A New Fully-Focused SAR Altimetry Processor in the ESA G-POD SARvatore Family: Validation and Applications on Inland Waters

Abstract

Satellite altimetry has successfully been applied to derive water level time series over lakes and rivers for many years. In the last decade major advancement within the field have been achieved. The SAR altimetry era, initiated with the launch of CryoSat-2, has made it possible to measure much smaller targets more accurately. The standard 20 Hz radar altimetry products generally allow to study targets down to a few hundred meters in optimal conditions. However, recent advancement in the low-level data processing, the Fully-Focused SAR (FFSAR) technique, where the footprint in the along-track direction can be as small as a few meters, has made it possible to measure the water level of even smaller targets.

FFSAR processed data from CryoSat-2 in SAR mode is made freely available by ESA through a newly integrated service on the Grid Processing on Demand (G-POD) platform, within the SARvatore Family of custom processors, where the user can order data from an area and specify various parameters to be applied in the processing. The FFSAR processor has been developed within the ESA ESTEC Sentinel-6 Project and adapted to CryoSat-2 for verification and validation purposes.

Here we demonstrate the value of FFSAR and evaluate the FFSAR product obtained from the ESA G-POD service over different inland water targets. The evaluation and validation is done via in situ water level data, external processed FFSAR data, unfocused SAR data and laser altimetry data from ICESat-2.

More specific, we perform an evaluation over the narrow ( < 100 m) American Rivers: Red River, Little River, and Canadian River located in the Mississippi basin. Where we compare the water level based on different retracker e.g. ALES+ and threshold retrackers.
We make a joint water level time series by considering data from CryoSat-2 and ICESat-2 over a reach, and validate the results against in situ data. Preliminary results show that we can capture the main part of the water level signal.

Over the Elbe River, we compare the FFSAR based river water levels with results based on unfocused SAR processing and in situ gauges.

We extend our analysis with a case study conducted over the Lake IJssel which is located in The Netherlands and covers an area of approximately 1100 km2. We examine the performance of the G-POD processor by evaluating the FFSAR lake height estimates using externally processed FFSAR data that have been validated against in situ data.