New approaches to processing radar altimetry waveforms over complex ice sheet topography

Sea level rise is among the most pressing environmental, social and economic challenges facing humanity, and requires timely and reliable information for adaptation and mitigation. The ice sheets of Greenland and Antarctica currently contribute approximately one third of global sea level rise, yet monitoring their coastal regions, which are often populated by numerous, highly dynamic outlet glaciers remains a challenge. One of the principal methods for monitoring ice sheet change is that of satellite radar altimetry, which provides a near continuous 30-year record of surface elevation and volume change. However, this technique can suffer from incomplete measurements and larger uncertainties over rugged coastal topography, where the instrument may fail to track the ice surface, or may record multiple distinct reflections within the illuminated ground footprint. In these situations, current Level 2 processing approaches can be sub-optimal, leading to inaccuracies being introduced into the resulting elevation measurements. Therefore, this study aims to explore new approaches to retrieving elevation measurements, comprising (1) multipeak waveform retracking and (2) a refined slope correction approach over complex regions. The developed approaches offer the potential for multiple elevation retrievals from a single waveform, and in turn the opportunity to increase both the reliability and quantity of elevation measurements.

Within the study, these processing techniques were developed and evaluated across Russell Glacier and the whole Greenland as two typical test cases, based upon Sentinel-3 SAR altimeter acquisitions over ice sheet regions that exhibit complex topography. Laser altimeters including Airborne Topographic Mapper (ATM) and Ice, Cloud, and Land Elevation Satellite-2 (IceSat-2) data were used as independent validation sources. Ice sheet wide analysis showed that the developed approaches were capable of delivering equally high accuracy for multiple elevation retrievals with comparable dispersion (~ 1 m) but much lower bias (~ 0.5 m) and outliers (~ 4%) compared to standard Level-2 products (~ 4 m bias and ~ 20% outliers). The developed approaches have the potential to further extend the capability of satellite radar altimetry over complex glaciological targets, and to improve the accuracy and coverage of altimeter measurements across these regions.