Robust Lake Level Extraction in Mountainous Areas By Retracking Cryosat Sarin Mode Waveforms.

Wednesday, 17 December 2014
Marcel Kleinherenbrink1, Pavel Ditmar2 and Roderik Lindenbergh2, (1)Delft University of Technology, Geoscience and Remote Sensing, Delft, 5612, Netherlands, (2)Delft University of Technology, Geoscience and Remote Sensing, Delft, Netherlands
Since the beginning of the 1990s, lake levels can be monitored using satellite altimetry.
This is notably an advantage for lakes located in remote areas, where no gauges are present.
Two disadvantages of traditional satellite altimetry however are the limited number of ground tracks and the pulse limited footprint size.
In 2010 Cryosat was launched with onboard a SAR Interferometric Radar ALtimeter (SIRAL), which has a dense ground track spacing of 7-8 km and a along-track resolution of approximately 300 m in the SAR Interferometric (SARIn) mode.
This potentially enables Cryosat to sample more lakes and obtain more reliable lake levels in near-shore regions.
Still, the accuracy and robustness of standard level 2 data is limited due to pollution of the waveform in near-shore regions.
We propose therefore to actively extract the water surface by a novel retracking algorithm based on cross-correlation of observed level 1b waveforms with a generic simulated waveform.
As a result, we obtain multiple elevations per waveform.
As the water level is contributing to each consecutive waveform over the lake, it can efficiently be extracted using a majority voting scheme.
By comparing the lake levels to those obtained with Jason-2 over Lake Nasser an RMS of differences of 0.30 m is found.
After this validation step, we applied the procedure to lakes in Tian Shan and Tibet.
In these areas 125 lakes are measured at least four times in two years, of which 30 are even sampled ten times in two years and 16 more than twenty times.
For the ones sampled more than twenty times an additional slope in the water surface could be estimated, which can be an effect of prevailing winds or errors in the geoid model.
Ultimately, we found a negative water balance in the natural lakes in Tian Shan and a positive balance in Tibet between February 2012 and February 2014.
These results clearly demonstrate the potential of Cryosat, as previous radar altimetry missions were only able to sample about 70 lakes over the same region.