Comparing Wind and Nuclear in Terms of Space

May 14, 2011

One of the reasons I support nuclear power is that it seems to require relatively little space to generate a huge amount of power. Some of Canada’s most powerful reactors can produce up to 881MW (electricity), or 7,717,560,000 kw/h annually. That’s enough to power about 643,000 households 24/7 (Average household consumption in Ontario is about 12,000 kw/h annually).

The Darlington nuclear power station – which has 4 such reactors – is about as big as one of the shopping malls in Whitby, ON (including the parking lots). Not bad, considering the plant produces power for up to 2.5 Million households, day in, day out.


Based on that alone, I always thought that nuclear power would be a pretty good option for replacing much of the electricity currently produced globally by burning coal and gas.

But, a recent story published at suggests that, according to a study written by Derek Abbott (a professor at the University of Adelaide), it would be impossible for nuclear energy to supply the entire global demand for energy because all these nuclear plants would take up far too much space.

Abbott addresses other factors,  too, but for the time being, I’ll just focus on the question on size.

I haven’t read the actual study, since it is not yet published (but will be soon in the Proceedings of the IEEE), so I have to go by what the reporter says about Professor Abbott’s findings.

Abbott estimates that

One nuclear reactor plant requires about 20.5 km2 (7.9 mi2) of land to accommodate the nuclear power station itself, its exclusion zone, its enrichment plant, ore processing, and supporting infrastructure.

I’m not entirely sure where he got this number from (I suspect the final article will provide the sources), but it seems he does not allow for multiple reactors on a single site.

The Darlington plant, for example, is a little less than 2 km long (including the parking lot), and roughly 800m or so across. That’s just about 1.6 km2, though my method of measuring that is – admittedly – a little crude. However, there are no less than 4 reactors at that site alone. Even after OPG is done adding another two, the site is not going to get much bigger.


Be that as it may, I will – just for the sake of argument – accept Abbott’s numbers for the time being.

Professor Abbott then calculates how many nuclear reactors it would take to supply the entire global energy demand of 15 Terawatt by generously assuming that each nuclear reactor can supply 1GW (e). That makes for easy math, and results in no less than 15,000 reactors globally. At 20.5 km2 each, the resulting space requirement is 307,500km2 – just a little less than Poland, or a little bit more than Italy.

That does, indeed, seem like a lot – all of Poland or Italy covered end-to-end in nuclear reactors, supporting facilities, fuel manufacturing plants, etc. etc. to supply the entire global energy demand (that is, all the power currently provided by fossil fuels, hydro electricity, nuclear , and other sources combined).

But how would that compare to other sources of energy under the same assumptions? While Prof. Abbott appears to like solar best, I’m going to do it for wind – simply because I have spent more time analyzing the spatial requirements for wind (mostly because wind power is the only low-carbon, non-hydro, source of electricity cost-competitive with nuclear).

Calculating the space requirements for wind is tricky business. The actual footprint of a wind turbine is not that much: if one includes the swept area, it’s anywhere from .2 – 2 acres (based on data from Enercon, and a little basic geometry. For those who want to dig deeper, the NREL has some good information on this).

Let’s assume we are going to use Enercon’s E101 turbine, which has a nominal capacity of 3,000 kW. Let’s further assume that we can expect an average output is about 25% of rated capacity ( though some studies indicate it is much less, and may be as low as 21%). The turbine has a diameter of 101m – or 331.4 ft – and therefore sweeps an area of about 1.98 acres. Since turbines need to be spaced several times their diameter apart, let’s assume we space them about 10x their diameter apart on average over a perfectly even plane, with nothing breaking the pattern (as I did with the nuclear plants above).

How big would a wind farm with such an arrangement have to be to generate 15TW of electricity?

16,023,693 km2 – a little less than the entire territory of Russia. Or about twice the size of Australia.


Or Canada and Greenland with a chunk of the US:


Even if we reduce the distance between the wind turbines to just 5x their diameter, we’d still end up with a space requirement of 4,005,923 km2 – 22% bigger than India.


What about a bigger turbine, like Enercon’s E126, rated at 7,500 kW a piece (spaced 10x diameter)?

Well, that would require 25,335,374 km2more than Russia and Australia combined.


If spaced only 5x diameter, it would still require 6,333,843 km2 – almost twice the size of India.


[The reason is that the E126 has a diameter of 127m, which results in much greater space requirements even though the output is that much greater than the smaller turbine].

But this would just be the size of the wind farm itself. It would NOT include:

  • all the infrastructure needed to supply the farms
  • all the land lost to mining for the materials from which to build the turbines
  • all the land needed for the manufacturing facilities
  • the housing for all the people who will have to work continuously to maintain the wind farm.

These numbers also assume that

  • no wind turbine will ever fail (because that would reduce the average output),
  • electricity can be stored without any loss of power (because sometimes the wind blows just right, and sometimes not so much – or too much -, and the surplus energy from when it blows just right has to be stored to make up for the other times),
  • electricity can be transmitted without any loss of power (which won’t be the case until we figure out cheap super-conductivity).

So, the space requirements I calculated significantly underestimate the territory required for wind farms, if we wanted to supply all global energy needs with wind alone, while Prof. Abbott’s calculations for the nuclear power seem to significantly overestimate the territory required.

While I admit that supplying all the world’s energy exclusively from nuclear would be a stupendous task, it pales before the challenges of trying to supply it with wind (the only other cost-effective low-carbon, non-hydro source of power).

My logo