Abstract
Conventional rainwater harvesting (RWH) performance depends on rainfall amount, seasonality, dry-period structure and temporal sequencing, yet estimates based on individual rainfall realisation may not represent the range of responses under climate variability. This study quantified rainfall realisation uncertainty and assessed the consistency of conventional domestic RWH operational regimes across Great Britain and Ireland. Hourly rainfall from twelve UKCP18 ensemble members was used to drive continuous RWH simulations over 99 hydrological years at 14,184 valid 5 km grid cells using a continuous yield-after-spillage (YAS) model and a fixed household benchmark. Performance was evaluated using water-supply efficiency (EWS), retention efficiency (ER) and four tank-state diagnostics describing near-empty, near-full, overflow and mean storage conditions. The driest and wettest ensemble members differed by 18.4% in spatial mean annual rainfall; corresponding Ews increased by 13.5%, whereas near-full and overflow fractions increased by 68.5% and 69.7%, respectively. Under the fixed benchmark, 39.3% of cells were inflow-limited, 42.0% storage-limited and 18.7% transitional. Mean and median member agreement were 0.897 and 1.000, respectively, indicating that the operational-regime classification was generally consistent across rainfall realisations. Threshold sensitivity altered regime proportions but retained median agreement of 1.000 across all tests. Storage sensitivity showed diminishing returns, with 75.5% of cells reaching the selected marginal-gain threshold within the 1.0–1.5 m3 range under ΔEWS < 0.020. The results show that rainfall-realisation uncertainty can substantially affect RWH performance magnitude while having a smaller influence on the diagnosed operating constraint. The framework provides a basis for climate-responsive RWH screening under multiple rainfall realisations.