Author(s): Jiandong Liu; Ngoc Son Nguyen; Srivatsan V. Raghavan; Minh Tue Vu; Shie-Yui Liong
Linked Author(s):
Keywords: Dynamic downscaling; Stochastic downscaling; Intensity-duration-frequency curves; Rainfall extremes; Urban flooding
Abstract: More intense storm occurrences have been observed in the recent decades and the trend is likely to continue with the changing climate. The changes in rainfall extremes could significantly impact storm drainage systems of urban cities in general, coastal urban cities in particular. To assess this impact on an urban/city scale, a high spatial and temporal resolution of rainfall is required. Although General Circulation Models (GCMs) have been often used to assess climate impacts, the coarse spatial and temporal resolutions (100-200 km spatial and 6-24 hour temporal resolutions) of these GCMs are not suitable for city scales. This paper applies a novel stochastic downscaling method that was originally developed for rainfall nowcasting using radar data. The data from GCMs was first dynamically downscaled to a regional scale using Regional Climate Model (RCM). The rainfall at high spatial and temporal resolutions (1 km and 5 minutes respectively) was further downscaled through a stochastic downscaling of this dynamically downscaled climate. The approach is demonstrated on an island state Singapore. The Intensity-Duration-Frequency (IDF) curves were then constructed from these high-resolution rainfall data. The findings are of immense interest to policymakers to assess the capacity of the existing urban storms drainage systems in light of the anticipated much more intense rainfall extreme due to changing climate. Adaptation and mitigation measures can then be more effectively and efficiently deliberated.
Year: 2018