Item

Abstract

The North China Plain (NCP) is a very important region, which is related to national water security and food security. Clarifying its water storage changes and spatial and temporal evolution patterns is a prerequisite to ensure sustainable development and diversified utilization of water resources. Due to the existing observation techniques, data processing methods, and data processing errors, there are differences between different GRACE solutions. Using a single GRACE solution may yield unreliable results when the true value cannot be confirmed. Therefore, to address the problem of unequal accuracy and systematic bias of GRACE data from different sources, this study proposes to use generalized three–cornered hat method (GTCH) and Bayesian model averaging method (BMA) to fuse multi-source GRACE data , and construct a unified high-precision water storage change model for the NCP. The numerical results indicate that (1) compared with other data fusion methods, the BMA method based on the a priori information of GTCH is more suitable for the NCP, and can obtain more accurate and reliable results of water storage changes; (2) for small and medium scale areas like the NCP, the fused GRACE spherical harmonic solution is more accurate than the MASCON solution; (3) the fusion results can more clearly show the spatial and temporal migration characteristics of groundwater storage distribution in the NCP.