Performance and estimation of solar radiation models in state of Minas Gerais, Brazil

Abstract

Solar radiation is one of the most important meteorological variables, as it is directly related to evaporation. Based on this variable, it is possible to develop models for estimating meteorological elements. However, when there is lack of data, estimates can be made using mathematical models. The objective of this paper was to perform the calibration and statistical performance evaluation of fifteen simplified models of global solar radiation estimation based on air temperature for 51 cities in the state of Minas Gerais, Brazil. The data were provided by the National Meteorological Institute using Automatic Meteorological Stations (EMA’s) located in the cities studied. The performance indexes used were the Coefficient of Determination for Linear Regression (R2), Root-Mean-Square Error and Mean Relative Error, and the Willmott d Index. Through the results obtained, it is possible to observe that the model with the best performance for the state of Minas Gerais was that of Donatelli and Campbell, because, based on the statistical analysis and ordering of the indexes, that is, the use of the position values (Vp) of the indicatives statistics to classify and define the best method for estimating global radiation, this model was the one that obtained the lowest Vp value.