Evaluation of design rainfalls using the before-after control-impact (BACI) analysis

Abstract

This study uses the BACI analysis to estimate the change of design rainfalls in south Korea due to the typhoon Rusa. In 2002, passing through the Korean penensula, the Rusa poured down beyond the historical rainfall. This typhoon renewed the recorded historic maximum rainfall data. Especially, the Gangwon Youngdong region and the coastal area of Kyungnam south eastern have been damaged seriously. In this study we collect observed rainfall data during the whole year at 56 gauge stations in south Korea. By using the BACI analysis we can divide the data period in the two parts: the first 20 years recorded data (before20) and the most recent 20years recorded data (after20). The rainfall by Rusa is included in after20. We set the Ganwondo Wonju, Jeonlado Namwon, and Kyungsangdo Daegu, which are located at inland area, as control sites, and choose the Gangwondo Daeguanryung, Gangloong, and Kyungsangdo Pohang, that is the region near the coast, as impact sites. In this method we can assess the amount of change of the design rainfalls and the amount of change of the parameter that influence to the probability density function between control and impact sites. To compute design rainfalls, this study choose the generalized extreme value distribution (GEV) so as to deal the maximum rainfall efficiently. Finally we can visualize the effect between before and after according to the generated typhoon, we can assess the autocorrelation, implement of the analysis of variance (ANOVA) between parameters that influence the probability model and design rainfalls. Through the BACI analysis we can estimate the statistical change in design rainfalls by typhoon Rusa.