Description

Abstract: Rapid land cover changes occurring in the Rift Valley of Kenya are altering the hydrologic response of critical watersheds. Four Landsat scenes from the past 18 years were used to develop a land cover classification scheme for the Njoro River watershed. These data were input to the Automated Geospatial Watershed Assessment (AGWA), a geographic information system (GIS) tool. AGWA was used to parameterize and run the Soil and Water Assessment Tool (SWAT), a hydrologic model suitable for assessing land cover change impacts on hydrologic response. The automated parameterization routines in AGWA are designed for US soil and land cover data sets and require inputs for terrain, soil, land cover, climate and rainfall. Climate, soil and terrain data were built for the watershed using historical data and field work, and classified land cover data were created using supervised and unsupervised classification and verified in the field. Techniques and methods were created to transform Kenya data sets into suitable formats for AGWA. Preliminary findings indicate the suitability of this type of analysis for watershed assessment in Kenya; changes in landscape and land use are reflected in significant changes to simulated hydrologic results.