Simulating flood impacts in Kampala, Uganda: when do land patterns matter

Abstract

The use of urban growth modeling has been found useful in searching for strategies to improve flood risk management in Kampala, Uganda. However, to generalize the approach in the context of Sub-Saharan Africa, a deeper understanding of the modeling’s implications is required. Building on a simple cellular automata model calibrated for Upper Lubigi catchment, Kampala, a set of scenarios is defined to explore their impacts on flood outcomes. These scenarios include variations on the spatial determinants of urban growth as well as on the level of built up land demand. Land demand is found to affect total hydrological measures but not the flooding patterns. Spatial patterns can introduce important differences in exposure of built up land cover to flood, especially when development is assumed to avoid recurrently flooded locations. Finally, it is not sufficient to consider only built up land cover; variations on the proportions of vegetation and bare soil are found to have large impacts on flood hazard and exposure.