Estimates of gridded relative changes in 24-h extreme rainfall intensities based on pooled frequency analysis

Abstract

The potential effect of climate change needs to be considered in urban infrastructure design and risk assessment to improve reliability. The present study proposes a methodology for obtaining grid-scale relative changes for updating 24-h extreme rainfall intensity, through the estimation of rainfall intensity quantiles from baseline and future simulations using a pooled frequency analysis approach. Coherence of relative changes over return periods and time horizons is analysed, and adjustments are proposed to facilitate the application of relative changes in practice. The approach is applied to Canada, using gridded daily precipitation series from model combinations belonging to the North American Coordinated Regional Climate Downscaling Experiment. Multi-model 10th, 50th and 90th percentile relative changes are provided for six return periods, considering two future scenarios (RCP 4.5 and RCP 8.5), and two horizons (2050 and 2080). Overall, estimated relative changes varied smoothly and formed a number of clusters of similar values across the country. Relative changes for RCP 8.5 are recommended for 2050, whereas either those for RCP 4.5 or RCP 8.5 could be used for 2080. As an example, median multi-model 50th percentile relative change over Canada is found to be 14%, 16% and 27% for RCP 4.5 – 2080, RCP 8.5 – 2050, and RCP 8.5 – 2080, respectively.