The Impact of Wind Speed Trends on Wind Energy Generation

Renewable energy investors, financial planners, and risk analysts require accurate estimates of future wind distributions to assess investment risk, anticipate energy revenue, and ensure grid reliability. However, empirical evidence from historical modeling and simulations of future wind speeds from CMIP6 global climate models suggest that wind speed distributions might not be static–even if the underlying physical drivers are not fully understood.

To investigate the effect of different climate trends on future wind generation, we calculated four possible future wind speed trends at each site in a sample of 100 representative US wind farm locations. One of these trends is the historical ERA5 wind speed trend from 1960-2023–reflecting the hypothesis that past historical trends will continue unchanged into the future. The other three climate-model-derived trends are calculated by first considering an RCP scenario and then averaging 2016-2046 trends from an ensemble of corresponding CMIP6 models. The three chosen pathways are RCP 2.6, RCP 4.5, and RCP 8.5.

At each site, we use the historical trend to detrend historical wind speeds to 2024-equivalents and obtain 64 years of hourly climate-equivalent wind speeds. For each year between 2025-2040 and each of the 4 wind speed trends calculated above, we then retrend wind speeds to obtain year-equivalent wind speeds–and pass the results through a generic wind farm power curve to obtain wind generation capacity factors.

As the relationship between wind speed and energy generation is non-linear, the change in wind energy production resulting from a wind speed trend is not simply proportional to the trend adjustment magnitude–but is dependent on the distribution of hourly wind speeds with respect to a non-linear turbine power curve. We analyze and compare the wind generation on the entire sample of US wind farms in different years and under different trend assumptions.

We take advantage of the preserved intra- and inter-annual wind speed variability of our experimental setup to calculate annual and monthly capacity factors and variances over the next 15 years under various trend assumptions. Finally, we discuss the implications on wind energy risk assessments and power grid resource adequacy.