Analysis of urban design storm based on stochastic storm transposition

Urban flooding is becoming more severe due to increases in extreme rainfall amounts and impervious land cover. Design storms are critical for developing infrastructure and land use solutions to these flood challenges. In this study, we examine urban design storms using Stochastic Storm Transposition (SST). SST is a technique that uses probabilistic resampling from a "storm catalog" of observed rainfall events selected from a pre-defined regional "transposition domain". Resampled storms are then spatially transposed to estimate distributions of extreme rainfall intensity, duration, and frequency (IDF). This study presents an SST-based rainfall analysis for Shanghai, China. Results show that there is spatial heterogeneity in rainfall across study's transposition domain which must be accounted for during the transposition step of SST. SST can reproduce reasonable IDF estimates and provide design storms with realistic spatiotemporal rainfall structures. Large variations in spatio-temporal structure of these SST-based design storms under different return periods call into question the conventional hydrologic engineering assumption of design storm rainfall that is spatially-uniform and temporally idealized.