Although satellite-borne sensors are now available to estimate cloud cover and incoming short-wave radiation across the Earth's surface, the study of climatic variation and its impact on terrestrial and marine ecosystems involves historical analyses of data from networks of weather stations that only record extremes in temperatures and precipitation on a daily basis. Similarly, when projections are made with global atmospheric circulation models, the spatial resolution of predicted radiation is too coarse to incorporate the effects of heterogeneous topography. In this paper, we review the development and set forth a set of general equations that allow both diffuse and direct solar radiation to be estimated for each month on the basis of mean daily maximum and minimum temperatures, latitude, elevation, slope, and aspect. Adjustments for differences in slope, aspect, and elevation are made by varying the fraction of diffuse and direct solar beam radiation. To test the equations on various slopes and under different climatic conditions, we drew on highquality radiation data recorded at a number of sites on three continents. On horizontal surfaces the set of equations predicted both direct and diffuse components of solar radiation within 1%-7% of recorded values. On slopes, estimates of monthly mean solar radiation were with 13% of observed values with a mean error of less than 2 MJ m(-2) day(-1) over any given month.