Evaluation and Regional Calibration of Solar Radiation Prediction Models in Southern Spain

Solar radiation ($R_s$) is a very important variable in agricultural meteorology, but it is measured at a very limited number of weather stations. In addition, $R_s$ data are usually low quality because of to sensor failure or lack of calibration. $R_s$ can be predicted from other meteorological measurements (e.g., air temperature), but the accuracy of these methods needs to be tested. In this work, the mechanistic model of Hargreaves-Samani, three modified versions of this model, a modified Bristow-Campbell equation, and the Mahmood-Hubbard model were evaluated at 56 sites in the Andalusian region in southern Spain. All of these methods use the daily temperature range of the air ($\mathrm{\Delta }T=T_{\max }-T_{\min }$) to estimate $R_s$. The estimated values of $R_s$ were compared with measurements of $R_s$ to check their suitability. The accuracy of the methods was primarily affected by the magnitude of $\mathrm{\Delta }T$, with larger $\mathrm{\Delta }T$ resulting in a greater accuracy. Finally, the parameter of the Hargreaves-Samani equation ($K_r$) was adjusted to regional conditions using $\mathrm{\Delta }T$, improving the estimations of incoming solar radiation for a wide $\mathrm{\Delta }T$ range.