Impact of spatial filtering to GRACE-FO-derived TWS changes: a case study of European river basins

Data provided by the gravity recovery and climate experiment (GRACE) and its successor follow-on (-FO) mission are noisy and thus an appropriate spatial filter is required to reliable assessment of the real (geo-)physical signals. This study applies traditional Gaussian smoothing approach and two different data-driven approaches, i.e.: 1) method of scale; 2) method of deviation to study the impact of spatial averaging of GRACE-FO data to leakage and total signal of total water storage (TWS) across major 24 European river basins. The greatest differences between filters are observed in regions indicating divergent TWS signals within small catchments. There are variations in trend values of ±1-2 cm/yr and amplitudes of up to 5 cm in the western and the northeastern European basins. Estimated true leakage represents at least 30% of total TWS signal. Moreover, TWS leakage error is at least 2-3 times larger for original than for both data-driven approaches.