Projected Precipitation Changes over Iran under SSP Scenarios: A Multi-model Ensemble Assessment Based on Robust Performance Evaluation

Abstract Accurate selection of Global Climate Models (GCMs) is essential for reliable climate projections and water-resource planning. This study evaluates the full CMIP6 ensemble (51 models) for monthly precipitation at 31 synoptic stations across Iran (1970–2014) to identify suitable models, then projects precipitation for 2025–2069 under SSP2-4.5 and SSP5-8.5 at five representative stations. Model performance used five statistical indices: R 2 , ANMBD), NRMSE, NNSE, NKGE. Ranking proceeded in two stages: an unweighted Simple Additive Weighting (SAW) baseline, then objective weights from the Entropy method to derive four weighted MCDM methods (WAT, Compromise Programming, Cooperative Game Theory, and TOPSIS). A group decision-making (GDM) approach reconciled method differences to yield a unified ranking. For future projections, a weighted ensemble used the top three GDM-ranked models per station; outputs for 2025–2069 were statistically downscaled via the change factor (CF) method. Evaluation shows NRMSE and R 2 are the most informative objective criteria, while ANMBD contributes least. The final GDM ranking identified IITM-ESM and MIROC-ES2L as the most reliable GCMs for precipitation in the study area. Among single-method alternatives, WAT correlated most strongly with the integrated GDM results, indicating it can serve as an efficient approximation. Future projections display spatially heterogeneous precipitation trends across stations and scenarios, underscoring the local nature of climate impacts. The results imply no single model is optimal across all regions and support combined MCDM approaches and region-specific GCM selection for robust precipitation projections.