Investigation of spatiotemporal characteristics of terrestrial water storage changes in the Yangtze River Basin using GNSS vertical displacement data

Continuous GNSS reference station network can monitor the vertical elastic crustal deformation caused by the change of surface hydrological loading in real-time. Therefore, using GNSS-observed hydrological loading vertical displacement to estimate the change in regional terrestrial water storage is an effective scheme. We collect the vertical displacement time series of 98 GNSS stations in the Yangtze River Basin from 2011 to 2020 and employ the Slepian basis function method to convert them into the corresponding displacement spectrum. We estimate the changes in terrestrial water storage in the Yangtze River Basin based on the mass loading theory and then analyze its spatiotemporal distribution characteristics and the hydrological driving mechanism by combining precipitation and data from GRACE and GLDAS. The GNSS-based inversion results and the changes in terrestrial water storage derived from GRACE and GLDAS all show obvious seasonal characteristics, and the spatial correlation coefficients (CCs) with GRACE and GLDAS in annual amplitude are 0.79 and 0.91, respectively. Three datasets all indicate a spatial pattern of dramatic change in water storage in the eastern and western parts and a subtle change in the central part. However, GNSS results show that the maximum annual amplitude of water storage change in the whole Yangtze River Basin is similar to 214 mm, which is significantly larger than the results of GRACE (similar to 121 mm) and GLDAS (similar to 107 mm). In addition, we investigate the spatial pattern and temporal characteristics of water storage changes in three subregions of the Yangtze River Basin (Jinsha River Basin, traditional upper reaches, and middle and lower reaches of the Yangtze River Basin). We find that the annual amplitudes of water storage change based on the three datasets show a decreasing trend from southwest to northeast in the Jinsha River Basin. The results of GNSS and GRACE show that the water storage changes significantly in the middle and lower reaches of the Yangtze River Basin. In the three sub-regions, the temporal CCs between GNSS time-varying results and GRACE, GLDAS, and precipitation data range from 0.48 to 0.84, among which the CC between GNSS and GLDAS in the Jinsha River Basin is 0.84. We also detect an irregular time-delay relationship between GNSS results and precipitation that GNSS peak time lags behind the precipitation by 2 months in the Jinsha River Basin and the traditional upper reaches, whereas there is almost no delay in the middle and lower reaches. The water storage surpluses and deficits estimated by GNSS results in the middle and lower reaches of the Yangtze River Basin correspond well with historical extreme hydrological events, implying that a sparse GNSS station network can capture the extreme hydrological conditions at the basin scale and provide a novel solution for the study of local-scale spatiotemporal variation of water storage.