Desempenho da radiação solar global diária e mensal do ERA5-LAND e terraclimate no nordeste brasileiro tendo por referência o GRID BR-DWGD.

Abstract

Studies related to Global Solar Radiation (GSR) are highly relevant to environmental research, and it is of great interest to evaluate how solar radiation influences climate-related processes. However, the acquisition of reliable data is hindered by the limited number of meteorological stations in the country, as well as the frequent occurrence of gaps in measurements. This study aims to: (a) evaluate the performance of daily global solar radiation (DGSR) and monthly global solar radiation (MGSR) estimates available from the ERA5-Land and TerraClimate databases for Northeastern Brazil by comparing them with the BR-DWGD reference dataset; (b) identify the presence of trends in the three datasets; and (c) verify whether changes occurred in the GSR climatological normals during the analyzed period. All datasets were extracted and processed using the Google Earth Engine platform for the period from 1961 to 2020. GSR data from Automatic Weather Stations (AWS) of the Brazilian National Institute of Meteorology (IN-MET) were analyzed for the period from 2007 to 2011 and compared with BR-DWGD data in order to assess the accuracy of the latter dataset. The performance of ERA5-Land and Terra-Climate data was evaluated using statistical indicators: Mean Absolute Error (MAE), Mean Relative Error (MRE), Root Mean Square Error (RMSE), Pearson’s Correlation Coefficient, and Systematic Error (Bias). The presence of trends in the time series of the three datasets was assessed using the non-parametric Mann-Kendall test, and changes in climatological normals were evaluated using the statistical test for differences in means for large samples. The results revealed greater consistency between the TerraClimate and BR-DWGD datasets, as evidenced by the MAE, MRE, RMSE, and Bias error metrics, as well as by the statistical tests for diffe-rences in means, demonstrating the reliability of the results. In the municipality of Cabaceiras, Paraíba State, for example, MAE values of 1.16 MJ/m², MRE of 5.61%, and RMSE of 1.81 MJ/m² were observed, indicating good agreement between the analyzed datasets. The results also indicated good consistency between INMET and BR-DWGD data, confirmed by the sta-tistical error metrics, which validates the representativeness of the observational data used as reference. Climatic analyses indicated an increasing trend in monthly global solar radiation in TerraClimate estimates for most of the analyzed locations, whereas ERA5-Land showed this behavior in a more limited number of areas throughout the evaluated periods.