Rainfall for engineering design is usually estimated from records that are short and sparse, variable in space and time, and increasingly affected by changing hydroclimatic regimes that challenge the stationarity assumptions behind traditional Intensity-Duration-Frequency (IDF) and Depth-Duration-Frequency (DDF) estimation. A structural mismatch compounds this: design rainfall must be resolved for the durations, return periods, and spatial scales applications demand, whereas short records, sparse networks, and ungauged sites rarely support these targets directly. Reliable estimation, therefore, needs models that represent rainfall variability across scales while accounting for the storm types and processes shaping extremes.

This session collects recent advances in modelling rainfall for engineering design: new theory, physically based frameworks, statistical and stochastic models, high-resolution observations (e.g., radar, satellite), physics-based simulations (e.g., convection-permitting models), and hybrid methods. Contributions span scales relevant to hydrological and hydraulic applications, from sub-hourly to multi-day durations and from point to areal estimates. Topics include but are not limited to:

– Methodological advances across spatiotemporal scales, including IDF or DDF derivation and use of physical information and covariates;
– Stochastic models, disaggregation, and continuous simulation for design-relevant estimates across scales;
– Areal estimation, areal reduction factors, and point-to-catchment-scale relationships;
– Multivariate analysis of extremes, including tail dependence, process heterogeneity, and cross-duration dependence, translated into joint design quantities and IDF or DDF surfaces;
– Estimation under changing climatic regimes, including covariate-dependent and trend-informed frequency models, and translation of climate-model outputs into design quantities (e.g., bias correction, change-factor and direct methods, ensembles);
– Process- and regime-conditioned estimation, including atmospheric rivers, monsoonal regimes, seasonality, and storm-type classification;
– Regional frequency analysis, pooling, and estimation at ungauged, data-poor sites;
– Emerging data and methods, including radar, satellite, reanalysis, and convection-permitting products, downscaling, machine learning, and hybrid frameworks;
– Design hyetographs, probable maximum precipitation (PMP), and quantification, propagation, and communication of uncertainty.