Wind turbines: a smaller footprint than you might think

 (Image: Pixabay CC0)
(Image: Pixabay CC0)

Wind power is an affordable and renewable energy source.

Yet decision-makers are reluctant to invest in this sector because they generally believe that wind farms require more land than fossil fuel power plants. A McGill University study assessing the extent of land occupied by nearly 320 wind farms in the USA - the largest study of its kind - paints a very different picture.

Do thermal power plants really take up less space?

Recently published in the journal Environmental Science and Technology , the study shows that when calculating the surface area occupied by a wind farm, it is generally assumed that the entire area is devoted to wind energy exploitation. However, the electricity-generating infrastructure (wind turbines, roads, etc.) usually occupies only 5% of the total area of the wind farm - the rest of the land is often used for other purposes, such as agriculture.

The study also shows that if they are sited in areas where there are already roads and other types of infrastructure, such as on farmland, wind turbines can be around seven times more efficient, in terms of energy produced per square metre of area directly affected, than if they are sited on virgin territory.

ÖLand use is one of the main reasons cited against wind power development," explains Sarah Jordaan , Associate Professor in the Department of Civil Engineering at McGill University and lead author of the study. "Yet, after calculating the surface area occupied by nearly 16,000 wind turbines in the western U.S., we found that gas-fired power plants take up no less space, if we take into account all the infrastructure associated with the natural gas supply chain, such as wells, pipelines and roads."

A promising new assessment method

Until now, it has been difficult to accurately assess the land use associated with wind energy development in the USA. Indeed, past studies only took into account infrastructure and land use on a relatively small scale, making it difficult to extrapolate their results. Other studies were based on estimates made for the wind farm as a whole, rather than for the land directly affected by the infrastructure.

Using geographic information system (GIS) data and machine learning models developed from nearly 2,000 images of wind farms located in the U.S. portion of the Western Interconnection (interconnected grids that supply electricity to 14 U.S. states, as well as parts of Canada and Mexico), the research team trained a deep learning model to obtain a portrait of land use by wind farms.Western Interconnection (interconnected grids that supply electricity to 14 U.S. states as well as parts of Canada and Mexico), the research team was able to train a deep learning model to obtain a portrait of land use by wind farms. It was thus able to assess several factors (pre-existing roads, location and age of wind turbines, etc.) that directly influence land use.

Sarah Jordaan adds: "The method we have developed could be used to carry out analyses for other energy technologies, whether for measuring an environmental impact or for setting up a carbon-neutral system. In fact, it could enable a consistent comparison of the environmental sustainability of various energy technologies, which was previously impossible."

The articleÖLandResources for Wind Energy Development Requires Regionalized Characterizations," by Tao Dai et al, was published in the journal Environmental Science and Technology.
DOI: 10.1021/acs.est.3c07908

Another article by the same research team:
"The life cycle land use of natural gas-fired electricity in the US Western interconnection", by Tao Dai et al, published in Environmental Science: Advances.
DOI: 10.1039/D3VA00038A

Backers