Projecting multi-scale sub-daily extreme rainfall distributions from Convection-Permitting Climate simulations: Implications for Intensity–Duration–Frequency curves in the West African Sahel

Abstract. Reliable characterization of extreme rainfall across a continuum of temporal scales is fundamental to hydrological applications, from flood risk assessment to the design of water infrastructure. In a changing climate, however, characterizing this continuum is particularly challenging, as both the distribution of extreme rainfall and its temporal scaling may evolve. Convection-permitting models (CPMs), which explicitly represent convection and provide precipitation at high spatial and temporal resolutions, offer promising opportunities to address this challenge. Yet, their added value for hydrological applications remains poorly explored, particularly regarding model biases and their ability to consistently represent and project extreme rainfall across timescales. We develop a methodological framework to project extreme rainfall distributions from from CPM simulations. The framework combines the identification of potential CPM biases with indicators of simulated changes in extreme rainfall distributions to guide the selection of analytical and empirical approaches for addressing these biases. Extreme rainfall is characterized across durations spanning 2 to 24 hours using synthetic intensity-duration-frequency (IDF) curves based on extreme value and simple scaling theories. The methodology is illustrated over the West African Sahel using CP4-Africa simulations to project locally observed IDF curves under future climate conditions. Onto the Niamey region, results indicate a substantial decrease in the return periods for given extreme rainfall intensities. The simulations are further analyzed to assess the applicability of the framework and guide projections across the Sahelian region. Its generic nature makes the methodology transferable to other regions, given the growing number of climate modeling initiatives based on CPM simulations worldwide.