海原活动断裂毛毛山－老虎山段现今震间变形：是蠕滑还是震后变形集中？

Along the Maomao Shan – Laohu Shan segment of the Haiyuan active fault, InSAR measurements showed a significantly interseismic strain localization across the fault, mainly near its connection to the rupture zone of the 1920 Haiyuan 8.5 great earthquake. Previous studies have proposed that the observed interseismic strain pattern is resulted from fault creeping, furthermore an aseismic creeping with its rate approximately consistent to the long-term slip rate of Haiyuan fault over several seismic cycles. Around the same segment of the active fault, two moderate strong earthquakes with magnitude of 5.8 and 5.6 had occurred in 1990 and in 2000, just ~10 years before the SAR data acquisition. These earthquakes, including a pair of M~8 great earthquakes in 1920s, clearly should cause post-seismic deformation around a certain place of the fault, where the interseismic strain will be time-dependently localized. . To better calibrate if the interseismic strain localization were result from creeping offset of the fault, we will deploy several dense continuous GPS sites along a key profile to precisely capture the velocity changes across the fault within a distance of 1-5 km. Then, based on available seismological observations and seismic waveforms recorded by local and global seismic network, we will invert the coseismic rupture model of both the 1990 and 2000 M~5.8-5.6 earthquakes, and the 1920s great earthquakes. Constrained by the cGPS observations and the coseismic rupture model, we will finally build a three-dimensional viscoelastic model to explore the possible mechanics of interseismic strain localization around the western portion of the Maomao Shan – Laohu Shan segment, by testing the available rheological parameters of the crust and upper mantle. The main objective is to decipher if the observed interseismic strain localization were result from aseismic creeping or from reloading of post-seismic deformations of the past earthquakes. We believe that such a study is crucial for better evaluating the potential hazard of earthquake especially around the Maomaoshan-Laohushan segment. In addition, the proposed work may also benefit to explore the mechanism of nonlinear behavior for interseismic strain accumulation.

在海原活动断裂的毛毛山－老虎山段，InSAR测量表明其在邻近1920年海原8.5级地震一侧有明显跨断层变形集中。前人认为该震间变形集中源于断层蠕滑，且近于无震蠕滑。同样沿该断层段落，1990年以来分别发生了5.8和5.6级走滑型地震。显然，这些（包括早期8级）地震的震后变形也会导致一定时间窗口、特定部位跨断层变形集中。本申请用跨断层组合短基线连续GPS测量来加确近断层地表运动方式；结合有仪器记录地震的波形和地表破裂数据，用模拟退火方法反演其同震破裂模型；基此，以现有地壳上地幔流变学参数及变化范围为约束，用非连续介质三维粘-弹性有限元技术来模拟海原断裂震间变形集中与驰去地震之震后变形的时空关联，以厘清毛毛山－老虎山段的变形到底源于无震蠕滑，还是有关地震震后变形集中。这关乎到评价该断裂关键部位会或不会发生地震之重大问题，同时对认识活动断裂震间变形累积非线性特征及动力学机制也有重要理论意义。

近10年来，通过应用空间大地测量InSAR技术，一些学者提出海原大型活动走滑断裂的西段—老虎山-毛毛山断层现今震间变形主要以~5mm/yr速率蠕滑。这暗示老火山-毛毛山断层没有明显弹性能积累，断层不会有强地震危险性。为了为回答“会”或“不会”地震这一核心科学问题提供充分证据，本项目提出开展跨断层应变的高精度时序观测，并通过反演和模拟从动力学上给出该断层可能发生特定震间变形的机制。于是，课题组从跨断层短基线连续GPS测量、历史大地震震源模型回溯、地震断裂多时空尺度动力学模拟等方向开展了大量研究。四年来，我们在关键部位布设了10个高频（1Hz）连续GPS观测基站，其基线长度在跨断层近场~10km、远场~100-120km，所投入的仪器、建站、维护、数据采集等日常经费近200万。通过收集同震地表破裂、地震烈度和国际台站在其地震地震图，对1920年M~8海原历史地震的同震破裂模型开展反演。在观测和反演数据约束下，通过构建三维粘弹性有限元模型，模拟了1920年M~8海原地震对老虎山-毛毛山断层在不同时-空窗口的荷载规律。项目研究的主要结论：1）反演了1920年M~8海原地震莫列模型，并获得同震走滑位移在断裂中段~15km深度范围14-18m。类似的研究对认识历史地震破裂模型是国内首次，具有示范意义。2）在跨断层小于~10km的连续GPS时序观测中，没有发现明显的，前人提出的断层蠕滑信号。3）通过多时空尺度模拟，发现1920年M~8海原地震至今还在对老虎山-毛毛山断层加载，且加载速率可以和远场构造加载速率相比。这些结论不仅是全新的，同时启示了更多科学问题：1）如果InSAR研究提出的老火山-毛毛山段蠕滑不是源于信号误差，则它可能暗示断层变形的某种瞬态行为，我们至今无解。2）从大量动力学模拟发现了大陆内部断层发生8级地震的一些重要线索，将有利于聚焦课题组后续研究。