Fact Check: Comparing wind and nuclear power

Posted on Updated on

Climate Scientist Kevin Anderson has criticised claims made by Sue Ion, a former technical director of British Nuclear Fuels, made on the BBC. Anderson makes the following quite strongly worded criticism of Ion:

Early on in the programme Sue emphasised how she is committed “to try and do more to help get facts across as opposed to just let the media run with whatever they thought … sometimes stories run when they actually do have no foundation in fact”.

Certainly the world of energy and climate change is awash with educated eloquence trumping quantitative analysis – and any attempt to rescue the latter from the former has to be welcomed.

However, despite Sue Ion’s concern about energy stories often having “no foundation in fact”,when it came to drawing comparisons between electricity generation from nuclear and wind power her comments only added to the misinformation that pervades energy debates.

Sue Ion suggests 1500 offshore wind turbines generate the same electricity as one nuclear power station; the real number is much lower – somewhere between 250 and 600.

So, Ion apparently had a figure that is three times higher than it should be. The comments by Ion can be heard about 20 minutes in. [Update:  in the comments section David has pointed out that Anderson has misunderstood what Ion was referring to. The question asked to her was in relation to a “nuclear power station,” not an individual reactor. Historically nuclear power stations in the UK have had two reactors, in France they often have four. The new nuclear power stations planned for the UK have more than 1 reactor, so this presumably is what Ion has in mind. So, Anderson should really be multiply his nuclear GW by 2.]

Below is the argument Anderson uses to counter Ion’s claim:

Calculations and Assumptions

The following calculations are premised on proposals for new-nuclear build,
assuming full operation by 2020 and assuming the load factor is significantly
improved from the UK’s experience of operating nuclear stations. The figures for
wind turbines similarly are premised on appropriately sited designs, with a good
capacity factor and assuming turbines at sizes equivalent to the larger models
now being installed and those likely to be installed before 2020.


Three reactor designs are now being considered for UK new-build.
§ Areva’s EPR – with a capacity of 1.6GW
§ Westinghouse’s AP1000 – with a capacity of 1.15GW
§ Hitachi-GE’s Advanced Boiling Water Reactor – with a capacity of 1.3GW
Assume an 85% load factor for all the nuclear designs
Note: this is 5% higher than is sometimes suggested should be the starting value
for nuclear load factors, and is 25 percentage points above the mean UK load
factor for nuclear power between 2007 ad 2011 (i.e. 60.1%).


Currently, installation of 6MW turbines is proceeding, with 8MW designs planned for
installation by ~2014, heading towards 10MW within a few more years. Some
companies are already proposing designs of up to 15MW per turbine. For comparing
with the nuclear designs operational by 2020, these calculations assume 6 to 10MW
turbine designs.
Assume a conservative capacity factor of 40% for offshore wind turbines.
For a well-sited large and offshore turbine, a 40% capacity factor is not
unreasonable figure to assume. It is worth noting, asthe capacity (MW per turbine)
increases, so does the hub height and hence the typical capacity factor. Moreover,
if sited off the West coast of the UK the capacity factor is likely to be higher still.


The three nuclear designs with a 85% load factor would generate between 8.6TWh and
11.9TWh each year
A 6MW and 10MW wind turbine with a 40% capacity factor would generate 21GWh
and 35GWh/year respectively.
Consequently, between 244 and 567 turbines are required to generate the same
quantity of electricity in a year as the three proposed new-nuclear designs

Who is correct?

Sadly, it appears that while Anderson is probably correct that Ion is making an inaccurate claim, he is also doing exactly what he has criticized Ion for doing, by correcting misleading information with misleading information.

First, consider that he uses turbine capacity of at least 6 MW. A quick look at Wikipedia will tell you that the average for existing UK offshore wind farms is slightly below 3.6 MW, with most wind farms using a 3.6 MW turbine. You can also see that 3.6 MW turbines predominate in the offshore wind farms that are currently being constructed. Ion’s remarks appear to clearly relate to existing wind farms, so exactly why Anderson is using a turbine capacity this large is unclear.

Anderson also claims that a 40% capacity factor for offshore wind farms is conservative. This may or may not be true for new offshore wind farms with 6 or 10 MW turbines, however for existing offshore wind farms it appears to be too high. Historic offshore wind farm capacity factors can be found from the UK government’s Department for Energy and Climate Change. These have been around 33-35% over the last few years.

So, let’s rephrase Ion’s claim and ask how many 3.6 MW turbines, at the UK average capacity factor of 35%, would provide as much electricity as a modern nuclear power plant. To get the same power as the 1.6 GW would take about 1,100 turbines. Matching Hitachi’s 1.3 GW reactor would take about 880 turbines.

What we are looking at is one of those rare occasions where the truth lies somewhere in between, and another example of people using numbers to make a point, and not to inform. [though I’ll now take back my comment about the truth lying somewhere in between, as Ion seems to have produced very reasonable numbers given that she was actually referring to a nuclear power station.]


21 thoughts on “Fact Check: Comparing wind and nuclear power

    David said:
    February 27, 2013 at 10:33 pm

    Well, not having listened to original broadcast I still find myself experiencing the uncontrollable urge to point out that Sue apparently said ‘plant’ – which to the experienced nuclear industry ear really refers to multiple units, most likely 2. I bet she’s probably thinking EPR – so 2 units each at Hinkley Point C and Somerset as per what’s on the books. They are two-unit plants. So tweak a few of the finer parameters and I think Sue is right back on the money. Kevin can suck it.


      Robert Wilson said:
      February 27, 2013 at 11:01 pm


      The question asked was about a modern “nuclear power station.” So, you might be correct. I know a lot of the French multiple unit set ups are referred to as nuclear power stations. So, in a European context a “nuclear power station” probably averages over 2 GW, as far as I can tell.


        David said:
        February 28, 2013 at 12:38 am

        Robert, that pretty well settles it. ‘Station’ from a historical UK perspective really means two units, and the only new reactor design with GDA approval is the EPR. Kevin pooched it and owes Sue an apology.


        Robert Wilson said:
        February 28, 2013 at 1:00 am


        I’ve updated the post. You’re right it’s common use to call something with multiple units a power station. Drax has six but is referred to as a power station. Single unit modern power stations are pretty rare, so it’s safe to assume she is thinking about something with at least two reactors. So at least 2.6 GW is what she had in mind. So, her numbers seem reasonable.


    timprobert said:
    February 27, 2013 at 10:42 pm

    The next generation of UK wind farms will use 6 MW turbines. Siemens and Alstom are already making them; DONG has installed two of Siemens’ machines at Gunfleet Sands and are now operational.

    While 10 MW is not yet a commerical prospect, it is no less fanciful to use 6 MW as a basis for “modern wind turbines” than using Areva’s EPR reactor for nuclear, which while under construction in Finland, France and China is yet to generate a single watt of power and would not in the UK for 8-10 years hence.


      Robert Wilson said:
      February 27, 2013 at 11:03 pm


      This may be true. However, having re-listened to it the point Ion was making was about the area required for the wind farm, which as far as I understand it is not particularly independent of turbine size.

      And your point about using the EPR reactor is more than little facetious.


        timprobert said:
        February 28, 2013 at 11:20 am

        Why facetious?

        My point is that using 6 MW for wind as a comparison is valid. Possibly more valid than using the 1.6 GW Areva EPR reactor, which as I say won’t be operational for 8-10 years in the UK,
        for the top end of the nuclear figures.


    smipp Ltd. said:
    March 1, 2013 at 12:24 pm

    Reblogged this on SMIPP Ltd..


    Kevin Anderson said:
    March 2, 2013 at 1:38 am

    I disagree with the response

    Fair comparisons need to compare like with like!

    I made absolutely clear my calculations were for comparing proposed new build nuclear in 2020 with an assessment of typical large offshore turbines in 2020; for which I assumed 6MW at the lower end (as these are already being built and operated) and 10MW for the upper end.

    The 2020 date here is pivotal and is being missed by some others’ comments and calculations, where they compare a new nuclear plant operating in 2020 (or possibly 2018) with current operating offshore turbines; this is not like with like!

    Personally, from a policy and investment perspective I think there is not much benefit in comparing today’s technologies; but if we are to do so it needs to be on a fair basis.

    So here goes:
    5MW turbines are already generating electricity offshore around the UK (e.g. Ormonde wind farm). So I suggest a fair comparison for today is the UK’s largest nuclear station (Sizewell B) at 1.191GW (according to British Energy) with the largest turbines at 5MW.

    In terms of capacity factors if, as some comments suggest, the last few years of offshore wind is to be used as a guide, then so should the last few years for nuclear generation, i.e. a five year mean load factor of 60% (see DUKES)

    Using these figures and the number of turbines is 408 to 433.

    I will finish by stating that I am agnostic about nuclear power, but hold strongly to the view that comparing the various merits of different options and portfolios of options needs to be done fairly. Moreover, in considering the future, whilst we need to be guided by history, assuming the future is the same as the past is unhelpful. This is why my original calculations took a positive view about nuclear and wind.

    Liked by 1 person

      Robert Wilson said:
      March 3, 2013 at 1:10 pm

      Thanks for the reply Kevin,

      I’ll just make a couple of points here, because I’m trying to finish a draft of something today.

      1. As I said in the blog post Ion was referring to modern nuclear power stations, the stations part being key. More or less every nuclear power station in the world has at least 2 reactors, and all of the planned nuclear power stations in the UK have aims to have 2 reactors. So, I believe it is fair to assume this is what Ion had in mind. So GW value of at least 2600 should be used, based on the EDF and Horizon reactor capacities.

      2. I disagree that 6 MW should be used as a low ball figure for new offshore wind. 6 MW turbines are currently in the testing phase. Considering that round 3 offshore wind farms are going under construction in 2016 (e.g. the Dogger Bank wind farm http://www.forewind.co.uk/projects/dogger-bank-creyke-beck.html) it is very difficult to see why you would assume 6 MW is a low ball figure and not a high ball figure.

      3. Talking in terms of turbines is not a good idea, instead we should be talking in terms of watts per square metre. Ion probably had the more common 3.6 MW turbines in mind, you can argue that we should use 6 or 10 MW. However does this really make a significant difference to the actual area of the sea that you will need to cover in turbines to get the power?


      David said:
      March 3, 2013 at 2:49 pm


      To be honest, the question of *precisely* how many offshore turbines match one new nuclear generating station is a dull one – especially when you are just playing around with your own (very questionable) preferred choice of inputs. However, I find myself rather vexed that you have leveled accusations at Sue on this, without bothering to first check the inputs she actually had in mind. Why not give that a go?

      The way I see it, you have misinterpreted her badly as Robert has now distinctly pointed out to you. Your statement here merely compounds your earlier mistake. By all means continue with this arbitrary and fairly irrelevant comparison, but have the decency to retract the comments you made against Sue lest you “inadvertently risk misleading both the public and policymakers.”

      You claim to be a nuclear agnostic, yet your comments on nuclear (unit capacity and capacity factor) strongly suggest otherwise. Why don’t you try applying some of the same fond future-gazing optimism to nuclear power as you do to offshore wind. Admittedly the development cycles are a bit slower, but you should find there is plenty of tech on the near horizon to get excited about.


    Kevin Anderson said:
    March 4, 2013 at 12:22 am

    In reply to Robert – and some of the other blogs.

    1. The question about stations, reactors or sites may certainly be open to interpretation; however the question posed by Jim Al-Khalili was: “the amount of energy produced by one modern nuclear power station – how may wind turbines would that need?”

    Given the Radio 4 programme was aimed at the lay public, I am unsure as to whether the typical interested rather than expert listener, or indeed the presenter, would have understood “one modern nuclear power station” to have meant ‘as many reactors as you can fit on a site’. But perhaps that is what they thought. Only asking the presenter and polling listeners could give an answer to this.

    2. Despite various comments and suggestions, I remain of the view that 6MW by 2020 is completely reasonable – if not conservative; please note the 2020 date! Are those that disagree, suggesting 6MW turbines are unlikely to succeed – or that this is some threshold of capacity – there are certainly turbine manufacturers that think differently. Again, please note I was and am talking about 2020.

    3. I’m very unsure about watts per square metre as being a useful measure. But if it is to be used I think it would need a time dimension and include all the fuel phase for options being compared, and, from an LCA perspective, the construction, decommissioning and waste phases too. So the exploration, extraction, enrichment, land transport, ports, ships etc, would need to be factored in, along with any space used in relation to the waste. Then this would need to be considered over time – to give a watts per square metre-year. As I say, I think this is unhelpful – but I certainly have no idea how the numbers would pan out – or how waste storage over decades to centuries should be factored in. In addition, some estimate of the frequency and severity of nuclear & wind accidents would need to be assessed and any changes in use of areas impacted considered – again measured in square metre-years. Personally, I’d stick with number of turbines being a least-worst proxy – after all, such figures are just a guide.

    Two final points.
    A) I get a sense (which I acknowledge may be incorrect) – that those arguing against my numbers are unreasonably favouring nuclear over wind in terms of load factors, capacities and operating dates. I took the view from the start that a new nuclear plant could be operating in the UK by 2020 at 85% load factor and assuming one of the three designs outlined. Similarly, I took the view that the 6MW wind turbines now being installed could be typical by 2020, and quite possibly could have reached 10MW. Moreover, that a 40% capacity factor was not unreasonable to assume for well sited arrays of large offshore turbines by 2020. For both nuclear and wind I took what I continue to hold to be a positive and achievable view of the future. However, I find it difficult to understand how the repeated disregard, by some bloggers, of the 2020 date along with their dismissal of both the 6MW capacities and the 40% capacity factor (as viable norms for 2020), can be said to represent a balanced contribution.

    B) Finally, on reflection I recognise that the tone of my original comment could have been construed as unnecessarily critical of Sue Ion. This was not my intention – rather I was and remain committed to trying to analyse and compare all energy options fairly, with the relative merits and drawbacks considered openly and honestly. So whilst I hold to the conclusion of my initial response to the Sue Ion interview, I wish to assure her, should she be following any of this, that my observations were not personal, but about the substance of her comments on the numbers of turbines and informing the media.


      Robert Wilson said:
      March 5, 2013 at 10:58 am


      On watts per square metre vs. number of turbines. While there is no perfect metric here I still believe watts per square metre is much more helpful than number of turbines. The key issue here is land use impact, and visual impact. Let’s say you go from 3.6 to 6 MW turbines this will roughly halve the number of turbines needed, however the total area needed for a particular wind farm may not change a great deal. Consequently objections about the impact of it on shipping lanes etc. will probably stay as they were before. Making the turbines bigger may even make planning more difficult because it will increase the visual impact of offshore wind farms.

      It’s not a perfect metric, and the importance of it will very depending on the interests of particular groups, but I really don’t see how number of turbines needed to match say a 1 GW nuclear power plant really tells us anything.


    jmdesp said:
    March 5, 2013 at 12:33 pm

    – The answer to “as many reactor as you can fit on a site” is 8, that’s the number Duke power has in Ontario. There’s nothing sophisticate in thinking that when saying “a new NPS” the lay public in UK will think the reference is Hinkley Point C, which is *the* power station that’s in the talk currently and comprises 2 reactors at 3200 MW.
    – OTOH 6 MW might be a quite correct value for wind power that will be completed in 2020. REpower has one already, BARD has a 5MW one, Areva and Alstom are both preparing one for their current projects.
    – About the load factor it’s clearly quite below 40% currently in UK (Denmark reaches it with better winds), this DECC reports assumes 29% for turbines above 5MW (page 25) :

    With those numbers, I actually exactly get 1500 turbines = (3200 * 0.85) / (6 * 0.29) = 1 563.
    I swear I didn’t modify anything to try to get that result.


    Jani-Petri Martikainen said:
    March 5, 2013 at 2:28 pm

    I would say W/m^2 is somewhat useful metric and especially so when it is very low. This is the case for bioenergy and sometimes a case for wind and solar. Why sometimes? Because the power density of both wind and solar fluctuate wildly with a high probability of power density being practically zero. One issue where power density has an indirect role is in materials consumption and on the size of the construction projects. A low power density source implies a construction work which typically requires much more concrete, steel etc. spread over much larger areas. This adds to the costs and makes it less likely that such power source will ever see costs coming down sufficiently to compete with fossil fuels (or nuclear) on a level playing field. This is “heuristic” argument, but there are reasons why societies have started using power density sources and moved away from wind, solar and biomass when the industrialized.

    Other useful metrics are….

    1) Does the power source deliver the kind of energy we use? This is relevant because if it doesn’t the gap will be filled most likely with fossil fuels. Demand could be reduced somewhat, but just check how little the demand moves during “Earth hour” , for example, and you see that most of the society has no interest in making their energy consumption conditional on weather. I certainly don’t.

    2) How much energy goes into building the energy infrastructure relative to how much is produced (EROEI)? This is terrible for liquid biofuels and pretty terrible for solar in bad places. Wind is not that terrible. As far as I remember EROEI less than about 8 implies difficulties since then energy sector itself starts to consume outsized share of all energy consumption in the society.


    Kevin Anderson said:
    March 6, 2013 at 4:02 pm

    Watts per square metre risks being a post hoc justification of a technology.

    How do you factor the fuel stream into the watts per square metre figure? Without a fuel stream you may have watts but you’ll certainly not have watt-hours – and it’s the latter we consume.

    How do you factor in time? So if you have some high-level nuclear waste that stops the use of an area for 100 years – or perhaps longer, how is that factored into the calculation?

    How do you factor in shipping – e.g. the UK’s large import of gas from Qatar being discharged at Milford Haven – do you include the port facility at both ends – as well as the gas production area? Or if coal – how about the train lines to the power stations?

    If coal or gas is assumed (with or without CCS), how you do factor in the areas of land lost through sea level rise, or desertification etc? And then how do you factor in time? If the sea level rise inundates a square metre which is then underwater for 10000 years – how is that factored in? (note – even gas CCS is unlikely to be below ~80gCO2/kWh for the full combustion and fuel stream)

    Isn’t there a risk we choose an ‘objective’ measure that favours our preferred technologies? Certainly, in the absence of some detailed methodology, including considerable discussion around factoring in time, along with a thorough analysis of the fuel stream, transport impacts, decommissioning, waste handling (nuclear) and waste impacts (fossil fuels), I’m unsure how useful a dangerously partial assessment of a contested measure really is?

    Please note: I’m not suggesting number of turbines is a particularly good measure – but for most lay folk it probably means more than a highly partial watts per square meter number.


      Robert Wilson said:
      March 6, 2013 at 5:42 pm


      2 points. I don’t see how this is a partial measure. I am not saying we base these decisions purely on W/m2. The real issue is looking at the potential problems each technology has. If you look at onshore wind then W/m2 is far more useful than number of turbines. A back of the envelope calculation will quickly show how much of the UK you would need to cover in wind farms to get say 50% of UK power supply. This calculation makes clear that truly large scale offshore wind in the UK is probably not likely to happen to due to local opposition. And it’s quite clear there is an unspoken consensus on that issue, otherwise we would not be having discussions about offshore wind farms, which are much more expensive than onshore wind farms.

      I can give another example which demonstrates just how useful the W/m2 can be: corn ethanol. The energy density of corn ethanol works out at about 0.22 W/m2 (http://www.oecdobserver.org/news/fullstory.php/aid/2083/21st_century_energy:_Some_sobering_thoughts.html). Again, a back of the envelope calculation shows that scaling corn ethanol up to provide a significant proportion of US transport fuel will cause serious land use problems. So, in this case attention to power density would have prevented significant environmental and humanitarian harm.

      Finally I still am not persuaded that the number of turbines is a helpful metric. What exactly does it tell us? Let’s say we want to get an average of 20 GW from onshore wind. Let’s say with one turbine size we could get this with 20,000 turbines, but with another we could get it with 10,000. Is 10,000 turbines better? Maybe from an economic point of view. However increasing the size of the turbines will increase the visual impact, and probably just increase opposition to wind power. Personally I don’t have any real problem with the look of wind farms, and can see Whitelee wind farm from where I live. However, most people are not like me. In a densely populated country land requirements are a serious issue for renewables, so we must have metrics that take them in to account.


    Kevin Anderson said:
    March 7, 2013 at 4:46 pm

    Given your and my concern over land use, how do you factor in the impact on land from climate change associated with the emissions from CCS coal or gas (say at lifecycle 80-150gCO2/kWh)?

    Moreover, if comparing ethanol with diesel/petrol on a w/m2 basis – how again do you factor in the climate impacts on land (& sea from acidification through raised CO2 levels)? If such w/m2 impacts are not included, isn’t there a risk of highly/unfairly skewing any conclusions?

    Going back to the original nuclear – wind comparison, I get the impression you think my time argument is without much merit– could you explain why that is? And also, do you consider the w/m2 metric is appropriate without considering the full fuel and waste streams? If so, could you explain why such a partial analysis is helpful – especially when comparing fossil & nuclear with renewables (at least non-bio), where the latter has little to no fuel or waste streams (at least from operation)?




    jmdesp said:
    March 12, 2013 at 6:26 pm

    About nuclear – wind comparison, I think you feel speed of construction is the definitive argument that dooms nuclear. But is it written that we are utterly unable to be as efficient as Chinese and Koreans ?

    Let’s take China that is strongly building both nuclear and wind :
    – In early 2010, it had 9GW of nuclear and 30 GW of wind
    – For 2015, the plan is 41GW of nuclear and 100 GW of wind. The added 32 GW of nuclear would on a year generated more power than the added 70 GW of Wind. Despite a very fast start on wind, the target is still 100GW, since there was integration problems.

    You think 41GW is impossible ? 5,5 have already been added from 6 units of an average construction time of 55 month, and construction of the 24 units remaining for 27 GW is definitively looking good even if 2 are now planed for 2016. It’s almost certain the 2 EPR will be completed in time with a 52/53 month construction time (operating test for the 1st are at end of this year, Areva is already completing the fuel assemblage).

    BTW about your newer comparison, Sizewell-B has a load factor of 82.5%, 60% is the number for the 30 years old reactors.


    digital foto frame said:
    May 27, 2013 at 11:52 am

    You know therefore significantly when it comes to this topic, produced me for my part believe it from a lot of varied angles. Its like women and men don’t seem to be fascinated until it’s one thing to do with Woman gaga! Your personal stuffs excellent. Always maintain it up!


Comments are closed.