Xiaoli Su, Zhicai Luo, Zebing Zhou

Institute of Geophysics, School of Physics, Huazhong University of Science and Technology, Wuhan, China

Received 31 July 2017, Revised 7 February 2018, Accepted 25 March 2018, Available online 27 March 2018.

Abstract Knowledge of backscatter change is important to accurately retrieve elevation change time series from satellite radar altimetry over continental ice sheets. Previously, backscatter coefficients generated in two cases, namely with and without accounting for backscatter gradient (BG), are used. However, the difference between backscatter time series obtained separately in these two cases and its impact on retrieving elevation change are not well known. Here we first compare the mean profiles of the Ku and Ka band backscatter over the Greenland ice sheet (GrIS), with results illustrating that the Ku-band backscatter is 3 ∼ 5 dB larger than that of the Ka band. We then conduct statistic analysis about time series of backscatter formed separately in the above two cases for both Ku and Ka bands over two regions in the GrIS. It is found that the standard deviation of backscatter time series becomes slightly smaller after removing the BG effect, which suggests that the method for the BG correction is effective. Furthermore, the impact on elevation change from backscatter change due to the BG effect is separately assessed for both Ku and Ka bands over the GrIS. We conclude that Ka band altimetry would benefit from a BG induced backscatter analysis (～10% over region 2). This study may provide a reference to form backscatter time series towards refining elevation change time series from satellite radar altimetry over ice sheets using repeat-track analysis.

doi: https://doi.org/10.1016/j.jog.2018.03.007