Estimativas da radiação de onda longa incidente na Região de Pampa Gaúcho

Abstract

The incident long wave radiation flux (L ↓) is a parameter that directly influences the radiation balance on the surface. Measures of this variable are scarce, where the use of classical equations may be an alternative to estimate it. In this way, this work aims to evaluate the performance of classic equations present in the literature and propose a new equation to describe L ↓ about the study areas located in Pampa do Sul do Brasil. First, twelve equations of L ↓ were evaluated for the experimental site of Santa Maria - RS, where we found errors above 20 W m-2. In the attempt to reduce the errors were associated to the models of L ↓ cloud coverage parametrizations. Only under cloudy conditions (NC) the errors were lower than the threshold. Thus, a local calibration of the coefficients of the equations of L ↓, as well as the cloud cover parameters, was necessary. However, the results found after these calibrations did not present values very different from those previously found. A new equation is proposed which uses relative humidity, air temperature and water vapor pressure without the need for cloud coverage parametrizations. Data from the year 2014 of the experimental site of Santa Maria - RS were used to calibrate the equation and data from the year 2015 to validate it. Data from the Pedras Altas-RS site were used to test the new L ↓. The equation proposed in this study (EAI) presented statistical values of RMSE ranging from 12 to 19 W m -2 and R2 between 0.65 and 0.84. We have tested models already described in the literature of L ↓ for regions of Pampa and the EAI equation presented better results. The proposed equation was tested to analyze the radiation balance of the SIB2 model. The results show that the use of the new equation considerably improves the values of the radiation balance of SIB2. Therefore, we recommend the use of the proposed equation for the calculation of L ↓, when presented vegetation cover and meteorological conditions similar to those of this study.