Interaction of climate and wind power
Dr. Robert Vautard is senior scientist at the laboratory of climate and environmental science in France. He has extensive experience in the field of climate modelling and extreme events. Also he contributed to the 5th IPCC report. In this symposium he will place renewable energy in the context of climate change and long-term climate variability. He will mainly focus on wind energy aspects, and address three important questions. First of all, how is does climate change impact wind power resources? Secondly, does wind power development have in impact on regional climate. And finally, combining the two: does wind power development have an impact on wind resource?
To answer the first question, Vautard starts with a brief introduction to regional climate modelling. Because long-term climate models run at a relatively course resolution, regional climate models are used to refine these simulations onto a smaller domain. This is important to assess the influence on climate change on regional circulation, especially for accurate projection of extreme events. It can also be used to assess the impact of regional climate policy. As such, it is suitable to investigate the effects of large wind farms. The EURO-CORDEX regional climate model is used in this study. To assess uncertainties, it is run in an ensemble configuration.
Overall changes in wind power potential are within 5%, but with clear spatial differences.
Vautard describes an experiment in which the power yield from wind turbines was computed, based on the EURO-CORDEX wind projections. A simple extrapolation from the 10m wind to hub height is used in conjunction with a reference power curve. This method is quite simple, but in the context of this study it is enough to give reasonable estimates of the changes in wind power. A complete dataset with turbine locations and characteristics is available here (not free). To account for future expansion of wind energy, several scenarios have been defined for the future wind turbine fleet. To construct these scenarios, the “CLIMIX” approach was used. Turbines are added over the model domain, with higher density in areas with better wind resources and taking care to avoid unsuitable land-use categories, until the projected wind energy production has been reached. This gives a seemingly realistic distribution of wind turbine locations over Europe.
The downscaled RCP8.5 climate scenario shows that 10m winds decrease in a very broad “coastal” zone along the Atlantic and the Mediterranean. Further inland and at the Baltic Sea, however, 10m winds increase. As a result, the overall changes in wind power potential are within 5%, but with clear spatial differences. The decrease of wind speed over the Mediterranean in the RCP8.5 scenario has been linked to a poleward shift of the Hadley circulation. This might give rise to an increased frequency of “wind droughts”: extended episodes of low wind speeds.
Results suggest that as more wind turbines are added to a windfarm, electricity generation initially increases, but levels off around 1 W/m2
With respect to the impact of wind farms on regional climate, Vautard metions that several studies have been conducted but no clear results were found so far. One particular study showed an increase of nighttime temperature between 0-1° in areas with wind farms. Subsequently, he shows the results of a number of WRF experiments, in which they used the available Fitch turbine parametrization for wind farms. The model is run on a 50 km resolution, forced by ERA-Interim data and run for 33 years. Again, the CLIMIX approach was used to project the spatial distribution of wind turbines for 2020 and 2050. Surface temperature changes within ±0.5°, but with that same distinct spatial variability. For sea level pressure, however, there is a clear significant increase up to 0.6 hPa centered over Poland, leading to a slight decrease in wind speed and precipitation. From the WRF simulations, it seems that this pressure change is related to the excitation of a Rossby wave: however, global simulations would be necessary to further investigate this feature.
Finally, to investigate the effects of wind farms on their own yield, Vautard describes a modelling study that was performed over Kansas. A very large wind farm was implemented in the WRF model, and several simulations were performed in which the turbine density was gradually increased. The results suggest that as wind turbines are added to a wind farm, the electricity generation initially increases, but levels off around 1 W/m2. This suggests that there is a fundamental limit to the amount of electricity extracted from the atmosphere.
As with the other talks, the presentation is met with many questions and feedback from the audience. One of the visitors notes that the wind resources in the Mediterranean area and Iberian Peninsula are driven by mesoscale circulations, and that the resolution of the regional climate model may still be too coarse to represent this well. This is definitely something to take into account for further studies.