Application of GIS-based spatially distributed hydrologic model in integrated watershed management: a case study of Nzoia basin, Kenya

Abstract

Application of GIS-based hydrologic models in management of watersheds in developing countries has lagged behind due to unavailability of adequate spatial datasets. However, use of spatial datasets from different sources, even if the datasets exhibit mismatch in scales may still go a long way in helping understand basin hydrologic processes for the purpose guiding broad policies in watershed management in developing countries. The overall objective of this study was to use Famine Early Warning System Flood Model to study effect of land use / land cover change on River Nzoia flows. The results are considered an important input in policy documents for integrated watershed management for the basin. Land use / land cover grid data were processed for the years 1986 and 2000 in IDRISI Kilimanjaro environment while other grid datasets were processed in case of rainfall from point observations in Arc View. Soil parameters were clipped from USGS continental scale datasets. Famine Early Warning System Flood Model runs as an extension in Arc View environment. The study revealed that the direction of land use / land cover change for the analyzed years was from forest and shrub land to farmland. Model calibration indicated reasonable correlation between observed and simulated flow during low flows and poor correlation for high flows, an anomaly which was attributed to poor rainfall data network, resulting in non-uniform sampling rainfall grid data. Analysis of change of land use / land cover between 1986 and 2000 on river flow revealed that change in land use / land cover resulted in increased hydro graph peaks which increases the chance of flooding.