CryoSat User Workshop 4

We report on the validation of CryoSat-2 elevations at the salar de Uyuni in Bolivia, which is one of the best-characterized flat surfaces on Earth.  CryoSat-2 elevations have been collected in both SARIN and LRM mode over the salar, allowing us to validate elevations from both modes using a 2400 km^2 GPS-derived digital elevation model (DEM).  The salar surface is halite underlain by a shallow water table that keeps the surface permanently wetted, resulting in miminal surface penetration at radar wavelengths.  We analyze data from 62 SARIN-mode passes over the salar DEM from 2010-2014 and an additional 20+ passes of LRM mode data acquired since.  We compare Baseline B and Baseline C versions of the pre-2015 data and evaluate the impact of processing improvements on the Baseline C dataset. Our analysis of Baseline B data has resulted in estimates of the CryoSat-2 range bias and range precision, identification and estimation of an ascending/descending elevation bias attributed to timing error, and a time-varying sinusoidal elevation anomaly with a period of ~1.5 years and amplitude of 7 cm. The sinusoidal anomaly does not align with the annual precipitation cycle and is much larger than ICESat and InSAR-derived surface change estimates over similar timescales, suggesting that it is not due to natural variations in surface elevation or reflectance.

Surface wind speed (WS) and significant wave height (SWH) products of the CryoSat-2 Fast Delivery Marine Mode (FDM) product produced from the Low Rate Mode (LRM) data of the Synthetic Interferometric Radar Altimeter (SIRAL) instrument have been monitored and validated against the corresponding parameters from ECMWF Integrated Forecast System (IFS), in-situ buoy and platform instruments and other satellites (specifically: Jason-2 and SARAL/AltiKa). The equivalent product produced by the National Oceanic and Atmospheric Administration (NOAA) is also used for the comparisons. The quality of the CryoSat wind and wave products and the timeliness of their delivery in near real time (NRT) is assessed.

In general, the CryoSat-2 FDM wind and wave products are of good quality. However, the wind speed product is about 1 m/s higher than the model and what other instruments (in-situ, Jason-2, SARAL/AltiKa and even the NOAA CryoSat product). It is also about 15% noisier than the other sources. On the other hand, the SWH product is very good. However, the product may need minor adjustment at low wave heights. The timeliness of data delivery is quite good with most of the data are received within 1 hour (of the end of the 6-hour time window).

The impact of assimilating CryoSat-2 SWH in the ECMWF Integrated Forecast System is assessed using several numerical experiments. The results show good impact. Operational assimilation of SWH at ECMWF model was realised in May 2015. The results will be presented and discussed.

This study addresses the processing and validation of water levels derived from CryoSat-2 data over rivers and lakes, for the preparation of upcoming Sentinel-3A SAR altimeter satellite. The study is part of the SHAPE project, the SEOM Sentinel-3 for Science Inland Waters activity.

Previous studies have introduced the potential benefits of the CryoSat-2 mission for the monitoring of rivers and lakes water levels. Compared to conventional satellite altimetry missions, CryoSat-2 flies on an unusual geodesic orbit characterized by a long orbit period (369 days) and dense spatial coverage (7km at equator). The onboard SIRAL altimeter implements for the first time SAR and SARin modes, together with conventional low resolution mode (LRM).

The mission's orbit characteristics makes the retrieval of water level time series over rivers and lakes challenging: (1) over rivers, because the measurements are spread along the river path (space) and time and (2) over lakes, in the absence of repeat tracks pattern.

The study will introduce two new frameworks to deal the processing of CryoSat-2 data time series prior validation:

(1) for rivers, dealing with the migration, filtering and editing of CryoSat-2 measurements along the river path thanks to river profile estimation based on Jason-2 and SARAL data;

(2) for lakes dealing with local geoid undulation estimation prior to filtering and editing of CryoSat-2 measurements spread over lake surface.

Validation results against in situ data will be presented for both rivers and lakes. Performances will be compared to those of Jason-2 and SARAL LRM altimetry missions.

Coastal Altimetry is expected to benefit significantly from the advent of SAR altimetry, due to the much higher along-track resolution and the better signal to noise ratio of the SAR mode, and data from Cryosat-2 can be used to assess this improvement. In the framework of the ESA-funded CryoSat Plus for Oceans Project (CP4O) we have carried out such an assessment, using SAR altimetry from CryoSat-2 around the UK coast from the CPP processor, courtesy of CNES, and from the ESA-ESRIN SARvatore GPOD processor.

The diagnostic that we use as a measure of the instrumental noise is the absolute value of the difference amongst consecutive 20-hz samples; over the short distance (~350m) of two consecutive samples, the contribution due to oceanographic phenomena and corrections is of the order of a few mm so the values observed (usually a few cm) are essentially due to the instrumental noise – and their median is an estimate of the instrument’s precision.

As the SAR mode resolution cell extends across-track it is appropriate to first consider this diagnostic as a function of the across-track distance from the coast. Analysis done over one year of CPP data shows that the 20-Hz noise is flat at <6 cm up to 6 km across-track from the coast. Application of a burst-weighting window (Hamming) as done in the GPOD data can reduce this further.

In terms of distance from the closest coastline the numbers are even better, as when the track is orthogonal or nearly orthogonal to the coast the height estimates are reliable to 1-2 km. This is demonstrated with some data along the UK south coast around Brighton (‘Brighton Box’), and also compared to previous results in the Northern Adriatic. With GPOD data the precision is flat at 4.5 cm up to 5 km and still < 6 cm at 3 km. With some screening based on retracking misfit we can still get 4.5 cm precision at 2km but with only about 40% of points valid; and ~4.0 cm at 5 cm with 80% of points valid. In the favourable track orientation conditions of the Brighton Box the noise statistics are flat up to 2 km from the coast. Validation results against tide gauge data  around the UK are also very encouraging, with rms difference values below 10 cm in most cases.

Finally we discuss how this work is being extended within the SCOOP Project, funded in the framework of ESA’s Scientific Exploitation of Operation Missions (SEOM) Program, which is paving the way for the scientific exploitation of forthcoming Sentinel-3 SRAL data in the coastal zone.

CryoSat-2 is a tremendous asset to the oceanographic community, and the exploitation of its data over the ocean represents a welcome additional return for ESA’s investment in a mission whose primary objective is to monitor the cryosphere. The CryoSat Project has approved, in the frame of the CryoSat routine phase, the generation of additional ocean products, which are available since April 2014. Here we present the results of a verification and scientific validation of the Geophysical Ocean Products (GOP), which have consolidated orbits and are available 30 days after acquisition. This assessment, carried out within the CryOcean-QCV project, is performed for the sea surface height (SSH), the significant wave height (SWH), and the wind speed. The mean value of the 20-Hz SSHA noise corresponding to a SWH of 2 m is 6.2 cm for LRM data and 10.2 cm for pseudo-LRM data. The standard deviation of the crossovers is 5.4 cm. The SSH is then validated at the coast against the sea level measured by a set of carefully selected and quality controlled tide gauges, and compared with Jason-2 observations. Correlations between SSH and tide gauge records are statistically significant at nearly all stations, with a mean value of 0.75 and 0.70 for CryoSat-2 and Jason-2, respectively. In the open ocean the SSH is compared globally with the steric heights derived from Argo float-derived temperature and salinity profiles. The mean normalized RMS difference between SSH and steric height is 40%. Regarding the SWH and wind speed, they are both validated against buoy observations. The SWH shows an RMS of 12 cm with virtually zero bias, however the performance of the CryoSat-2 for wind speed is worse with a bias of almost 2 m/s. In addition, the SWH is also compared with the values provided by the Wavewatch III model. Differences between CryoSat-2 and the WW3 are less than 1/5 of the SWH over most areas of the ocean.