Characterizing the Seasonal Hydrological Loading Over the Asian Continent Using GPS, GRACE, and Hydrological Model

Abstract

Based on combined data of the Global Positioning System (GPS), Global Land Data Assimilation System (GLDAS), and Gravity Recovery and Climate Experiment (GRACE), the seasonal hydrological loading over the Asian continent is characterized in this study. The hydrological loading effects over the Asian continent display strong latitude dependence. The significant hydrological loading effects appear at the GPS stations situated in the coastal areas, some regions near large rivers and lakes, and high-latitude areas in Russia, as evidenced by the fact that a large root mean square (RMS) and high percentage of the variance related to the annual signal modeled by singular spectrum analysis (SSA) for each measurement are cumulated at the stations located in these regions. In contrast, the hydrological loading effects are not pronounced in mid-latitude areas of the Asian continent (e.g., Central Asia, northern and plateau regions of China), which is due to the high topographical variability and scarce water resources in these regions. Then, the cross wavelet transform (XWT) is used to quantify the consistency between different data sets. For the data sets of GPS/GLDAS, the XWT-based semblance for 64% of the stations reaches above 0.8, while it reaches above 0.8 for 48% for the data sets of GPS/GRACE, indicating that the data sets of GPS/GLDAS present better consistency. In addition, we also discuss the effects of hydrological loading on GPS observations from the RMS value, noise characteristic, and velocity uncertainty. After applying the hydrological loading correction, the RMS values of almost all GPS observations are reduced with different amplitudes, implying that the hydrological loading correction can reduce the RMS values of most GPS observations in the Asian continent. Meanwhile, the variations of noise and velocity uncertainty suggest that hydrological loading has changed the noise characteristic of almost all GPS observations, and thus lead to the overestimation of velocity uncertainty.