Permanent displacement characteristics of landslide considering frequency and energy behaviors of historical seismic ground motions

Landslides often occur along expressways in earthquake-prone areas, leading to casualties and property loss. Most previous studies applied the simplified Newmark method to assess landslide susceptibility at a regional scale, ignoring the characteristics of local ground motions. In this study, we investigated permanent displacement characteristics of three potential landslide areas along the Dayong expressway using the rigorous Newmark method, considering the frequency and energy characteristics of the historical seismic waves. First, we analyzed the frequency characteristics of historical earthquakes in the study areas using the fast Fourier transform method and adopted them as input ground motions in calculating the Newmark displacement. Next, we computed the critical accelerations of the three study areas. Parameters such as the elevation, soil layer thickness, and soil strength required for critical acceleration calculations were obtained using high-precision (approximately 3 cm) unmanned aerial vehicle mapping technology, drilling operations, and laboratory tests. Finally, the Newmark displacement was calculated, and landslide susceptibility of the study areas was evaluated. The results showed that the predominant frequency of historical seismic waves in the study areas was 0.38–3.36 Hz, indicating low-frequency characteristics. The results also indicated that the frequency and energy characteristics of the seismic ground motions significantly influenced the Newmark displacement. The maximum Newmark displacement under different waveforms at Damieju ranged from 2 to 18 cm. Under the most dangerous conditions, all three areas showed potential for landslides. The maximum Newmark displacement positively correlated with the energy of the predominant frequency. Large energy and low frequency characteristics of seismic wave corresponds to a large displacement. The simplified Newmark method in regional landslide susceptibility analysis did not reflect this effect and may be unsuitable to determine the size of landslides.