Refinement of ICESat-2 derived inland water surface levels with the TG20 local geoid model: In the case of Türkiye lakes

Abstract

This study examined the contribution of the T & uuml;rkiye Geoid 2020 (TG20) local gravitational model to water height estimation using ICESat-2 ATL13 data. We tested eight separate lakes in T & uuml;rkiye to achieve this aim. The EGM2008 values were determined by a comparison of the orthometric heights derived from ATL13 data and the ellipsoidal heights derived from ATL03 data. Subsequently, the corresponding latitude and longitude values within the local gravitational model were obtained and integrated into the ICESat-2 data. The root mean square error (RMSE) for the unfiltered and filtered ATL13EGM2008 data was 0.66 m and 0.53 m, respectively. The filtering method in the ATL13TG20 data resulted in an average improvement of the RMSE from 0.62 m to 0.27 m. The local gravity model refinement tested in 8 different lakes showed an RMSE improvement ranging from 17% to 71%. This study demonstrates that local gravity models significantly improve the vertical accuracy of ICESat-2.