Convection-permitting analysis of sub-daily extreme rainfall events in the Awash River Basin

Study region Awash River Basin, Ethiopia. Study focus Extreme rainfall at sub-daily scales is a trigger of flash floods, yet its future behavior remains poorly analyzed in data-scarce regions. Convection-permitting climate models provide an improved representation of short-duration rainfall extremes, but their short simulation length limits the applicability of traditional extreme-value approaches. This study analyzes sub-daily extreme rainfall in the Awash River Basin using high-resolution CP4-Africa simulations representing present-climate (1997–2007) and future-climate conditions (2095–2105). Extreme rainfall events are identified using end-aligned rolling accumulations at durations of 1–24 h within a peaks-over-threshold framework, with temporal declustering applied to isolate independent storm events. Future changes are quantified using empirical quantile-based change factors and bootstrap uncertainty estimation without assuming parametric tail behavior. In addition to basin-mean intensity changes, the analysis examines seasonal modulation, changes in event frequency and intensity, and storm-scale spatial organization using footprint-based diagnostics. New hydrological insights for the region Results indicate consistent intensification of rainfall accumulation within extreme events across durations, accompanied by duration-dependent changes in storm spatial extent. Spatial expansion is more evident for longer accumulation windows, while short-duration events remain more localized. These findings suggest greater flood-generating potential and increased flash-flood risk in the basin. The analysis demonstrates the value of event-based diagnostics for interpreting convection-permitting climate projections and provides new regional insight into how sub-daily rainfall extremes may influence future flood hazard in data-scarce regions.