H21I-1510
Disaggregation of Daily Rainfall using Modified K-Nearest Neighbors Resampling Method
Tuesday, 15 December 2015
Poster Hall (Moscone South)
Gunhui Chung, Hoseo University, Department of Civil Engineering, Cheongcheongnam-do, South Korea and Heeseong Park, Korea Institute of Civil engineering and building Technology, Hydro Science and Engineering Research Institute, Gyeonggi-do, South Korea
Abstract:
As the city is developed, the protection of the city from the severe flood becomes important due to the possible massive economic damage and fatality. In a mega city, such as Seoul, the travel time in an urban area is less than 2 hours, which causes the difficulty in the estimation of runoff and requires a set of short time step rainfall data. Therefore, in this study, the daily rainfall is stochastically disaggregated into hourly data using Modified K-Nearest Neighbor (MKNN) resampling method. When MKNN resampling method is applied, 7 rainfall patterns are developed using historical hourly data and utilized these patterns to reproduce the similar hourly patterns of the disaggregated data. The proposed method is applied in the 54 years of Seoul precipitation data.
