Estimation of Daily Solar Radiation from Measured Air Temperature Extremes in the Mid-Mediterranean Area

Daily solar radiation $R_s$ at ground level is a necessary input variable required for the evaluation of evapotranspiration and crop growth, development, and yield-simulation models. Nevertheless, it is measured in few weather stations and at many locations it is not observed; also, available $R_s$ temporal series are generally no longer than a few years. A valid surrogate of $R_s$ measurement is the diurnal air-temperature range ($\mathrm{\Delta }T$); indeed, $\mathrm{\Delta }T$ is inversely proportional to cloudiness and therefore could be a good indicator of atmospheric transmittance. As opposed to $R_s$, daily maximum and minimum air temperatures are measured at many locations and their observations in developed countries began in the 19th century. For this reason, several models that permit $R_s$ indirect evaluation from air-temperature data have been suggested in the literature. The most famous models are the simple Hargreaves-Samani (HS) formula, many times recalibrated by the authors, and the Bristow-Campbell model, which has recently been improved. In this paper, the suitability of each proposed model is tested by comparing $R_s$ real data, recorded in 33 Sicilian agrometeorological stations in the period 2003–2008, typically representing the Mid-Mediterranean area, with $R_s$ estimates obtained by the models from $\mathrm{\Delta }T$ data. In addition, a regional relationship is obtained for the scale coefficient K(T) of the Hargreaves-Samani formula: the relationship found improves $R_s$ data prediction with respect to the original HS formulation.