The incident solar radiation (Rs) is an important variable in agricultural studies, particularly for the estimation of evapotranspiration and yield models. However, its measurement is not commonly performed in conventional meteorological stations. The aim of this study was to evaluate nine empirical models to estimate Rs from the temperature for the Metropolitan, Vale do Rio Doce and Zona da Mata areas in Minas Gerais State, Brazil. The models used were Hargreaves, Annandale, Chen, Bristow & Campbell, Donatelli & Campbell and Hunt. Data used were obtained by Rs daily automatic weather stations installed in these regions and belonging to Instituto Nacional de Meteorologia (INMET). For all models local calibration coefficients were derived. The performance of each method was evaluated using the following statistical indicators: coefficient of determination (R²), root mean square error (RMSE), mean bias error (MBE) and test-t. The little difference between the models evaluated suggests that any of these models may be used. However, given the simplicity, performance and significance, the model of Hargreaves, calibrated and with two coefficients, is the most suitable for estimating incident solar radiation.
calibration; empirical equations; Hargreaves; Bristow-Campbell
