Frontiers in Environmental Science (Aug 2022)
Reconstructing the Historical Terrestrial Water Storage Variations in the Huang–Huai–Hai River Basin With Consideration of Water Withdrawals
Abstract
The Huang-Huai-Hai River Basin in eastern China has suffered from severe water scarcity during recent decades due to the effects of climate change and human activities. Quantifying the changes in the amount of terrestrial freshwater available in this region and their driving factors is important for understanding hydrological processes and developing a sustainable water policy. This study proposed an ensemble learning model to reconstruct historical variations in the terrestrial water storage (TWS) of the Huang-Huai-Hai River Basin, China. The model was trained using the observations of the variations in TWS from the Gravity Recovery and Climate Experiment mission (GRACE) satellites, climatic driving, and human withdrawal datasets produced on a monthly scale. The variations in the reconstructed TWS were compared with the results of several land surface and hydrological models with a variety of in situ measurements of the soil water content. The contributions of the climate and human activity to the ensemble learning model were also quantified. The results show that the proposed approach generally outperforms the land surface and hydrological models examined in this study, matches the patterns in the GRACE solutions, and reconstructs past changes in TWS, which are consistent with the GRACE observations. Climatic variables are the most important in the ensemble learning model, with precipitation over the prior month being a critical factor. The model that includes human intervention tends to perform better than without it. Irrigation, industry, and domestic water withdrawals contribute equally to the model. This study provides a flexible and easily implementable model that can bridge the gap between GRACE observations and past changes in TWS. The model is applicable in areas with intense human activities, and the results have the potential to be assimilated into and enhance hydrological models.
Keywords
