Abstract:

An M_S7.4 earthquake occurred in Maduo County, Qinghai Province on May 22, 2021, which is the largest magnitude event following the 2008 Wenchuan M_S8.0 earthquake in China. A total of 54 aftershocks occurred up until 3 June, 2021, including 15 M4.0~4.9 earthquakes and one M5.0 earthquake. These earthquakes primarily nucleated at a depth of 8~10km, and ruptured a NNW-SSE-trending fault with a length of~170km. Based on the post-earthquake field damage survey, the seismogenic fault information, aftershock spatial distribution, as well as the strong-motion and seismic-intensity observation recordings, the Ministry of Emergency Management of the People’s Republic of China officially released the macroseismic intensity map for this M_S7.4 earthquake. The maximum intensity reaches X(Chinese seismic intensity scale)in a narrow area close to the fault. Due to the strong earthquake-resistance and the low population density, the Maduo M_S7.4 earthquake causes few casualties and some injuries only.
The sparse strong-motion observation stations in this region, distributed mainly to the northeast of the epicenter, collected a few number of strong-motion recordings only at far field(over 150km to epicenter)during the M_S7.4 event. In order to evaluate the spatial distribution of ground motion intensity, the stochastic finite-fault method for simulating three-component(two horizontal and one vertical components)seismic ground motions were applied to reproduce the ground motions at 4 461 dummy observation positions during the Maduo M_S7.4 earthquake. The inverted source rupture process from USGS and Zhang Yong of Peking University, source slip distribution model from Institute of Geology of China Earthquake Administration, and the source kinematic rupture models stochastically generated were used separately to represent the source model of this earthquake in the simulations. The average stress drops for these source models are mainly in the range of 3~4MPa. The V_S30 scaling was considered to represent the amplification effects of the local site conditions. The peak ground accelerations(PGAs)and velocities(PGVs)for the simulated ground motions were first compared with those from the far-field observations and the predicted medians by the ground motion prediction equation applicable for Qinghai-Tibetan seismic zone. The PGAs and PGVs for the near-fault simulations are generally up to ~300cm/s² and ~30cm/s, respectively. The simulated PGAs and PGVs are in good agreement with those predicted medians, and well stand for the distance decay. However, the simulated peak values are, in general, smaller than those far-field observations. Moreover, the acceleration and velocity time histories, and the 5%-damped pseudospectra accelerations (PSAs)at some typical dummy observation positions, e.g., epicenter, Maduo County, urban district of Guoluo, were provided. It was found that the amplitudes of the time histories and the PSAs at vertical component are much lower than those at both horizontal components for the simulated recordings. The amplifications on the horizontal ground motions caused by the local site condition, and the much sharp source spectral decay rate for P-wave may by the possible reasons.
Finally, the simulated three-component ground motions were applied to calculate the seismic intensities. The spatial distribution of seismic intensity was then displayed. Results indicated that: 1)the maximum seismic intensity is up to Ⅸ or Ⅷ dependent on the specific source model; 2)the principal axis of seismic intensity is well consistent with the strike of the seismogenic fault; 3)the distribution of meizoseismal area is just concentrated in a narrow area along the ruptured fault; and 4)the isoseismal lines are irregular ellipses due to the considerations on the local site conditions. Compared with the officially released macroseismic intensity map, the seismic intensities based on the simulated ground motions are generally underestimated by about 1~2 degrees. We infer that the stress drop statistically averaged used for simulations bears the responsibility. Considering the uncertainty of the seismic stress drop, the uncertainty of the seismic intensity based on the simulated ground motions is about plus or minus 1.2. At the end, we re-simulated ground motions using the source rupture model of USGS, while the value of stress drop was set to mean plus one standard deviation(about 13.59MPa). In this case, the seismic intensity based on simulated ground motions show good agreements with the macroseismic intensity map, especially for those regions with intensity equal to or greater than Ⅶ, and the maximum intensity reaches Ⅹ.

Key words: Maduo M_S7.4 earthquake, stochastic finite-fault method, ground motion simulation, stress drop, seismic intensity

摘要：

2021年5月22日青海省玛多县发生M_S7.4地震, 这是2008年汶川M_S8.0地震后中国发生的震级最大的地震。 由于该地区强震动观测台站稀疏, 在此次地震中仅获得少量远场强震动记录。 为评估此次地震的地震动影响场, 文中利用青海玛多M_S7.4地震的多个震源破裂模型、 断层滑动模型以及震源破裂随机模型, 基于随机有限断层三维地震动模拟方法给出了震中附近一定区域内4 461个虚拟观测点的模拟地震动三分量加速度时程。 与远场观测记录峰值(PGA、 PGV)及青藏地震区地震动峰值预测方程的预测中位值进行对比, 发现模拟记录峰值及其距离衰减与预测中位值基本一致, 但远场模拟记录峰值普遍低于观测值。 文中给出了典型场点(震中、 玛多县城、 果洛藏族自治州市区等)的加速度时程、 速度时程及拟加速度反应谱(PSA)。 最后基于模拟记录峰值计算给出了地震烈度的空间分布。 基于模拟记录的地震烈度长轴方向与破裂面走向一致, 极震区沿破裂面分布, 局部场地效应导致烈度等震线呈不规则的椭圆形。 与正式发布的地震烈度图进行对比可知, 基于模拟记录的地震烈度整体上偏低1~2度, 极震区烈度仅为Ⅸ或Ⅷ度。 我们推断模拟中采用的统计平均应力降可能导致模拟地震动普遍偏小, 考虑地震应力降的不确定性及其与地震动不确定性的关系, 模拟记录估计的地震烈度的标准差约为1.2。

关键词: 青海玛多M_S7.4地震, 随机有限断层, 地震动模拟, 地震应力降, 地震烈度