Thirty-year time slices of a regional climate model with 0.5° resolution for two ICCP climate change scenarios (A2, B2) were analyzed to produce stochastic time series of annual precipitation. We used these time series to drive the spatially explicit landscape model WADISCAPE with 1.5 km × 1.5 km resolution. WADISCAPE integrates results of a shrub model and an annuals model with finer spatial and temporal resolution. The model generates projections of annual biomass production that can be used for grazing. These projections are also used to estimate changes of the carrying capacity of small livestock. The decrease of herbaceous biomass (annual plants and shrub leaves) and carrying capacity with mean annual precipitation (MAP) was strongest in the semi-arid climate, which therefore is most susceptible to a decrease of MAP. We applied the sigmoid relationship between biomass and MAP to precipitation maps for different categories of landscape topography in a GIS to produce maps of carrying-capacity and change of food resources in the Jordan River catchment.