CHARACTERISTIC EARTHQUAKE MAGNITUDE: MATHEMATICAL VERSUS EMPIRICAL MODELS
In the literature concerning the characteristic hypothesis, one basic question is widely discussed: is it possible to justify (by statistical tests) favouring the characteristic magnitude model for the interpretation of available catalogues? No generally accepted answer has been given now a days. In a previous paper (Grandori et al.,2008) we analyzed a different question, perhaps more useful from the engineering point of view: is it possible to judge (on the basis of statistical tests) which one of two competing magnitude models is more reliable (all other things being equal) for the evaluation of a specific hazard quantity at a given site?In that paper we described a method which can give an answer to this question, and we studied the controversysurrounding the comparison between characteristic-type magnitude models and the classic doubly truncatedexponential model. We found that in many cases a characteristic magnitude model is more reliable than theexponential model.In the present paper we recall the main features of the method and we apply it to the comparison between amathematical model FM and an empirical (non parametric) distribution F*. The aim is to find an empirical F*which is more reliable than FM, thanks to the substantial reduction of possible errors due to the use of a wrongmodel FM.We do not give a general method for the construction of such F*, nor we maintain that it exists in all cases. Wesimply show how, in a study case, we found the way to construct a very satisfactory F*.
Magnitude distribution credibility of the model comparison between competingmodels.
G. Grandori E Guagenti L. Petrini
Dept. of Structural Engineering , Politecnico di Milano, Milano, Italy
国际会议
14th World Conference on Earthquake Engineering(第十四届国际地震工程会议)
北京
英文
2008-10-12（万方平台首次上网日期，不代表论文的发表时间）