Multistage rupture mechanism of gypsum mine collapse seismicity revealed by moment tensor inversion and surface deformation monitoring

Moment Tensor analysis plays a pivotal role in characterizing both tectonic and mining-induced earthquakes, particularly for understanding rupture dynamics and associated hazards. Between 2015 and 2020, 4 high-energy mining-induced earthquakes (up to ML4.0) occurred at some gypsum mines in Shandong Province, China, causing severe casualties, surface subsidence, and infrastructure damage. Through Moment Tensor inversion and decomposition, we systematically quantified the source mechanisms of these events, revealing distinct collapse characteristics dominated by negative crack components. Sensitivity tests under different velocity models and station configurations confirmed the robustness of the inversion results. Integrating seismic analysis with geological surveys and accident investigations, we reconstruct a multi-stage rupture process for the 2015 Pingyi event (Event 01): progressive degradation of gypsum pillars and roof cavities leads to sudden brittle failure of overlying limestone strata, producing the mainshock and subsequent collapses. The optimal source depth (0.2–0.3 km) from Moment Tensor inversion closely matches field-observed roof collapse depths. InSAR analysis reveals multiple localized subsidence centers further support the collapse-type source mechanisms. Our findings highlight the critical importance of seismic monitoring and analysis technologies for detecting and characterizing mining-induced earthquakes. Even in scenarios lacking immediate field investigations, these methods provide essential real-time data that can be directly support emergency response and disaster mitigation efforts.