Probabilistic seismic hazard analysis for the northern segment of the North-South Seismic Belt in China based on improved spatial smoothing and fault source model integration
-
Abstract
This study conducted probabilistic seismic hazard analysis (PSHA) for the northern segment of the North-South Seismic Belt, which is characterized by intense crustal deformation and frequent large earthquakes. The research integrated historical earthquakes, active faults, and paleoearthquake data to construct a composite seismic source model. A logic tree approach was employed to quantify uncertainties in seismic activity and ground motion models, enabling comprehensive seismic hazard assessment in the region. To accurately characterize the spatial distribution of seismic activity, improvements were made to the fixed-radius circular spatial smoothing model (CSSM) and adaptive spatial smoothing model (ASSM). These enhancements consider the spatiotemporal completeness of historical earthquake data through maximum likelihood estimation for optimal parameter determination. Results demonstrated that when earthquake catalog data were limited, the cross-validation method for determining optimal correlation distances in CSSM proved more robust. Evaluations using probability gain, concentration plots, and receiver operating characteristic (ROC) curves confirmed that both improved models effectively captured spatial seismic characteristics, with ASSM exhibiting superior predictive performance. Significant discrepancies emerged among different models: single point-source models underestimated hazards, whereas combined simple fault-source models showed markedly increased hazard levels. The characteristic fault-source model provided more reasonable estimates for large earthquake hazards, particularly for faults approaching their recurrence intervals for major seismic events. Comprehensive analysis integrating improved spatial smoothing point-source, simple fault-source, and characteristic fault-source models identified high-hazard faults in the region, including the Huangxianggou, Zhangxian, and Tianshui segments of the Xiqinling northern edge fault; Maqin-Maqu segment of the Dongkunlun fault; Longriqu fault; Maoergai fault; Elashan fault; Riyueshan fault; eastern segment of the Lenglongling fault; Maxianshan segment of the Northern Margin fault of Maxianshan; and Maomaoshan-Jinqianghe segment of the Laohushan-Maomaoshan fault. This study indicates that high-hazard active faults should be given priority in monitoring and mitigation, and that urban planning and critical engineering projects in the affected areas should accordingly adopt higher seismic-design standards.
-
-