中国大陆强震的早期余震概率预测效能评估与制约因素

Effectiveness evaluation and constraints of early aftershock probability forecasting for strong earthquakes in continental China

为系统地考察中国大陆地区强震之后的早期余震概率预测效能及制约因素,以期开展适合板内地震活跃地区的“可操作”强余震概率预测,服务震后抗震救灾、应急救援等工作,本文采用充分利用震后早期小震信息的Omi-R-J模型对中国大陆的86个强震序列进行系统性的预测效能评估.通过连续分段滑动、拟合和余震发生率预测,对地震序列的早期模型参数、余震预测限度及其制约因素进行了讨论,并利用N-test检验方法开展了分级分段分档预测结果的效能评估.通过对比1天和30天数据拟合参数的相关性发现,b、k、c具有较强的相关性,差别较小,而p值的相关性较弱,这或许和p值表示余震活动在长期内的衰减特性有关,Omi-R-J模型可在震后早期记录相对不完整的阶段更早的获得较为稳定的序列参数;利用Omi-R-J模型对中国大陆强震序列的预测结果显示,Omi-R-J模型对中国大陆早期的强余震具有较好的预测能力,总的预测有效率为81.16%,平均预测有效率为83.82%,预测过少的比例大于预测过多的比例,早期序列发育较差或监测能力有限的区域,其预测效果受到很大制约;地震序列参数k是影响地震数目的关键因素,而地震序列发育程度可能是影响震后0.05天地震预测效率的重要因素之一.

In order to systematically investigate short-term aftershock forecast efficiency and constraints following strong earthquakes in continental China,so as to develop“operational”aftershock forecasting strategies tailored to intraplate earthquakes in seismically active area,we adopt the Omi-R-J model which can fully utilize the information of early small earthquakes to evaluate the forecasting effectiveness of 86 strong earthquake sequences in continental China.By means of continuous sliding and multi-period fitting,we discussed the early model parameters of earthquake sequences,the limit of aftershock forecasting and its constraints.By comparing the correlation of fitting parameters of 1-day and 30-days data,we found that b,k and c have very strong correlation,whereas the correlation of p value is weak.This may be related to the characteristic that p value represents the attenuation rate of aftershock activity in the long term.Using the Omi-R-J model,we can obtain more stable sequence parameters from relatively incomplete catalog records in the early stage after the mainshock.Early forecast results after mainshock show that the Omi-R-J model has a good forecasting ability for the early strong aftershocks in continental China.The overall forecasting efficiency is 81.16%,the average forecasting efficiency is 83.82%,and the proportion of the forecasting for“too low”is larger than that for“too high”.In some areas with fewer early aftershocks or limited monitoring ability,the forecast efficiency is greatly restricted.Parameter k is the key factor affecting the forecast results of aftershock number,and the record quantity of aftershocks may be one of the most important factors affecting the forecast efficiency during 0~0.05 days after the mainshock.