Climate-informed spillway design for Horuwila Wewa under future rainfall uncertainty

Abstract

Intensity-Duration-Frequency (IDF) curves for water infrastructure are traditionally developed assuming climate stationarity. However, human-induced climate change has altered extreme rainfall patterns in recent decades, rendering historical curves untenable and raising global flood risk. Horuwila Wewa, an ancient reservoir in Anuradhapura with a catchment area of 26.4 km2 within Sri Lanka's Dry Zone tank cascade system, faces growing challenges managing extreme rainfall under future climate change. This study aims to quantify climate projection uncertainty in future extreme rainfall and assess its implications for Horuwila Wewa's spillway design. Using observed rainfall (1995–2024) and downscaled projections from three Coupled Model Intercomparison Project Phase 6 (CMIP6) general circulation models (GCMs) (CNRMCM6- 1, HadGEM3-GC31-LL, MRI-ESM2-0) under shared socioeconomic pathways (SSPs) SSP2-4.5 and SSP5-8.5, 50-year return-period IDF curves were developed for 2041–2060 and 2081–2100. Daily extremes were estimated using the Gumbel distribution and disaggregated into sub-daily intensities using regional temporal distributions. The resulting IDF curves were incorporated into an uncalibrated HEC-HMS model to simulate reservoir inflows and determine the spillway length required to maintain target peak-pool stage under a 0.8 m bundheight increase. Peak inflow varied from −13.1% to +57.8% relative to the 283.8 m³/s baseline, reflecting substantial inter-model uncertainty. The most severe scenario (MRI-ESM2- 0, SSP5-8.5, 2081–2100) required an 84% increase in spillway length. The results demonstrate that climate projection uncertainty can substantially influence spillway design requirements. Incorporating multiple GCMs and SSPs into climate-informed IDF analysis can provide a more robust basis for adapting reservoir infrastructure to future extreme rainfall.