Wind power potential and intermittency issues in the context of climate change

Wind power is developing rapidly because of its potential to provide renewable electricity and the large reduction in installation costs during the past decade. However, the high temporal variability of the wind power source is an obstacle to a high penetration in the electricity mix as it makes difficult to balance electricity supply and demand. There is therefore a need to quantify the variability of wind power and also to analyze how this variability decreases through spatial aggregation. In the context of climate change, it is also necessary to analyze how the wind power potential and its variability may change in the future. One difficulty for such objective is the large biases in the modeled winds, and the difficulty to derive a reliable power curve. In this paper, we propose an Empirical Parametric Power Curve Function (EPPCF) model to calibrate a power curve function for a realistic estimate of wind power from weather and climate model data at the regional or national scale. We use this model to analyze the wind power potential, with France as an example, considering the future wind turbine evolution, both onshore and offshore, with a focus on the production intermittency and the impact of spatial decorrelations. We also analyze the impact of climate change. We show that the biases in the modeled wind vary from region to region, and must be corrected for a valid evaluation of the wind power potential. For onshore wind, we quantify the potential increase of the load factor linked to the wind turbine evolution (from a current 23% to 30% under optimistic hypothesis). For offshore, our estimate of the load factor is smaller for the French coast than is currently observed for installed wind farms that are further north (around 35% versus 39%). However, the estimates vary significantly with the atmospheric model used, with a large spatial gradient with the distance from the coast. The improvement potential appears smaller than over land. The temporal variability of wind power is large, with variations of 100% of the average within 3-10 h at the regional scale and 14 h at the national scale. A better spatial distribution of the wind farms could further reduce the temporal variability by around 20% at the national scale, although it would remain high with respect to that of the demand. The impact of climate change on the wind power resource is insignificant (from +2.7% to -8.4% for national annual mean load factor) and even its direction varies among models.