Monitoring access to piped water and sanitation infrastructure in Africa at disaggregated scales using satellite imagery and self-supervised learning

arXiv:2411.19093v4 Announce Type: replace Abstract: Access to drinking water and sanitation services is essential for health and well-being, yet large global disparities persist. Sustainable Development Goal (SDG) 6 sets targets for universal access to these services, but progress toward these targets is hindered by existing monitoring systems that rely heavily on costly, infrequent, spatially uneven household surveys and censuses subject to substantial reporting delays. To address this gap, this study develops a scalable remote-sensing framework for estimating access to piped water and sewage systems at approximately 2.56 km spatial resolution. The framework combines Sentinel-2 imagery, Afrobarometer survey responses, 30 m population data, and Vision Transformer representations learned with DINO self-supervised learning. The best-performing model achieves held-out AUROC values of 91.54\% for piped water and 93.24\% for sewage system access across African survey locations. Applied to gridded inference across 50 African countries, the resulting population-weighted estimates closely track WHO/UNICEF JMP statistics for piped water access ($R^2=0.92$) and show meaningful agreement for sewage-related sanitation access ($R^2=0.72$). In countries without Afrobarometer survey coverage, the model achieves population-weighted MAEs of 9.5\% for piped water and 10.7\% for sewage system, with estimates falling within 15\% of JMP values for 121.4 million and 159.7 million people, respectively. A Nigeria application across 767 Local Government Areas (LGAs) shows how our framework's fine-scale predictions reveal subnational spatial inequality relevant to environmental justice.