A physics-based approach for simulating future extreme design storms to assess flood risks

Design storms are often used to assess flood risks in urban and rural catchments. These synthetic storms are not replicas of real extreme rainfall events but rather simplified simulations of them. Using rainfall intensity-duration-frequency curves, these storms are parameterized to follow extreme rainfall conditions. To construct design storms for the future, these curves must first be recalculated to reflect future climate conditions. We propose a framework for adjusting short-duration intensity-duration curves and storm designs to future climate conditions that only requires projected temperature changes during rainy days. To do this, we first utilize the TENAX (TEmperature-dependent Non-Asymptotic statistical model for eXtreme return levels) model, a novel physically-based statistical model that can estimate future rainfall short-duration return levels. It is then possible to simulate future rainfall intensities (i.e., a design storm) using the duration-intensity curves for the future climate. In most cases, the information from climate models at a daily scale can be used to construct design storms at a sub-hourly scale without any downscaling or data bias corrections. We illustrate our approach by re-parameterizing the Chicago Design Storm (CDS) in light of climate change. Using the city of Zurich (Switzerland) as a case study, we demonstrate how we can calculate changes in the intensity-duration curve for durations ranging from 10 minutes to 3 hours by applying the TENAX model to the 100-year return level. We will then show how we can construct a synthetic 100-year return period design storm using CDS based on the present and future climates, as well as produce flood inundation maps to assess the changes in flood risk in the city.