A B ayesian Spatial Disaggregation Approach for Fine‐Scale Mapping of Water Access Disparities

Access to clean water remains a critical challenge for achieving Sustainable Development Goal 6: Clean Water and Sanitation. While national statistics agencies report water access at coarse municipal scales, this data obscure critical hotspots where vulnerable populations lack essential services. This study presents a Bayesian spatial disaggregation methodology to downscale municipal‐level water access indicators into 1 km 2 raster estimates, enabling precise localization of water‐insecure populations. Applied to 85 of the 87 municipalities in Santander, Colombia, our model combines a latent Gaussian model, whose spatial random field uses the stochastic partial differential equation (SPDE) approach of the integrated nested Laplace approximation (INLA) framework, with spatial covariates—including travel time to cities, nighttime lights, and distance to inland water—to predict fine‐scale population counts lacking drinking water. In municipal leave‐one‐out cross‐validation, predicted and observed municipal totals were strongly correlated ( r = 0.89). The resulting maps enable targeted spatial prioritization for water infrastructure investment and SDG 6 localization. This approach provides a replicable methodology for translating global development targets into actionable local interventions across tropical and developing regions.