Key points are not available for this paper at this time.
Over-exploitation of groundwater is the principal anthropogenic trigger for land subsidence in the North China Plain (NCP). With the promulgation of policies on groundwater extraction restrictions and bans, seasonal fluctuations and uplift have begun to occur in some typical subsidence areas in the NCP. In order to explore the land subsidence-rebound drivers, this study takes Dezhou City, a typical subsidence area in the NCP, as the study area and selects the main drivers in climate change: precipitation and evapotranspiration and the main driver in human activities: groundwater extraction, which together forms the background of the regional dualistic water cycle. In this background, this study selects the corresponding datasets and uses the improved Water Balance constraints Recurrent Neural Network (WB-RNN) method to establish the link between precipitation data, evapotranspiration data, groundwater extraction and groundwater level. Then uses the Gate Recurrent Unit-Convolutional Neural Network (GRU-CNN) method to establish the link between groundwater level and Interferometric Synthetic Aperture Radar (InSAR) deformation, and then obtains the link between the drivers of the dualistic water cycle and land subsidence. Comparative analysis and changing model inputs found that 1) Seasonal fluctuations in land subsidence-rebound in Dezhou City are mainly controlled by the groundwater levels in Aquifer Groups III and IV, with amplitudes up to 15 mm per year. 2) Groundwater level changes in different layers lag precipitation at different times. Excluding the influence of human extraction factors, the maximum ground surface deformation fluctuation caused by monthly 260 mm precipitation can be up to 21.44 mm. 3) The impact of short-term heavy precipitation caused by Typhoon Lekima and Doksuri in 2019 and 2023 can cause deformation fluctuations to a certain extent.
Wang et al. (Mon,) studied this question.