The purpose of this research is to further develop statistical models of local temperature changes, using Landsat-5 satellite remote-sensing data. The temperature at any location and for any land use is modeled as a function of the pattern of land uses around this location. These models are estimated with data pertaining to the Columbus, Ohio, metropolitan area (CMA). Their applicability to land-use planning and regulation is illustrated by simulating hypothetical land-use changes in part of the CMA, and computing the resulting temperature effects. The results clearly demonstrate that it is possible to reduce temperatures in residential and urban areas through a judicious siting of green areas.