In this study, a spatially explicit hydro-ecological model (SEHEM) has been developed and validated over a semi-arid grassland sub-watershed in Arizona. The model combines a plant growth sub-model to simulate the seasonal dynamics of root and aboveground biomass, and a hydrological sub-model to simulate soil moisture and temperature dynamics, energy and water budgets for the soil and the vegetation. In addition, the model has been coupled to radiative transfer models (RTMs) in the visible, near infrared and thermal infrared (TIR) bands so that canopy reflectance and directional radiative surface temperature are simulated. Landsat Thematic Mapper (TM) images obtained during a six year period were used to adjust some spatially variable model parameters by minimizing the différence between model simulations and remotely sensed data. Comparisons between observations and model estimates of above ground biomass, net radiation, sensible and latent heat flux, component temperatures are presented.