ABSTRACT This study investigates the dynamic changes in land subsidence in Zhengzhou—a densely populated city in the North China Plain severely affected by this phenomenon—following the extreme precipitation event of July 20, 2021. An innovative methodology was developed, combining slope units and road network segmentation to partition evaluation units (EUs), thus more accurately capturing the complexity of terrain and human activity impacts. SBAS‐InSAR and Sentinel‐1A data were used to calculate cumulative land subsidence between 2020 and 2022. Building on this, a novel analytical framework was proposed, integrating the Deep Forest algorithm with SHapley Additive ExPlanations to assess the dynamic influence of various driving factors on maximum monthly subsidence for each EU, before and after the extreme precipitation event. The results indicate that land cover, temperature, normalized difference vegetation index (NDVI), and groundwater depth are critical factors affecting land subsidence in Zhengzhou. In the monthly models, the relative importance of NDVI and temperature was higher during the period associated with the July 2021 extreme precipitation event than during the corresponding periods in the reference years. Furthermore, a complex nonlinear relationship between groundwater depth and land subsidence was observed. This data‐driven approach effectively elucidates the dynamic changes in subsidence and its underlying drivers, while highlighting the need for cautious interpretation of post‐event causality and uncertainty.
Peng et al. (Fri,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: