Zhou Zheng, Hu Jiyuan, Wang Jiayao, Wang Lijun, Qiao Tianrong, Li Zhen, Cheng Shiyuan
College of Geography and Environmental Science, Henan University, Kaifeng, 475004, China.
Henan Industrial Technology Academy of Spatio-Temporal Big Data, Henan University, Kaifeng, 475004, China.
Sci Rep. 2025 Jan 31;15(1):3848. doi: 10.1038/s41598-025-87789-9.
Zhengzhou city (China) experienced relatively significant land deformation following the July 20, 2021, extreme rainstorm (7·20 event). This study jointly utilised Multi-temporal synthetic aperture radar interferometry (MT-InSAR), eXtreme Gradient Boosting (XGBoost), and hydrogeological analysis to quantitatively assess the extent and trends, as well as the causes of land deformation before and after the 7·20 event in Zhengzhou city. The findings detected three major subsidence zones and two uplift zones within the city. The most significant subsidence occurred in the northern part of Zhongmu (- 28 mm/year), the northwest of Xingyang (- 16 mm/year), and the western region of Gongyi (- 6 mm/year). Conversely, a notable uplift was observed in the central city district (13 mm/year) and Xinzheng Airport (12 mm/year). The accuracy assessment of in-situ measurements (GNSS and levelling) yielded an overall root-mean-square error (RMSE) of 2.2 mm/year and an R-square of 0.948. Subsequently, the feature evaluation results based on the XGBoost method suggest that road density and precipitation are the dominant factors affecting land deformation in the entire study area or in the subsidence and uplift zones individually. Nevertheless, the other five factors (groundwater storage, soil type, soil thickness, NDVI, and slope) also act on land deformation individually and are intricately intertwined with each other. Furthermore, hydrogeological analysis from six groundwater wells reveals a synchronous relationship between groundwater level decline and land subsidence. The building load analysis shows a significant correlation between build-up density and subsidence rates, especially for those severe subsidence areas, with the maximum correlation coefficient reaching 0.6312. Finally, the geographic patterns analysis of post-event demonstrated a northeastward trend in land deformation, with a gradual reduction of deformation impact from 2018 to 2022.
中国郑州市在2021年7月20日的极端暴雨（“7·20”事件）之后经历了较为显著的地面变形。本研究联合运用多期合成孔径雷达干涉测量技术（MT-InSAR）、极端梯度提升算法（XGBoost）和水文地质分析，对郑州市“7·20”事件前后地面变形的程度和趋势及其成因进行了定量评估。研究结果在该市探测到三个主要沉降区和两个隆起区。沉降最显著的区域位于中牟北部（-28毫米/年）、荥阳西北部（-16毫米/年）和巩义西部地区（-6毫米/年）。相反，在市中心区（13毫米/年）和新郑机场（12毫米/年）观测到明显隆起。原位测量（全球导航卫星系统和水准测量）的精度评估得出的总体均方根误差（RMSE）为2.2毫米/年，决定系数R²为0.948。随后，基于XGBoost方法的特征评估结果表明，道路密度和降水量是影响整个研究区域或沉降区与隆起区地面变形的主导因素。然而，其他五个因素（地下水储量、土壤类型、土壤厚度、归一化植被指数和坡度）也分别对地面变形产生作用，并且它们之间相互交织。此外，对六口地下水井的水文地质分析揭示了地下水位下降与地面沉降之间的同步关系。建筑荷载分析表明建筑密度与沉降速率之间存在显著相关性，特别是在那些沉降严重的区域，最大相关系数达到0.6312。最后，对事件后地理格局的分析表明，地面变形呈东北方向发展趋势，从2018年到2022年变形影响逐渐减小。
