A new paper has been published which raises questions about the amount of electricity that can be generated by large scale wind farms. You can get a free copy of it here, as well as watch a video by the lead author David Keith. The conventional view is that, on average, you can generate between 2 and 3 Watts per square metre in wind farms (note this is electricity generated not capacity of wind farm per square metre). Adams and Keith provide references in their paper that can be checked, but for a quicker (and paywall free) explanation you may want to read this blog post by David MacKay.
What Adams and Keith suggest is that for wind farms greater than 100 square kilometres the power density appears to saturate at about 1 watt per square metre, essentially the turbines have a slowing effect on the wind. They use mesocale modelling for this, however I am not a climate scientist so will not embarrass myself by either trying to explain the model or criticize it.
For some context energy use in a densely populated country such as the UK is just over 1 watt per square metre. A lower power density of large scale wind will mean a large chunk of land must be covered in wind turbines to meet a significant portion of energy demand. Simple back of the envelope calculations show that this could lead to problems in some countries. Take Japan, which is 60% forested. If large scale wind does saturate at 1 watt per square metre, then there probably is not enough land in Japan for onshore wind to provide half of its energy.
Let’s however consider what this could possibly mean in practical terms. 100 square kilometres is in fact the exact size of the soon to be opened first phase of the London Array offshore wind farm, which is currently the world’s largest offshore wind farm. The total capacity of the wind farm is 630 MW. If the power density only turns out to be 1 watt per square metre then it will have a load factor of 16%. This is about half of what I expect the wind farm is hoping for, and much less than the current average of about 27% for UK wind farms. The London Array will be operational in a couple of months according to its website, so we can probably know for sure within a year if the model of Adams and Keith is something we should be immediately concerned with.
What about existing and operating windfarms, can they tell us anything? Europe’s largest operational windfarm is Whitelee Windfarm, just outside Glasgow, and I can see it every day on my walk in to work. The operational part of it has a capacity of 332 MW, and covers 55 square kilometres. Its current load factor is 28.7% (taken from the anti-renewables lobby group REF’s website, though their data seems reliable). So the energy density is about 1.7 watts per square metre, and if we considered the 100 square kilometre region it is contained within it works out at 0.92 watts per square metre. The wind farm is currently expanding from 332 to 539 MW, with the full 539 MW expected to be fully operational next month. Again, we should know in a year if the density will be much higher than 1 watt per square metre, but it would be surprising if it was not. And certainly the finances of the wind farm will not be in good shape if it is not.
There are a few wind farms in America with total capacities greater than 500 MW. Most of them appear to wind farms split over multiple sites, so aren’t particularly useful for testing Adams and Keith’s results. However Shepards Flat Wind Farm is more or less perfect. It covers 78 square kilometres and has a total capacity of 845 MW. A power density of 1 watt per square metre would give a load factor of about 12% for this wind farm, when something like 30% might be expected. Shepards Flat opened in September 2012, so again we should know quite soon if Adams and Keith’s result holds for functioning wind farms.