Abstract: The bedrock scarps are believed to have recorded the continuous information on displacement accumulation and sequence of large earthquakes. The occurrence timing of large earthquakes is believed to be correlated positively with the exposure duration of bedrock fault surfaces. Accordingly, cosmogenic nuclides concentration determined for the bedrock footwall can offer their times, ages, and slip over long time. In general, multiple sites of fault scarps along one or even more faults are selected to carry out cosmogenic nuclide dating in an attempt to derive the temporal and spatial pattern of fault activity. This may contribute to explore whether earthquake occurrence exhibits any regularity and predict the timing and magnitude of strong earthquakes in the near future. Cosmogenic nuclide ³⁶ Cl dating is widely applied to fault scarp of limestone, and the height of fault scarp can reach as high as 15~20m. It is strongly suggested to make sure the bedrock scarp is exhumed by large earthquake events instead of geomorphic processes, based on field observation, and data acquired by terrestrial LiDAR and ground penetration radar (GPR). In addition, it is better for the fault surface to be straight and fresh with striations indicating recent fault movement. A series of bedrock samples are collected from the footwall in parallel to the direction of fault movement both above and below the colluvium, and each of them is~15cm long,~10cm wide, and~3cm thick. The concentrations of both cosmogenic nuclide ³⁶ Cl and REE-Y determined from these samples vary with the heights in parallel to fault scarps. Accordingly, we identify the times of past large earthquakes, model the profile of ³⁶ Cl concentration to seek the most realistic one, and determine the ages and slip of each earthquake event with the errors. In general, the errors for the numbers, ages, and slips of past earthquake events are ±1-2, no more than ±0.5-1.0ka, and ±0.25m, respectively.

Key words: normal faulting, bedrock fault scarp, cosmogenic nuclides, long-period earthquake record, fault slip rates

摘要： 正断层基岩断面可以连续地记录大地震位移累积的过程以及地震事件序列。该地震序列发生的时间与地震位移导致的基岩断面的暴露时间呈正相关。因此，通过测定断面基岩表层的宇宙成因核素浓度，可获得过去大地震发生的次数、时间和滑动量。目前研究工作中广泛选用灰岩的宇宙成因核素³⁶Cl测年方法，最佳采样基岩断面的高度距地表15～20m。选择采样剖面时，需确保基岩表面记录的是构造事件而并非地貌剥蚀过程，断面平直新鲜。基本采样策略是，在崩积楔之上和之下平行断层滑动方向连续采集一系列基岩样品，单件样品的长×宽×厚约15cm×10cm×3cm。基于样品的宇宙成因核素³⁶Cl和稀土元素-钇（REE-Y）浓度随着样品距地表高度的变化情况，识别出强震次数及其发生时间范围，然后运用数值模型模拟³⁶Cl剖面来获得最可能的强震事件发生时间和滑动量以及相应的误差。

关键词: 正断层, 基岩断面, 宇宙成因核素, 长周期地震记录, 断层滑动速率