Study On Regional Forecasting Strategies And Forecasting Effectiveness Evaluation

In the analysis of regional earthquake hazard, the evaluation of earthquake prediction efficiency, the construction of reference model and the study of regional earthquake prediction strategy are three important basic problems. While in China's earthquake forecasting research and prediction work of the current system upgrade process, the urgent need to build forecasting model of different time scale and efficiency evaluation of earthquake, to study on regional forecasting strategies and forecasting effectiveness evaluation. Accordingly, the current international is being carried out which based on strict statistical tests,"competition" to establish the forecasting model "Collaboratory for the Study of Earthquake Predictability"(CSEP)planand "Operational Earthquake Forecasting" (OEF) work have important significance for related work in China.In view of the above three problems, this paper choosesthe North South seismic belt (21.0°?43.0°N,97.0°?107.0°E) and "the junction of Shanxi,Hebei and Inner Mongolia"(37.0°?41.0°N,110.0°?116.0°E) as the study area,the North South seismic belt for global "Collaboratory for the Study of Earthquake Predictability"(CSEP) plan Chinese Testing Region, meanwhile, "the junction of Shanxi,Hebei and Inner Mongolia" has been designated as an important earthquake hazard area of China Earthquake Administration for more than 20 times, and it is the most important experimental site for natural earthquake forecasting in North China. The date adopted in this study are based on the earthquake catalogue of "National unified official catalogue" since 1970 were provided by the China Seismological network. In the data processing, the spatiotemporal ETAS model and the stochastic declustering method were used to obtain thebackground sesimicity. To select catalogue of earthquake sequence at the junction of Shanxi,Hebei and Inner Mongolia region,adopt the space-time distribution of natural boundaries methods of "natural boundary", which can distinction between earthquake sequence and the surrounding background sesimicity, by means of latitude-time diagram, longitude-time diagram and epicenter distribution diagram combining approach. In order to ensure the catalogue completeness of earthquake sequences, the "magnitude-number method" is used to estimate the magnitude of completeness for each earthquake sequenceAiming at the forecasting efficiency evaluation and its technical problems,taking into account the current global "Collaboratory for the Study of Earthquake Predictability"(CSEP) planand "Operational Earthquake Forecasting"(OEF) trends,carried out the introduction and application of inspection technology forecasting plan CSEP universal the Molchan diagram, the area skill score, the number test, the student's test and so on. The method will solve the probability forecasting and classification of hazardous area such as test of “black/white forecasting" results,continuous multi period forecasting of the overall evaluation, the incidence of earthquake forecasting test, different forecasting model of the comparative study of the problem. The test method in the optimization and construction plays an important role in the forecasting model, on the other hand, the quantitative evaluation forecasting model and forecasting results, the introduction of domestic relative single method forecasting evaluation results play an important role in the effectiveness of these forecasting scoring methods.Aiming at the reference model needed inthe construction of forecasting modelproblems, in this paper, the global CSEP plan, Chinese Testing Region—the North- South seismic belt area are taken as the research areas, and a model for evaluating the efficiency of seismicity in the area is constructed. According to the principle of simple and clear physical meaning of the construction principle, choose the simple smoothing model (Triple-S model) and relative intensity (RI) model as the reference model, the "reference model" for other earthquake forecasting model to develop the "competitive target",and considering the stochastic declustering methodunder the influence, comparative study with "earthquake accelerating moment release" (AMR) forecasting modelwas carried out. In addition, the spatiotemporal ETAS model is used to obtain the background seismic catalogue after the stochastic declustering. It is used for the forecasting efficiency testof the AMR model. The related results show that the two reference models (Triple-S model and RI model) can be applied to the middle scale forecasting of the North-South seismic belt areas. The actual forecasting effect of the accelerated moment release (AMR) model is lower than that of the reference model, indicating that the current model design of the model can not meet the needs of regional earthquake forecasting.Aiming at the problem ofthe regional earthquake forecasting strategy, to evaluate the forecasting efficiency systematically by continuous sliding forecastusing the epidemic type aftershock sequence (ETAS) model and the ETAS-based thinning algorithmat the junction of Shanxi, Hebei and Inner Mongolia. The results show that the ETAS model and ETAS-based thinning algorithm have a certain ability to forecast the aftershock occurrence rate in this region, with a lower ratio of forecasting fewer occurrence rate and a higher ratio of forecasting more occurrence rate; the forecasting efficiency is good in the early stage after the main-shock, however, it reduces significantly with sequence duration; a modest increase in time window can improve the problem of forecasting fewer occurrence rate, but has little improvement with the problem of forecasting much occurrence rate . In addition, the excitation ability of big aftershocks is weak, which may be one reason for the low efficiency of the ETAS model and ETAS-based thinning algorithm in this region. As possible strategies of forecasting the short-term aftershock occurrence rate, it may be more realistic to use 3-day forecasting time window, to use this algorithm only in the early stage of the sequences,and focus on "no less than" the corresponding forecasting earthquake number during the application of the ETAS model and ETAS-based thinning algorithm in this region.