Capability of pulse-limited satellite radar altimetry to monitor inland water bodies

Abstract

Pulse-limited satellite altimeters were originally designed for oceanographic observations but they have been extended to monitor inland water bodies. So far, studying water level variations of inland water bodies, e.g. lakes, has been a challenge for this type of altimetry in terms of data quality. The returned altimetry waveforms can be seriously contaminated by topography and environmental error sources. Retracking is an effective method against this contamination to improve the accuracy of range measurement and, consequently, to determine more accurate water level. In addition, the design of an optimal retracking algorithm appropriate for a specific inland water body is very important in this respect. In this study we processed 1 Hz Geophysical Data Record (GDR) of Envisat RA2 altimetry data by on-board tracker and retrackers. We also analyzed 18 Hz data of this mission, i.e. Sensor Geophysical Data record (SGDR), with respective different retrackers.First we processed GDR data to determine water level variations from ALL, MEDIAN and MEAN values of water level in each satellite pass over an inland water body. In this step we analyzed to find the best on-board tracker and retrackers. In the next step, the whole waveform, called full-waveform, was processed to estimate retracked water level variations using OCOG, Threshold and β-parameter retrackers. Then we assumed that the reflecting surface inside the radar foot print is a complex surface with different responses. Therefore a given waveform was considered as a combination of a number of small waveforms, called sub-waveform. Each sub-waveform was processed by all of the mentioned retrackers to determine water level variations. The largest salt lake in the Middle East, Urmia Lake, has been selected as a testing area in this study. We found out that between on-board tracker and on-board retrackers the MEDIAN values, processed by ice-1 retracker, provide the most accurate water level variations. Finally the result of different retracked water level were compared with ice-1 retrackers, and with available in-situ gauge data. Our analysis shows that retracking on the sub-waveform outperforms the retracking on the full-waveform. The minimum RMS, 18 cm, was achieved by sub-waveform, retracked by Threshold 50% algorithm. Therefore sub-waveform retacked by threshold 50% is the best retracking scenario to retrieve the water level variations of Urmia Lake.