier mass budgets in the Karakoram and Pamir mountain ranges revealed possible balanced or even positive budgets over the last decade. While all studies are consistent in the Karakoram, results for the Pamir deviate. In addition, no long-term mass balance measurements exist for both regions except for Abramov Glacier in the Pamir-Alai located between the Pamir and the Tien Shan. Information about the long-term behaviour of the glaciers in Karakoram and Pamir are mainly based on front variation or area change measurements. Length and area changes, however, show a delayed reaction to climate change and cannot be directly linked to atmospheric conditions or hydrology. The latter is in particular the case for the abundant surge-type glaciers in both regions.

We used two declassified stereo Hexagon images from the mid 1970s for geodetic mass balance investigations and, hence, to extend existing time series. In addition, available digital elevation models (DEMs) from ASTER stereo data (AST14DEM), as well as DEMs generated from ALOS, SPOT and Cartosat-1 covering the year ~2010, and the SRTM DEM acquired in February 2000 were used. We concentrated our analysis on Hunza Valley in central Karakoram and Vakhsh Catchment in north-central Pamir with a glacierized area of ~2500 km² and ~4500 km² respectively. Glacier extents are derived within the Glaciers_cci project and are based on Landsat ETM+ scenes from ~2000. They served as baseline information for glacier coverage and were manually adjusted for the respective years based on the utilized Hexagon and ASTER images.

Our results confirm the high spatial and temporal variability of glacier changes in Karakoram and Pamir with frequent surges also in the period prior to ~2000. We reveal slightly negative mass budgets for the 1973 to 1999 period and slightly positive values for the investigated glaciers afterwards for both catchments. However, the differences are not significant. In agreement with available in-situ measurements, the geodetic mass budgets for Abramov Glacier are clearly negative in our study. The largest uncertainties are related to the penetration of the C-band RADAR into snow and ice for the SRTM DEM and to data voids and outliers in the accumulation regions where low contrast due to bright snow is common on the utilised optical images.

The 2013 Intergovernmental Panel for Climate Assessment (IPCC) Fifth Assessment Report (AR5) concluded that the observed and explained geophysical causes of global geocentric sea-level rise, 1993–2010, is much closer towards closure at 3.2 [2.8→3.6] mm yr^-1 versus 2.8 [2.3→3.4] mm yr^-1.  However, the discrepancy reveals that circa 1.3→37.5% of the observed sea-level rise remains unexplained, despite contemporary reports on reconciled mass balance estimates ofice-sheet and mountain/peripheral glaciers during the early 21^st century. This discrepancy is primarily attributable to the relatively wide range of estimates of respective contributions of Greenland and Antarctic ice-sheets and mountain/peripheral glaciers to sea-level rise, at 0.60 [0.41→0.79] mm yr^-1 and at 0.76 [0.39→1.13] mm yr^-1, respectively.  In particular and for example, the Himalayan glacier system, as part of the Asian High Mountain glacier systems remains a focus of public and scientific debate, as the uncertainty of its mass balance estimates and its future projection have a significant implication of water resource problems potentially affecting 5 billion people, or ~43% of the world’s population in the region.  It is also not clear, for example in the case of the Asian High Mountain glacier system, glacier ablation instantaneously contribute to sea-level rise, as melt water is largely dammed up in an anthropogenic activity in the region.  In this contribution, we provide an updated estimate of mountain glacier mass balance and its contribution to sea-level rise, 2002–2015, primarily using observations including GRACE gravimetry, satellite altimetry, in situ mass balance data and other data sets.

We present an overview of ice velocity measurements produced from data acquired during the first 12 months of the Sentinel-1 mission. Satellite observations acquired over the past 25 years have shown marked ice velocity speed up on individual Antarctic ice streams, with ice velocity increases of over 42% observed on Pine Island Glacier. In Antarctica, areas of ice velocity speed up are dynamically unstable and comprise the largest component of ice sheet sea level rise contribution. However, despite a clear long term trend for increasing ice velocity in many regions, speed up has not been constant through time and multiple years with no significant change have also been observed. It is necessary to make present day measurements of ice velocity to provide an independent means of measuring ice mass loss from the most rapidly changing ice sheet regions. However the spatiotemporal coverage of historical ice velocity measurements has been limited by a paucity of suitable data over the full Antarctic ice sheet and to date, parts of east Antarctica have been observed only a few times during the last 25 years. We present 12 months of ice velocity measurements on 10 key Antarctic ice streams, produced from the normalised cross-correlation of real-valued intensity features in Interferometric Wide Swath (IW) mode Sentinel-1a data. A time series of ice velocity measurements produced from short 12-day repeat Sentinel-1a data over Pine Island Glacier shows that in 2014 and 2015 the ice surface speed has remained constant at ~4 km/year. A Sentinel-1a ice velocity map of the Antarctic Peninsula demonstrates that good quality measurements can be obtained along the full length of the Peninsula using Sentinel-1a. TOPS mode SAR Interferometry (InSAR) results shows that interferometric coherence can be preserved over the 12-day repeat period on stable slower flowing ice covered terrain, such as the Transantarctic Mountains in East Antarctica. We conclude that the short 12-day repeat period and large scale coverage of Sentinel-1a is an invaluable new resource, which in the future will be a critical tool for monitoring ice velocity change and dynamic instability of the Antarctic ice sheet.