Abstract: Urban stormwater management requires a transparent, updateable method for prioritizing documented water-accumulation hotspots when detailed engineering investigations cannot be performed simultaneously at every location. This study presents an uncertainty-aware municipal screening framework for Riyadh using 1,104 hotspot records and four infrastructure-mapping benchmarks (2017, 2019, 2023, and 2025). Eight expert-informed spatial criteria were applied in GIS multi-criteria maps; six physically oriented criteria were separately used for Monte Carlo ranking and comparison against exploratory surface-flow outputs, excluding land-use importance and main-road proximity because they are contextual priorities rather than direct simulated hydraulic responses. Four hundred weight realizations (±30%) yielded MC_mean, MC_std, P_high and TailMean90, defined as the average of the 40 largest scores. A location-grouped XGBoost surrogate reproduced TailMean90 with out-of-fold R² = 0.9718 (RMSE = 0.0747; MAE = 0.0451). An exploratory, uncalibrated, 2D-only InfoWorks ICM surface-flow exercise used 18 spatially matched points under one six-hour, 100-year storm; no 1D pipe-network routing was simulated, while catchment-derived runoff entered the mesh through 2D point sources. This limited comparison illustrates a possible extension of municipal screening, not independent hydraulic validation. The framework supports updateable prioritization for further engineering investigation, without implying causal effects of network construction or optimal allocation of budgets.