On the quality of drought characteristics by GPS and GRACE signal errors

In recent years, drought events have become more frequent and severe, affecting human life, the environment, and industry. As a result, monitoring drought characteristics such as patterns, occurrences, intensity, categories, and duration presents a crucial challenge for scientists. These characteristics are usually estimated using hydrological or climate models, which, however, frequently fail to capture actual changes. Consequently, droughts are under-, overestimated or not captured. As remote sensing advances, a near real-time drought assessment would be successfully enabled using data provided by geodetic techniques such as the Gravity Recovery and Climate Experiment (GRACE) and the Global Positioning System (GPS). However, the limitations in GRACE and GPS techniques, data products or their quality, such as the spatial resolution, leakage effect, background models used for GRACE observations processing or systematic errors of GPS technique may face limitations in accurately capturing essential information on drought characteristics. In our study, we assess the impact of errors embedded in GRACE and GPS data on determined droughts. We calculate uncertainties of the Drought Severity Index (DSI) determined from GRACE-derived and GPS-observed vertical displacements. We investigate a number of ways to designate errors, starting with spherical harmonic coefficients errors, TWS errors associated to gridded GRACE mascons, errors in the positions of permanent GPS stations, by GRACE TWS variance-covariance matrices to errors in combining field using the Three-Corner-Hat (TCH) method. We find that maximum error values occur in nearly 30% of drought periods, showing that they are over- or underestimated by geodetic data. For the variance-covariance method, uncertainty of DSI determined from GRACE are identical for the entire European region. On the other hand, we observe that uncertainty of DSI determined from GRACE for both SH errors and mascon TWS errors are coherent in time. Values of GPS-DSI uncertainty are mostly close to zero, although we also identify significant peaks in series over drought and flood periods as sensed by GRACE-DSI. The results obtained for several different methods of error assessment are the next step in examining the reliability of drought characteristics, which can be valuable for decision makers.