Uncertainty Analysis of Future Design Floods for the Yongdang Reservoir Watershed using Bootstrap Technique

Abstract

To estimate design floods for hydraulic structures, statistical methods has been used in the analysis of rainfall data. However, due to the lack of rainfall data in some regions, it is difficult to apply the statistical methods for estimation of design rainfall. In addition, increased uncertainty of design rainfall arising from the limited rainfall data can become an important factor for determining the design floods. The main objective of this study was to assess the uncertainty of the future design floods under RCP (representative concentration pathways) scenarios using a bootstrap technique. The technique was used in this study to quantify the uncertainty in the estimation of the future design floods. The Yongdang watershed in South Korea, 2,873 ha in size, was selected as the study area. The study results showed that the standard errors of the basin of Yongdang reservoir were calculated as 2.0~6.9 % of probable rainfall. The standard errors of RCP4.5 scenario were higher than the standard errors of RCP8.5 scenario. As the results of estimation of design flood, the ranges of peak flows considered uncertainty were 2.3~7.1 %, and were different each duration and scenario. This study might be expected to be used as one of guidelines to consider when designing hydraulic structures.

Uncertainty Quantification of Future Design Rainfall Depths in Korea

“…Yoon and Cho [53] performed uncertainty analyzes of rainfall extremes through multiple model ensembles of 9 GCMs. Lee et al [54] estimated the design rainfall depth using a traditional bootstrap technique and evaluated the uncertainty of future design flood. In their study, the range of estimated peak discharges was used as an evaluation criterion for uncertainty.…”