Spatial and temporal analysis of daily terrestrial water storage anomalies in China

The spatial–temporal evolution of terrestrial water storage anomalies (TWSA) is crucial in monitoring floods and sustainable water management. Unlike monthly gravity models, daily models can obtain TWSA at daily resolution, which demonstrates advantages in monitoring short-term floods. Moreover, with sufficient observations it is possible to capture the temporal characteristics of TWSA. In this paper the TWSA of nine major drainage basins in China spanning from January 2003 to August 2016 are estimated. The spatial variations of the Yangtze drainage basin which is taken as example accurately reflect the 15 July, 2010 flood. The variation of Wetness Index (WI) agrees well with that of discharge of DaTong gauging station. Meanwhile, WI shows four days lead-time prior to the flood, which can be regarded as early warning indictor in ungauged basin. For the temporal analysis, noise characteristics of TWSA are assessed, which show that the optimal noise model is autoregression moving average noise (ARMA) but with different orders for different basins. With the optimal ARMA noise, the uncertainties of estimated parameters can reach up to 28 times that considering only white noise. Therefore, to get the comprehensive temporal features of daily TWSA, its time-correlated characteristics cannot be neglected.