APPLICATION OF EXTREME VALUE ANALYSIS MODEL TO ANNUAL MAXIMUM SEISMIC MAGNITUDE DATA USING HISTORICAL INFORMATION AND ITS ACCURACY EVALUATION
-
- Ota Mutsuki
- 新潟県 上越地域振興局地域整備部
-
- IZUMIYA Takashi
- 新潟大学 工学部建設学科
Bibliographic Information
- Other Title
-
- 歴史データを用いた地震津波規模の再現確率値の推定法とその精度に関する研究
- レキシ データ オ モチイタ ジシン ツナミ キボ ノ サイゲン カクリツチ ノ スイテイホウ ト ソノ セイド ニ カンスル ケンキュウ
Search this article
Abstract
We use Monte Carlo simulation to assess an accuracy of quantile and its standard deviation estimated by a maximum likelihood method to annual maximum data with historical information. The historical information is categorized into three classes: “censored” data where the seismic magnitudes are known, “binominal” data, where only threshold exceedance information is available and “range” data, where the seismic magnitudes of historical seismic data have error within predetermined range. The exponential distribution for annual maximum data of seismic magnitude in the eastern margin of the Japan Sea is selected as a candidate distribution. The extreme values and its standard deviation are estimated for the seismic magnitude with historical information. The introduction of historical information in the extreme value analysis can produce quantile estimates with higher accuracy when a proper evaluation of historical information is made.
Journal
-
- Journal of Japan Society of Civil Engineers, Ser. B2 (Coastal Engineering)
-
Journal of Japan Society of Civil Engineers, Ser. B2 (Coastal Engineering) 72 (2), I_295-I_300, 2016
Japan Society of Civil Engineers
