In the transition from fossil fuels to renewable energy, electricity from wind farms and solar farms will play a key role. Even in a country as big as Canada, the question arises: will there be enough land where to place all these new wind farms and solar farms? Exactly how much land is needed to generate a Megawatt of electricity? This is a question about energy density. The land that is needed is not just any land: it needs to be in the right spots where winds are strong, and it needs to be close enough to major power lines to be economical. This means that land use for wind farms competes with land use for other uses. That in turn will make the cost of land grow in importance, especially if wind farms that encroach on urban settlements will be asked to compensate owners for easements. Land use is a non-trival consideration as we are expanding the use of renewable energy amidst our climate emergency.

The National Renewable Energy Laboratory has estimated how much land is need for a modern wind farm in the United States. Their report from August 2009 found that the answer is about 34.5 hectares (ha) per Megawatt (MW) of nameplate capacity, plus minus 22.4. This figure includes land that is impacted directly as well as land that is needed to surround the turbines. The latter is often identified as lease area or project area in environmental permit applications. Given the relative wide standard deviation, it would be safe to work with a number of mean plus one standard deviation to allow for increasing complexity of additional sites. Therefore let us work with a number of 57 ha/MW. Let us also assume that a wind farm has a typical capacity utilization of 32%—another conservative assumption even though advertised utilization rates are typically in the 40-45% range, but do not account for maintenance outages and output curtailments. The purpose of the exercise is to get a realistic sense of the land that would be needed.

The land use of 57 ha/MW can be expressed as an energy density as well. Because the wind does not always blow, the 32% utilization rate for one Megawatt of wind farm capacity translates into 320 kW of average output. As we need roughly 570,000 square meters (57 hectares) per Megaawatt of capacity, the energey density comes to about 0.56 Watts per square meter (W/m²).

The Canadian Wind Energy Association (CanWEA) reported that wind farm capacity had reached 13,413 MW of nameplate capacity by the end of 2019. Statistics Canada reports in CANSIM Table 25-10-0015-01that wind farms produced a total output of 32.8 Terawatthours (TWh) during 2019—or 5.1% of total electricity output. As the year has 8,760 hours, and a typical wind farm has a capacity utilization of 32%, this is equivalent to 11,700 MW of nameplate capacity. To replace the roughly 20% of generation that is still from combustible fuel sources (some 132 TWh), Canada would need four times as many wind farms as we have today—a total of 46,800 MW of nameplate capacity. How much land would be needed for this?

As we had seen above, with 57 ha/MW (equivalent to 0.57 km²/MW), we require some 26,676 km² of land. How can one visualize this land size? Our smallest province in Canada, Prince Edward Island, has a land size of 5,660 km². If we covered every square kilometer of beautiful PEI with wind farms, it would not be enough. We would need five PEIs, or about half of Nova Scotia. This is an enormous amount of land. Finding this amount of land in the most economical locations will be difficult. If land suitable for wind farms is getting scare, c could offshore wind farms come to the rescue?

Some of the largest offshore wind farms in the North Sea have over 600 MW nameplate capacity. To provide 46,800 MW of nameplate capacity, we would need 78 farms of 600 MW capacity—or roughly 7,000 wind turbines of 7MW capacity. A tall order, but not inconceivable. The question is of course: would it make economic sense to do so, or are there cheaper or better alternatives?

Onshore wind farms are meeting increasing opposition from local residents, and available land in good wind locations is getting scarcer. As a result, much of the recent expansion of wind energy in European countries has therefore occurred offshore. Historically, the cost of offshore wind has been higher than for onshore wind, but the latest projects that moved ahead in Europe no longer require subsidies to be profitable. A projection form Bloomberg New Energy Finance put offshore wind at $78/MWh, although this is still significantly more expensive than best-in-class onshore projects.

Even though it appears that Canada has ample land where to place onshore wind turbines, in practice they are numerous constraints. As mentioned above, they need to be in good wind locations (often close to lake shores) and they need to be close to transmission lines. They can't be too close to settlements, as there are increasing demands to keep turbines out of sight and at a safe distance to avoid noise (in particular, infra-sound). Finding enough land for onshore wind farms will be harder than many realize. Similarly, offshore wind farms face geographic limitations mostly because ocean shores are often nowhere near major transmission lines. Building these transmission lines is rather costly. While offshore installations are now commonplace around North Sea countries, there are none at all in Canada. (One project was proposed in British Columbia, the Naikun project, but it competes against cheap hydro and is dormant at this point.)

The renewable energy future in Canada looks rosier than south of the border because Canada relies already on a large share of clean hydro electricity. Converting the remaining 20% of electricity generation from fossil fuels to renewable sources is a much easier task than what is needed in the United States, where about 63% of electricity is generated from fossil fuels. To replace all coal-fired power generation in the United States with wind farms requires a much larger amount of land. In 2019, the United States generated 966 TWh from coal, according to the Energy Information Administration. Using the numbers from above, that translates into 344.6 GW of wind farm nameplate capacity and thus nearly 200,000 km² of land. That is about the size of Nebraska! Given this sort of land requirement, it is hard to imagine that wind farms will solve the clean power needs of the United States.

Does solar power fare better? A typical solar panel produces about 100–150 Watts per square meter when receiving full sun—a conversion efficiency of around 10–15%. However, power density must also take into account extra space needed to separate panels if they need to be tilted on a flat rooftop, for example. Another way to capture energy density is to measure the annual output in kWh per square meter of land. A study by Denholm and Margolis find values of 35-146 kwh/m²a, equivalent to 4-17 W/m², with an average for the Unite States of about 9 W/m². It is immediately apparent that the power density of solar energy is larger than wind by an order of magnitude. Denholm and Margolis continue to calculate the per capita solar footprint in square meters per person based on the assumption that electricity needs in each state are met by solar power alone. Unsurprisingly, southern U.S. states get more sunshine and thus people there need less land to power their grids. Roughly, the numbers range from about 100 square meters per person in California to 300 square meters in Wyoming. According to the authors, it would require about 0.6% of the total area of the United States to provide sufficient solar electricity, which is less than 2% of the land dedicated to cropland and grazing. As the efficiency of solar panels improves gradually, less land will be needed. A solar future is not utopia—it is a question of economics.

Even though the land use picture for solar is much better than for wind, neither land use is trivial. Keep in mind that natural gas and nuclear power plants have much higher energy densities (including upstream mining operations). For example, a 1,000 MW nuclear power station (excluding upstream mining) requires just 340 hectares of land—a power density of roughly 300–1,000 W/m². These power sources have a gigantic advantage in terms of their land footprint. The challenge will therefore be to make land use from renewable energy more compatible with current land use: not exclusive use, but co-use. Rooftop solar installations are the obvious example as they don't require exclusive land use, but sit on top of residential or industrial land use. (For example, a recent study estimated the rooftop potential in Lethbridge, Alberta.) Dedicating more surface to such co-use is also ideal in terms of connections to the grid, as small on-site storage can be combined with short distances to the grid, which also benefits from the two-way use of the transmisison lines. The electric grid has to become "smarter" to manage these varying loads and intermittent supplies. The electric grid was built for reliability, and maintaining the integrity of the grid will require investments into infrastructure that go well beyond wind farms and solar panels. The challenges that lie ahead are outlined in a 2011 report that explores many of the issues. Importantly, changes to the way the power grid is operated are needed to direct investments into where it improves efficiency and keeps costs low. For example, local marginal prices need to reflect local abundance and scarcity of electricity.

Bhadla Solar Park in India is currently the world's largest solar power plant, with 2,245 Megawatt of peak capacity. It covers 4,500 hectares of land. Assuming average output of 12% of peak capacity over the course of the year, this comes to a power density of roughly 6 W/m² (equivalent to 52 kWh/m²a, or about 16 ha/MW land use).

Lastly, an important caveat needs to be mentioned. Power generation from intermittent sources does not displace power from dispatchable sources one for one. The problem is the lack of electricity storage over longer durations, as electricity demand fluctuates over the course of the day but also seasonally. Land use calculations for intermittent sources ultimately must also take into account land needed for electricity storage, which is not yet widely available.