Near-real-time prompt assessment for regional earthquake-induced landslides using recorded ground motions

Abstract

Earthquake-induced landslides (EQILs) can cause severe economic losses and casualties. The prompt assessment of EQILs can provide an important reference for emergency rescue. Based on recorded ground motions and the Newmark sliding block method, a near-real-time prediction method for EQILs is proposed, and the corresponding prompt assessment system is developed. First, the seismic network is used to obtain the ground motion records near the epicenter immediately after an earthquake. The inverse distance-weighted interpolation and the continuous wavelet transform method are then adopted to generate the ground motion input field based on recorded ground motions. Subsequently, the slope of the target region is determined based on a global digital elevation model. Further, the mapping relationship between the Global Lithological Map database and the engineering lithology classification is established to determine the geological parameters for the target region. Finally, the predicted EQIL results are obtained using the Newmark sliding block method. The commonly used Newmark displacement models and the Newmark sliding block method are compared in this work to demonstrate the necessity of time-history analysis. The proposed method is validated using data from the 2011 Tōhoku earthquake and the 2016 Kumamoto earthquake. The 2018 Hokkaido earthquake is used as a case study to illustrate the detailed procedures and advantages of the proposed method. With high efficiency and convenience, the characteristics of the ground motion can be fully considered using the proposed method, which provides an important reference for post-earthquake emergency rescue.