A novel solution for outlier removal of ICESat altimetry data: a case study in the Yili watershed, China

Abstract

Due to the influence of cloud and saturated waveforms, ICESat data contain many contaminated elevation data that cannot be directly used in examining surface elevation and change. This study provides a novel solution for removing bad data and getting clean ICESat data for land applications by using threshold values of reflectivity, saturation, and gain directly from ICESat’s GLAS (Geoscience Laser Alteimeter System) 01, 05, and 06 products. It is found that each laser campaign needs different threshold compositions to assure qualified ICESat data and that bad data removal rates range from 9.6% (laser 2A) to 62.3% (laser 2B) for the test area in the Yili watershed, China. These thresholds would possibly be used in other regions to extract qualified ICESat footprints for land applications. However, it is recommended to use the steps proposed here to further examine the transferability of threshold values for other regions of different elevations and climate regimes. As an example, the resulting ICESat data are applied to examine lake level changes of two lakes in the study area.