RESUMEN

Water scarcity is a global issue resulting from rapid urbanization, increasing population growth, industrial development and expansion of human activities over time and space. Water shortage affects every continent and is listed as one of the largest global risks hence the need for proper management of water resources. Municipalities and cities worldwide are struggling to maintain a steady supply of water to meet the increasing water demand. The study used Geographic Information System (GIS) techniques and Multi-Criteria Decision Analysis (MCDA) to develop a decision support model that can be applied to improve the utility water demand management for the Lodwar Municipality in Turkana Kenya. The data comprised remotely sensed data, population density, spatial plans, utility infrastructure maps and metered water connections data. The AHP pairwise comparison matrix was applied to assign weights for the 8 criteria influencing water demand in the area. The population density, proximity to water network facilities and land use criteria were equivalent to 30 %, 25 %, and 23 % respectively whereas 22 % of other criteria were dependent on each other. The analysis of satellite images showed the expansion of built-up areas and emerging human activities in regions towards the South and Western of Lodwar Town. The resulting model outcome identified the potential demand priority sites within the region of which some are underserved. The model efficiency was assessed through the application of statistical indicators as well as graphical and map presentations. Consequently, the addition of more input variables affecting demand is likely to improve the results over changing aspects within the zones. Municipality water utility managers and decision-makers can therefore employ the model information to highlight suitable areas for network expansion as well as infrastructure management planning within the municipality. This method offers an alternative hybrid technique for mapping potential utility water demand in rural municipalities with inadequate data.