Click here to the Online Version
A recent article published in Energies covers the groundbreaking wind speed forecasting model, which has been developed to enhance the accuracy of short-term operational predictions in the energy sector, addressing the challenges posed by complex local topography in Costa Rica. The methodology, designed to meet the needs of the Costa Rican Institute of Electricity (ICE), leverages artificial intelligence (AI) and large-scale climate indicators to downscale coarse numerical model outputs into reliable localized forecasts.
This innovative approach, rooted in the World Meteorological Organization’s (WMO) full-value chain framework for weather, water, and climate services, achieved a remarkable reduction of approximately 55% in root mean square error (RMSE) compared to the baseline GFS grid values. By integrating regional climatological knowledge with AI-based downscaling models, the solution demonstrates an effective strategy for generating accurate wind forecasts up to 10 days ahead, directly benefiting wind energy production at local power plants.
As a proof of concept, the study presents a replicable model for WMO members and underscores the potential of AI-driven methods in overcoming the limitations of coarse-resolution numerical models in areas with complex terrain. This advancement marks a significant step toward more reliable and efficient energy sector operations globally.
