Mechanisms of groundwater recovery and land subsidence mitigation in a piedmont plain

Abstract

Groundwater recovery and land subsidence mitigation have been widely observed, but their mechanisms remain poorly known. Here, the northeastern Beijing piedmont plain is used as a natural lab to explore the mechanisms based on data of groundwater level and InSAR. Time series of groundwater level in two decades show groundwater recovery after the operation of the South-to-North Water Diversion Project since 2014. By using an explainable machine learning method, we find pumping restriction has contribution to groundwater recovery in the whole plain, whereas managed aquifer recharge (MAR) has effect in the impact zone with a width of 12 km along the river. Due to the superposition of MAR, groundwater recovery inside the impact zone is generally higher than that outside the impact zone. The higher groundwater recovery inside the impact zone in Aquifer I is a result of direct infiltration of MAR, whereas in Aquifer IV is caused by the loading effect of increased storage in shallow aquifers resulting from MAR. InSAR displacement data reveals obvious land subsidence mitigation since 2018, in the form of reduced subsidence rate or land rebound. The area of land rebound accounts for 3 % during the stage from 2018 to 2020, and increased to 47 % during the stage from 2021 to 2022. The rate of land rebound from 2021 to 2022 has a linear correlation with the proportion of sand layers, and land rebound occurs when the proportion of sand layers exceeds 50 %, which indicates the importance of elastic expansion of sand layers in land rebound.