A study was performed to assess the accuracy of a daily solar radiation model for horizontal or sloping sites in terrain where surrounding ridges and tall trees block both direct beam and diffuse sky short wave radiation. To simulate adequately the major effects of these environments on radiation, the model incorporated (1) standard treatment of solar geometry (after Harris, 1983); (2) separation of direct and diffuse radiation by considering scattering and absorption due to aerosols, water vapor, air molecules and ozone (after Iqbal, 1983); (3) anisotropy of diffuse radiation modeled as a circumsolar component (15–50% of the total clear sky diffuse) plus an isotropic background; and (4) proper calculation of the proportions of sky radiation and ground reflected radiation sensed by both horizontal and inclined sensors. The model was calibrated using data from both horizontal and inclined pyranometers at six sites over four years. Values for the Ångstrom turbidity coefficient and the percentage of circumsolar diffuse radiation were determined using an iterative best fit technique. Mean monthly values of these coefficients and mean monthly ozone and water vapor concentrations from nearby weather service stations could be used for predicting total, cloud free, daily radiation with acceptable accuracy. Required site measurements include slope, aspect, latitude, longitude, date, time, effective horizon for the site, and an assessment of albedo for the surrounding terrain (including vegetation).