I’ve been doing some sums and I can’t quite believe the answers I’ve come up with, so I wonder if some kind soul could help me understand where I’ve gone wrong.
Around the middle of this month I read an article in the Telegraph which gave me pause. It claimed that when the Walney offshore wind farm extension was completed, the 288 square kilometer installation would generate 750Mw of electricity. A quick back-of-the-envelope indicated that, for this installation at least, the electricity generated per square kilometer was around 2.6Mw.
I received quite a bit of help from various people on twitter and in comments to this blog post and have managed to get hold of some figures which I hope are accurate relating to the installed nominal UK wind generation capacity over time, and also five-minute snapshots of wind generation.
The figures I have:
Total nominal installed capacity (on- and off-shore) is very close to 8.5Gw based on tables found at ref.org.uk. Sorting by accreditation date this figure has been reasonably constant since November 2012.
Five minute data taken from Gridwatch.
Wind generation density (calculated as above): 2.604 Mw/square Km. I assume this is a good as it gets, the wind off Walney Island is legendary, and I further assume that offshore installations pack the turbines in as efficiently as possible.
I wanted to work out how much land area we’d need if we were to actually want to produce the claimed 8.5Gw installed capacity. The naive calculation is to divide 8.5Gw by 2.604Mw to arrive at a figure of 3,264 square kilometers. This is incorrect because wind is intermittent, and so we need to reduce that figure by taking the average actual generation which, I am led to believe, is generally assumed within the industry to be 40%. This means we must multiply the area required by 2.5, which gives an area of 8,160 square kilometers. But even than this is just an average. I wanted to know what we’d need to generate this amount of power 24/7, 365 days a year.
Using the data downloaded from Gridwatch, and the following formula, I’ve calculated the area requirement for each five minute snapshot just for the first 12 days of March 2013. The results are incredible.
The highest figure for generation in that period was 4.339Gw, just shy of 51% of nominal capacity, at 15:20 on 10th March. The lowest was 29Mw, at 22:00 on 2nd March. The graph below gives an idea of the spread over that 12 days.

So what area of land would be required to generate 8.5Gw, 24/7, over that period? The calculation is:
Area required = (nominal / actual) * (nominal / density)
where nominal is the nominal capacity, i.e. 8.5Gw, actual is the actual generation at that moment, and density is the Mw/square Km, computed above as approximately 2.604.
Plugging our high/low figures into that function:
when wind generation was briefly at 4.339Gw, the area requirement would have been 6,449 square kilometers. When wind generation was at 29Mw, the land requirement would have been a ridiculous 964,872 square kilometers - almost four times the land area of the UK.
If we are determined to commit to generating any substantial part of our electricity needs from wind power, then we have to choose one of three options:
1) Accept that there will be times - sometimes for quite extended periods - when we are unable to generate sufficient power for our needs, and the lights are going to go out somewhere.
2) Accept that we will need to keep our traditional means of power generation as a backup for when wind fails
3) Accept that every inch of our countryside will eventually be covered in wind turbines.
Of course, this whole argument rests on my calculations and assumptions being correct. I would be grateful for any and all advice, criticism and corrections to my calculations or assumptions. And frankly, I can’t believe I haven’t made a mistake. No sane person who has access to this data - as our government and its advisors surely must - would allow even one wind farm to be build if the return is as poor as this.