衰减
峰值地面加速度
地质学
震级(天文学)
加速度
地震学
地震动
强地震动
大地测量学
光谱加速度
非线性系统
物理
天体物理学
光学
经典力学
量子力学
作者
Bin Zhang,Xiaojun Li,Yanxiang Yu,Jize Sun,Mianshui Rong,Su Chen
标识
DOI:10.1016/j.jseaes.2023.105853
摘要
Based on strong motion and seismic velocity records from small-to-moderate earthquakes in the capital circle region of China, we developed a new ground motion prediction model (GMPM) that incorporates magnitude, geometric attenuation, anelastic attenuation, and linear/nonlinear site response terms. The GMPM was subjected to evaluation through residual analysis. We compared the median predictions of the GMPM with those of two commonly used local GMPMs, as well as the observed ground motions in this region. Furthermore, we conducted an investigation into the differences between the GMPM and four NGA-West2 models. Additionally, we explored certain model parameters that could potentially explain these differences. The results indicate that the GMPM effectively captures the magnitude-dependent attenuation and nonlinear average effect of soil sites. The GMPM demonstrates the capability to predict horizontal peak ground acceleration (PGA), peak ground velocity (PGV), and spectral acceleration (SA (T = 0.01–3.0 s)). The model is applicable for surface wave magnitudes (MS = 3.1–5.1), hypocentral distances (Rhyp = 10–200 km), and VS30 = 103–1070 m/s in the capital circle region of China. There are significant differences in anelastic attenuation at rock sites between southern California and the capital circle region of China. These differences decrease with increasing periods and largely diminish in far sources and soil sites. This is primarily because the effect of anelastic attenuation is neutralized by the greater site amplification of ground motion in southern California, especially over medium and long periods. The difference in ground motion observed in small-to-moderate earthquakes across different regions is predominantly influenced by long-distance attenuation and dissimilarities in site responses.
科研通智能强力驱动
Strongly Powered by AbleSci AI