Updating the intensity-duration-frequency curves in major Canadian cities under changing climate using CMIP5 and CMIP6 model projections

Abstract

Characteristics of extreme rainfall events are often presented using the Intensity-Duration-Frequency (IDF) curves. As Canada gets warmer, the credibility of using historical IDF curves for flood risk analysis and design of various urban infrastructure becomes questionable. This study aims at updating and intercomparing the IDF curves in 30 Canadian cities using the simulations of Climate Model Intercomparison Project Phase 6 (CMIP6) under Shared Socioeconomic Pathways 2–4.5 and 5–8.5 and their predecessor CMIP5 under Representative Concentration Pathways 4.5 and 8.5. Quantile-Quantile Downscaling is utilized to estimate extreme rainfall values at fine spatiotemporal scales. Overall, more frequent storms with an average increase of around 30%-40% in their intensity are projected in most cities based on both CMIP5 and CMIP6 models during 2071–2100. For instance, the fast-growing cities of Vancouver, Toronto, and Montreal could expect around 4–11 times more frequent 5-min storms with a 100-year return period during 2071–2100. Moreover, although IDF curves based on CMIP5 and CMIP6 individual models diverge, the expected IDF curves based on their multi-model ensemble mean are somewhat similar if models with a high climate sensitivity based on CMIP6 are excluded from the analysis. This study underlines the need to update IDF curves in Canadian cities.