Abstract Medium-term wind energy production forecasting, spanning time scales from days to several months, is essential for optimizing wind farm integration into electricity generation systems. This optimization leverages renewable energy resources advantages, including a reduced environmental footprint and greater energy independence, while tackling energy price volatility and reducing exposure to potential supply crises. Accurate forecasts are fundamental for modern power systems, enabling effective power generation coordination, maintenance planning, and resource management. As systems increasingly incorporate higher shares of renewable energy sources, challenges such as reduced system flexibility, constrained storage capacity, and escalating operational complexity become more pronounced. This growing complexity amplifies the need for reliable medium-term forecasts, which enable informed planning and decision-making to manage variability and uncertainty and support more flexible and resilient system operation. Wind energy production is inherently variable and highly dependent on local weather conditions, underscoring the necessity for advanced forecasting methods. This study presents a novel methodology that integrates seasonal climate forecast ensembles with wind turbine characteristics and historical wind speed data to generate probabilistic energy forecasts, thereby quantifying and constraining uncertainty. To assess its real-world performance, the methodology was applied to a substation in South Euboea, Greece, which aggregates production from four wind parks. Over a 42-month testing period with seven distinct seasonal forecast datasets, the approach demonstrated a mean absolute percentage error under 17% and produced actionable operational signals, such as a guaranteed minimum production level. These findings underscore the value of seasonal forecasts for developing effective medium-term prediction tools to aid operational decisions.
Tzanes et al. (Thu,) studied this question.