Desempenho de modelos de estimativa da radiação solar e evapotranspiração de referência

Abstract

Solar radiation (Rs) is the most important source of energy in nature and directly influences the evapotranspiration process. There are many ways of estimating solar radiation and evapotranspiration, however, they require a range of data, often unavailable. Given that, in the absence of observed Rs data, it is possible to estimate them using mathematical models. Overall, the models vary in degree of complexity as well as in the adjustment coefficients, and they must be calibrated to the place of interest to obtain the best performance. In this sense, the main purpose of this study is to evaluate the performance of models for the estimation of global solar radiation and reference evapotranspiration on a daily scale, based on limited meteorological data, for the region of Santa Maria - RS. The data were taken from the automatic meteorological station (EMA) of the Instituto Nacional de Meteorologia (INMET), located in the city of Santa Maria - RS. Ten models were evaluated in the estimations of global solar radiation, where five models with calibrated coefficients were used and adjusted through the software TableCurve 2D and, five models had the coefficients determined by the authors. Then, the reference evapotranspiration was estimated using the Penman Monteith method, adopting the solar radiation data measured at the station, which were considered as standard and compared with the solar radiation results estimated by the ten models. The models proposed by Ball (BA), Chen (CH) and Meza and Varas (MV) presented a performance index of 0.80, indicating "optimal" performance for the models with calibration of their coefficients. In this way, they are recommended when there is no availability of measured data at the study site. For the non-calibration models, except the Donatelli and Campbell (DC) model, in which obtained a "good" performance, the other models had an average index of 0.77, indicating a "very good" performance. Considering the estimation of reference evapotranspiration, nine of the ten models presented "excellent" performance. Therefore, it is concluded that there is no need to calibrate the coefficients and all models are indicated for Santa Maria - RS.