Climate Migration and Sustainable Urban Growth in Lake Victoria

Global South cities are likely to receive large volumes of climate-induced migration, putting additional strain on existing urban areas by pushing up housing prices and overloading infrastructure systems. Nevertheless, spatial data gaps continue to limit planning in these cities. This study develops a geospatial decision-support approach for Kampala (Uganda), Eldoret (Kenya), and Mwanza (Tanzania) in the Lake Victoria Basin. We integrated multi-source open and remotely sensed datasets, including building footprints, population density, residential built-up areas, impervious surface products, transport networks, and terrain-hydrological variables, aggregated into 1-km hexagonal units. Urban development is defined as a binary outcome based on impervious-surface change, to create a site-suitability model that captures the status quo of urban expansion over the past 25 years. Logistic regression is used to model the probability of development. Results show strong predictive performance (AUC: 0.911-0.990) and are used to map development potential for scenario analysis. Across cities, development is more likely in areas with greater developable land and lower density, while avoiding flood-prone zones, indicating outward expansion rather than infill. Stakeholder engagement through participatory GIS, in general, substantiates these findings while also raising context-specific considerations, such as socio-demographic dynamics, land-use planning, and institutional constraints.