Can solar and wind power Britain? An update of David MacKay’s numbers
These technologies have developed a lot since 2008. Should we be any more optimistic about their potential to power Britain?
This Substack was inspired by the work of Sir David MacKay. I wrote about this on the blog’s About page.
Debates around climate, energy and sustainability are often charged with emotion. That’s fine. But we also need to put the numbers on these problems.
This is what David MacKay did for energy: in his book ‘Sustainable energy – without the hot air’ [which is online for free] he used back-of-the-envelope engineering and maths to work out if Britain could power itself from renewables. His answer in 2008 was, unfortunately, no. Far from it. In his final interview – he sadly passed away from cancer – he said the “idea that renewable energy can power the UK is an appalling delusion”. It should focus on nuclear and carbon capture and storage instead.
But the world has changed a lot since 2008. Costs have plummeted in the last decade. New wind designs have come online. And public support for climate action and clean energy has shifted.
David’s approach is still invaluable, but his numbers are out-of-date. This sector is moving quickly, and we can’t use analyses from 2008 to guide policy decisions.
I’ve always wanted to do an update, but thankfully someone else did the hard work for me. Last month, Brian O’Callaghan and colleagues from the University of Oxford published a policy paper looking at the potential for solar and wind to meet Great Britain’s energy needs.1 Here’s their summary.
Far from being an “appaling delusion” they think that its “wind and solar resources are more than sufficient to meet all its energy needs, both practically and economically”.
In this post I’m going to work through their numbers, and how they compare to MacKay’s.
I’ll try to explain clearly where their numbers come from so that you can pick them apart for yourself. I think it’s important that we get some solid numbers because it really does affect policy decisions and public perceptions. I’d be interested to hear your strongest criticisms of the paper.
How much energy could we get from solar and wind?
Let’s start with the headline results from O’Callaghan et al. (2023).
They estimate that it could produce 2,895 TWh of electricity each year from solar and wind. That’s almost double its estimate for final energy demand in 2050. See the chart below.
Both of these numbers are said to be conservative (the supply number is conservatively low, and the demand number is high).
We can see this when we look at other estimates of energy demand from the literature. The National Grid FES projects that Britain will need just 900 TWh in 2050. The UK’s Department for Industrial Strategy says 1250 TWh.2
Total final energy demand today is 1599 TWh. We would expect this to fall with the energy transition because electrification and decarbonisation result in large efficiency gains.
O’Callaghan says that their 1500 TWh demand figure – which is basically unchanged from today – is conservative. It would leave room for new demand for the production of cultivated meats, direct air capture, or intensive manufacturing. I also think it’s important to leave room for hydrogen production for industries that can’t be electrified. Green hydrogen production needs low-carbon electricity: it’s just not the final product.
In any case, they think that Britain could produce more than enough solar and wind to meet this demand.
This wouldn’t come for free. We’d need land (and some ocean) to do it. How much?
The summary of these sources is in the table below.
They think there is a large potential for offshore wind. This would be spread over 10% of the UK’s exclusive economic zone. Onshore wind could be used on 5% of British lands, and combined with farmland. 2% of British land would be used for solar PV, and could also be combined with farmland using a technique called ‘agrivoltaics’. Rooftop solar doesn’t add much – the output is quite small, even if 8% of British rooftops are covered. Definitely still a good option for individuals, but maybe not for the nation as a whole.
You might think some of these requirements are too optimistic. That’s fine: you can quickly adjust the numbers to something more reasonable. If you think the public would only accept 1% being used for solar, half the 544 TWh to 272 TWh. If you think only 4% of the UK’s exclusive economic zone would be usable for floating offshore wind, half the output to around 800 TWh.
Above, our supply was double our 2050 demand, so you could technically half every number in this table and solar and wind would still be sufficient.
How do these estimates compare to David MacKay’s?
Later I’ll go through each of the technologies one-by-one to explain O’Callaghan et al.’s assumptions, and how they differ from MacKay’s. But let’s first compare the big picture.
MacKay calculated figures for Britain’s ‘technical’ resource. He then calculated its ‘practical’ resource once costs and public acceptability were taken into account. This is actually where the main source of difference comes from.
His stack of resources (which includes other renewables, not just solar and wind) was close to meeting demand (which was too high, as I’ll explain later). This is on the left.
But he concluded that almost all of it was practically unfeasible. Solar PV was removed for being too expensive. No one would accept a wind farm near them. Britain could not have offshore wind due to concerns about birds. I think these were reasonable presumptions at the time (although the way he decided to cut them was a bit “hand wavy”). Solar PV was expensive in 2008. And there was not the public support for renewables that there is today.
In the end, he concluded that Britain could deploy just one-seventh of its technical resources.
Here’s the comparison to the recent update. MacKay’s technical resource is very similar to what O’Callaghan thinks is practical.
And here’s the comparison of different sources. O’Callaghan et al. (2023) think that floating offshore wind is now feasible, so that’s much higher. But MacKay’s technical resources for every other source are bigger.
What has changed, then, since MacKay’s book in 2008?
Energy demand will be much lower. MacKay primarily used primary energy demand – which includes all of the inefficiencies that would disappear with electrification and thermal inefficiencies from burning fossil fuels.
By my calculations, his ‘demand’ stack totalled up to 2700 TWh for the UK (around 2600 TWh for Great Britain).3 That’s 2 to 2.5-times the expected demand for 2050. Energy demand in the UK has already fallen by around one-quarter since 2008, and that’s before mass electrification.
Later in the book he does look at consumption scenarios in 2050: these add up to around 1700 TWh of energy demand.4Solar and wind costs have plummeted. In the last decade the cost of solar power has fallen by around 90%, and wind by 70%. They have gone from being very expensive to being cheap.
These technologies are not just cheaper, they’re better. Solar panels are more efficient, and bigger and more effective turbines have been developed.Support for renewables is higher. In the latest survey by the UK Government, 85% of the public supported wind and solar projects. Only 1% were opposed to renewables. Three-quarters think renewables are good for the economy.
Just 12% said they would be unhappy about having an onshore wind farm in their area, and just 7% for a solar farm.5Renewables are being paired with farmland. In the past, there was often the assumption that renewables would have to be placed on new, additional land. We’d have to ‘give up’ farms or other land uses to make space for them. But, wind power is now being combined with farmland, and solar on farmland (agrivoltaics) has growing support.
If MacKay had written his book in 2023, he wouldn’t have waved away as much solar PV for being ‘too expensive’ or wind power for NIMBYism (“not-in-my-backyard”).
At this point, I’m going to go through each of these technologies one-by-one. Some of you might not be interested in the specific details. If so, feel free to drop off here.
The conclusion you should take away from this update is that the prospects for renewable energy in Britain are not as bleak as they were in 2008. Solar and wind power could outstrip our demand – and we wouldn’t need to cover the country with turbines and panels.
This is true, even if you think the updated numbers are too optimistic. You could assume that there was no floating offshore wind, and supply would still match demand.
To be clear: this does not mean that this is the ‘optimal’ electricity mix in 2050. Not least because energy storage costs would be very high. We would probably want to diversify a bit, not least to help with grid balancing. Before all of the nuclear fans get mad: I think there’s room for nuclear in there too.
But the point still stands: it seems we have a lot of untapped solar and wind resources and they could make up a large chunk of our grid, even if they’re not 100% of it.
Technical comparisons
Offshore wind
Recent updates suggest that our potential for offshore wind is much higher than David MacKay estimated. This is because of the development of floating offshore wind.
A major limitation to fixed offshore wind is water depth: they can’t be installed in very deep waters because they need to be fixed to the seabed. Instead, floating offshore wind is built on a platform, which is then anchored to the seabed. Here’s an idea of what they look like. This means we can have wind farms further offshore, where there is even more wind potential.
The differences in estimates are shown in the chart below.
First, let’s start with MacKay’s numbers. He assumed that 120,000 km2 of water was available for offshore wind. He removed two-thirds of this area because it conflicted with shipping corridors and other possible uses. That means 40,000 km2 as his technical estimate.
He then reduced this to 5,000 km2 as his practical estimate as a result of cost and poor public support. Why he divided by 8 to get this final figure is not clear.
O’Callaghan et al. (2023) estimate that there is between 2,500 to 12,000 km2 available for fixed offshore wind, and 62,000 km2 for floating offshore. For their fixed offshore estimates, they removed any farms within 25 miles of the shore due to the visual impact.
That equates to 2% of the UK’s exclusive economic zones for fixed offshore wind, and 8% for floating offshore.
The differences in area estimates are given in the table below.
The authors give three reasons for their discrepancy with MacKay’s estimates.
Improved fixed turbine technologies. MacKay described all waters deeper than 30 metres as “not economically feasible”. But fixed turbines could soon be commercially feasible to around 80 metres of depth.
Floating offshore turbines. The advent of floating turbines means that these can extend into much deeper waters. Note that O’Callaghan et al. (2023) already account for many other competing uses such as fishing areas, military zones, shipping routes, and low-wind areas.
Improved social and political support. O’Callaghan et al. (2023) assume there is less public resistance to offshore wind, which seems appropriate.
Finally, a few details on how O’Callaghan et al. (2023) calculate their wind output. They assume an average capacity factor of 50%, a turbine size of 15 MW, and a spacing of 7 times the turbine diameter between them.
Onshore wind
David MacKay estimated that around 10% of British land could technically power wind turbines. That would produce around 449 TWh per year. Based on concerns about public support, he reduced this to 1.5% for his ‘practical’ estimate. That gives 67 TWh. See the chart below.
Various studies have suggested that much more of Britain has sufficient wind potential: ranging from 18% to 44% of land.6
In their update, O’Callaghan et al. (2023) take a conservative estimate of 5% of British lands having wind farms – all of this would be on shared agricultural land. That would mean one-thirteenth of the country’s farmlands would be shared with wind projects.
To get this 5% figure, they did the following. Took the average of previous studies looking at the area of land with high wind potential: that gave 23% of land. They then cut this figure by three-quarters to account for the fact that many landowners would be resistant to the installation of wind, and some local communities would oppose it too.
Most of the ‘land’ used for farming is actually open space: it’s the unused spacing between the turbines. The amount physically impacted by the installation of the turbines, roads, and infrastructure would only be 0.05% of British land. For context, around 0.9% of land in English is used for quarrying.
Finally, a few details on how O’Callaghan et al. (2023) calculate their wind output. They assume an average capacity factor of 38%, a turbine size of 7 MW, and a spacing of 6 times the turbine diameter between them.
Solar PV
Solar PV has developed rapidly since 2008. Cells have become more efficient and prices have plummeted.
For his technical estimate, MacKay assumed that 10% of British land was used for panels with a 10% efficiency. Irradiance in the UK was assumed to be 110 W/m2. For rooftop solar, he assumed 10% efficient cells on south-facing roofs.
As you can see in the chart, his ‘practical’ estimate was reduced to almost zero. He thought solar was far too expensive (which was true in 2008, but not today).
O’Callaghan et al. (2023) assume that 2% of British lands could be used for solar. This would be combined with agricultural lands – ‘agrivoltaics’. The average solar irradiance in the UK is 101 W/m2. This accounts for night-time and cloud cover. The authors cut this 101 by 55% to account for the spacing of the panels and added 15% to adjust for optimal tilt. That gave them 52 W/m2. They assumed a panel efficiency of 25%, covering 2% of British land, which is 4,740 km2.
Multiply these numbers together, and the number of hours in a year (8,760) and we get 544 TWh. For rooftop solar, they estimate just 25 TWh.
What are the major differences to MacKay?
Lower cost: the price of solar PV has fallen by 90% in the last decade. MacKay’s main concern was it was too expensive: this is not the case today.
Improved cell efficiency: MacKay used a cell efficiency of 10%, and thought an efficiency of 30% would be “quite remarkable”. Last year, Fraunhofer ISE achieved 47.6%. Efficiencies greater than 30% have also been achieved using perovskite solar cells. O’Callaghan uses a cell efficiency of 25%. See the footnote for a chart showing the differences in efficiency over time.7
Combined use with agriculture: MacKay assumed an inherent trade-off between agricultural land and solar. He thought the Brits would never give up farmland for solar panels. But this trade-off does not always exist: there are now a range of projects where solar and agriculture work in tandem (‘agrivoltaics’).
Comparison to our energy mix today
Regardless of how much potential solar and wind have, we are still very far away from a low-carbon energy mix.
In the chart below I’ve compared the UK’s electricity generation from low-carbon energy sources (which is broader than solar and wind – it has decent amounts of nuclear and bioenergy) in 2022, to the various estimates of final energy demand in 2050. Supply data comes from Ember Climate.
We need to produce 5 to 7 times as much low-carbon energy as we do today. Better get building quickly.
O’Callaghan, B., Hu, E., Israel, J., Llewellyn Smith, C., Way, R. and Hepburn, C. (2023). Could Britain’s energy demand be met entirely by wind and solar? Smith School of Enterprise and the Environment, Working Paper 23-02.
The IEA’s estimates are based on announced policies, not a scenario where net-zero is achieved so energy demand is higher (remember that the energy transition will reduce energy demand).
He estimated 125 kWh per person per day.
Multiply this by 365, and we get 45,600 per person per year.
The UK population at the time was around 59.5 million. For Britain only, it was around 2 million less.
Multiply 45,600 by 59.5 million, and divide by a billion = 2715 TWh.
Or 2623 TWh for Great Britain.
He estimates 68 kWh per person per day.
For a 2050 population of 71 million, this works out at 1700 TWh.
Note that around 45% said they would be happy about having these technologies in their local area. So, a large chunk of people were in the middle: they didn't oppose or strongly support it.
McKenna, R. V., Hollnaicher, S., vd Leye, P. O., & Fichtner, W. (2015). Cost-potentials for large onshore wind turbines in Europe. Energy, 83, 217-229.
Eurek, K., Sullivan, P., Gleason, M., Hettinger, D., Heimiller, D., & Lopez, A. (2017). An improved global wind resource estimate for integrated assessment models. Energy Economics, 64, 552-567.
Enevoldsen, P., Permien, F. H., Bakhtaoui, I., von Krauland, A. K., Jacobson, M. Z., Xydis, G., ... & Oxley, G. (2019). How much wind power potential does europe have? Examining european wind power potential with an enhanced socio-technical atlas. Energy Policy, 132, 1092-1100.
Ryberg, D. S., Tulemat, Z., Stolten, D., & Robinius, M. (2020). Uniformly constrained land eligibility for onshore European wind power. Renewable energy, 146, 921-931.
A question on storage. One of the big issues with solar/wind is variability of supply - hence the need for base load generation or storage. I was looking at the eye-watering estimates for the new nuclear stations and wondered how that compared with, say, enough batteries to store a week's worth of energy from a power station. I appreciate there will be loads of practical considerations as well (e.g. how big would it be??) but as a starting point how do the costs compare?
Some of the comments have alluded to this. Some in a nice way and some in a not so nice way. I think this adage applies, "Just because you can, doesn't mean you should."
I continue to believe that questions like these are really best left to systems experts that do tradeoff studies, e.g. Net-Zero America Project. Here is a link to their summary, https://netzeroamerica.princeton.edu/img/Princeton%20NZA%20FINAL%20REPORT%20SUMMARY%20(29Oct2021).pdf .
Because of variability both seasonal and daily for renewables, you have to overbuild. This is seen clearly in the graph on pg. 25 where the annual generation for an all renewable system is the largest. This also comes with increased cost. An important point is that this study is already a couple years old and the results would probably be different today. The point being that experts in system studies are continually reassessing the situation. However, as Dr. John Bistline wrote in an NYT op-ed concerning the U.S, we shouldn't worry so much about the final mix of technologies but we should do the things we know we need to do,
"For the coming decade, rapidly reducing coal electricity and building extensive wind, solar and storage systems are low-cost strategies in many places, regardless of how much energy might or might not eventually come from renewables."
https://www.nytimes.com/2022/04/10/opinion/environment/ipcc-report-climate-change-debates.html .