Abstract
Jincheng City's mining areas have long been plagued by surface subsidence, posing significant threats to local residents' safety and impacting the region's economic and social stability. Understanding and effectively monitoring the driving factors and mechanisms of surface subsidence are crucial for devising scientific prevention measures and promoting the sustainable development of mining areas. This article aims to comprehensively reveal the large-scale surface subsidence phenomenon in Jincheng City's mining clusters by utilizing advanced remote sensing technology and machine learning models, identifying its main driving forces, and predicting future subsidence trends to provide scientific evidence for geological disaster prevention in mining areas. The study employs Multi-Temporal Interferometric Synthetic Aperture Radar (MT-InSAR) technology, using both Permanent Scatterer Interferometric SAR (PS-InSAR) and Small Baseline Subset Interferometric SAR (SBAS-InSAR) techniques for cross-validation, to confirm the existence of surface subsidence. Further, by integrating Variational Mode Decomposition (VMD), Singular Spectrum Analysis (SSA), and Long Short-Term Memory (LSTM) networks, a high-precision time series prediction model (VMD-SSA-LSTM) was developed. The results indicate that from 2018 to 2021, the surface subsidence rates in Jincheng City ranged from − 34 to 34 millimeters per year, with significant variations in subsidence levels across different areas. Gaoping City exhibited the highest subsidence, with rates ranging from − 34 to 5 mm per year, while Yangcheng County showed the most pronounced subsidence changes. These variations are primarily attributed to mining activities, land use changes, and adverse geological conditions in Jincheng City. This study unveils the large-scale surface subsidence phenomenon in Jincheng City's mining clusters, marking the first comprehensive ground deformation monitoring analysis of small mining clusters across four cities in Jincheng. The development of a high-precision surface subsidence prediction model provides new insights for scientifically understanding geological disasters in mining areas. These findings are significant for formulating effective geological disaster prevention measures and land management policies.