A spatially distributed thermal model has been developed that simulates thermal processes at the surface of the tundra, within the active layer, and in the underlying permafrost. This model was developed and applied to simulate processes on the Kuparuk River watershed on the North Slope of Alaska. Gridded meteorological data came from seven stations. Meteorologic data to calculate the surface energy balance at each of the nodes were distributed across the watershed using kriging. The kriged air temperature was also adjusted to account for elevation differences using the dry or wet adiabatic lapse rate as appropriate, and incident shortwave radiation was adjusted to consider slope effects. The equations describing the thermal processes of the surface energy balance were solved simultaneously for the surface temperature. This calculated surface temperature was then used in the subsurface finite element formulation to calculate the temperature profile and depth of thaw in the soil. Thermal properties of the soil were estimated spatially on the basis of measurements collected in typical landform vegetation units and then distributed on the basis of a vegetation map. Performance of the model was judged on the basis of comparison to measurements of soil temperatures and thaw depths. The model performs quite well in areas where subsurface thermal properties are well known. The model explains greater than 80% of variance at the surface when comparing predicted subsurface temperatures versus measured soil temperatures, and it increases in performance at greater depths. The model explains 82% of variance when comparing predicted thaw depths versus thaw depths measured over 1 km2 grids. ¿ 1998 American Geophysical Union