日本列島20世紀の降水量時系列の経年的非定常性とその確率降水量の評価値に及ぼす影響

概要

論文の詳細を見る
As a preparatory step of moving partial data (sample size: 31years) frequency analysis, the stationary process in object precipitation time series was assumed, and the best fitting probability distribution of the above data was searched by calculating SLSC. This resulted in the selected Pearson family distribution in many cases. Next, the trends of mean and standard deviation were investigated by using moving partial data. Here, linear regression analysis was used for the distinction of the standard of nonstationarity. As a result, not only the mean, but also the standard deviation was found to have nonstationarity. Therefore, the 10 year (non-exceedance) return period values of monthly, seasonal and annual precipitation, and the 100 year (exceedance) return period values of annual maximum daily precipitation were estimated. The moving partial sample frequency analysis proposed here would be useful for water shortage disaster management and flood disaster management.
日本自然災害学会の論文
2008-02-29