Regional seismic topographic effects and seismic landslide hazard assessment during the 2023 Ms 6.2 Jishishan Earthquake in Gansu, China

On December 18, 2023, the Ms 6.2 Jishishan earthquake in Gansu Province triggered widespread landslides, collapses, and slope failures. Recent seismological studies indicate that the event was caused by an east-dipping fault rupture. The extensive ground motion recordings obtained from this earthquake provide a solid foundation for refined assessments of coseismic landslide hazards, incorporating both fault rupture characteristics and regional topographic amplification effects. This study utilized acceleration records from strong-motion and intensity monitoring stations across Gansu and Qinghai provinces. Through systematic baseline correction to eliminate DC offsets, key ground motion parameters — including Arias intensity and peak ground acceleration (PGA) —were derived. Using nonlinear least squares regression, we developed topography-sensitive attenuation models for both parameters, incorporating fault distance, VS30 site conditions, and topographic relief as predictive variables. Subsequently, the Newmark cumulative displacement model was applied to assess coseismic landslide hazards associated with the Jishishan earthquake. The modeled displacements were systematically compared with the spatial distribution of coseismic landslides near the epicenter obtained, obtained through remote sensing interpretation. Key findings reveal a quasi-linear relationship between topographic amplification coefficients and terrain relief. Specifically, statistical analysis demonstrates that every 10-meter increase in topographic relief lead to amplification factors of 1.18 for Arias intensity and 1.09, respectively for PGA, indicating a slightly stronger topographic amplification effects of Arias intensity. The proposed hazard assessment methodology, which explicitly incorporates topographic amplification into ground motion prediction, effectively captures the regional distribution patterns of seismically induced landslides. However, the accuracy of local-scale predictions still requires enhancement through the acquisition of high-resolution lithological and topographic data.