Comparing and Evaluating Ensemble Generation Techniques from Multi-Model Climate Data for Wind Speed Projection in Rio Grande do Norte for the Present and Future

This study presents a comparative evaluation of ensemble generation techniques for projecting wind speed in the state of Rio Grande do Norte, Brazil, utilizing regional climate models from the CORDEX initiative. Two approaches—Arithmetic Mean (AM) and Convex Combination (CC)—were assessed for the historical period (1994–2023) and for future projections (2031–2060) under the high-emission RCP 8.5 scenario. The findings demonstrate that the AM method consistently outperforms CC, exhibiting higher correlation coefficients and lower root mean square error (RMSE) values across all subregions analyzed. Specifically, the AM ensemble achieved correlation coefficients of 0.88, 0.86, and 0.80 in the northern, central, and eastern regions, respectively, exceeding those of the CC method (0.85, 0.84, and 0.78). Relative to present-day conditions, projected future wind speeds increase by approximately 12.2% in the northern region, 23.5% in the eastern region, and 19.6% in the central region. A notable seasonal shift was also observed, with peak wind speeds occurring later in the year across all areas. These projected increases, when considered in light of the cubic relationship between wind speed and energy production, suggest that wind power potential may rise by over 40% in certain regions. It is also important to acknowledge that such results are subject to uncertainties inherent in climate modeling, including the structural differences among regional and global models and their associated physical parameterizations. Nonetheless, the projected enhancement in wind speed holds significant implications for strategic renewable energy planning in Rio Grande do Norte and reinforces the utility of multi-model ensemble techniques in climate-based energy assessments.