Wind Blowing Nowhere

In much of Europe energy policy is being formulated by policymakers who assume that combining wind generation over large areas will flatten out the spikes and fill in the troughs and thereby allow wind to be “harnessed to provide reliable electricity” as the European Wind Energy Association tells them it will:

The wind does not blow continuously, yet there is little overall impact if the wind stops blowing somewhere – it is always blowing somewhere else. Thus, wind can be harnessed to provide reliable electricity even though the wind is not available 100% of the time at one particular site.

Here we will review whether this assumption is valid. We will do so by progressively combining hourly wind generation data for 2013 for nine countries in Western Europe downloaded from the excellent data base compiled by Paul-Frederik Bach, paying special attention to periods when “the wind stops blowing somewhere”. The nine countries are Belgium, the Czech Republic, Denmark, Finland, France, Ireland, Germany, Spain and the UK, which together cover a land area of 2.3 million square kilometers and extend over distances of 2,000 kilometers east-west and 4,000 kilometers north-south:

Figure 1:  The nine countries

We begin with Spain, Europe’s largest producer of wind power in 2013. Here is Spain’s hourly wind generation for the year. Four periods of low wind output are numbered for reference:

Figure 2:  Hourly wind generation, Spain, 2013

Now we will add Germany, Europe’s second-largest wind power producer in 2013. We find that Spanish low wind output period 4 was more than offset by a coincident German wind spike. Spanish low wind periods 1, 2 and 3, however, were not.

Figure 3:  Hourly wind generation, Spain + Germany, 2013

Now we add UK, the third largest producer in 2013. Wind generation in UK during periods 1, 2 and 3 was also minimal:

Figure 4:  Hourly wind generation, Spain + Germany + UK, 2013

As it was in France, the fourth largest producer:

Figure 5:  Hourly wind generation, Spain + Germany + UK + France, 2013

And also in the other five countries, which I’ve combined for convenience:

Figure 6:  Hourly wind generation, nine countries combined, 2013

Figure 7 is a blowup of the period between February 2 and 15, which covers low wind period 2. According to these results the wind died to a whisper all over Western Europe in the early hours of February 8th:

Figure 7: Wind generation, nine countries combined, February 2013

These results are, however, potentially misleading because of the large differences in output between the different countries. The wind could have been blowing in Finland and the Czech Republic but we wouldn’t see it in Figure 7 because the output from these countries is still swamped by the larger producers. To level the playing field I normalized the data by setting maximum 2013 wind generation to 100% and the minimum to 0% in each country, so that Germany, for example, scores 100% with 26,000MW output and 50% with 13,000MW while Finland scores 100% with only 222MW and 50% with only 111MW. Expressing generation as a percentage of maximum output gives us a reasonably good proxy for wind speed.

Replotting Figure 7 using these percentages yields the results shown in Figure 8 (the maximum theoretical output for the nine countries combined is 900%, incidentally). We find that the wind was in fact still blowing in Ireland during the low-wind period on February 8th, but usually at less than 50% of maximum.

Figure 8:  Percent of maximum wind generation, February 2013

But even Ireland was not blessed with much in the way of wind at the time of minimum output, which occurred at 5 am. Figure 10 plots the percentage-of-maximum values for the individual countries at 5 am on the map of Europe. If we assume that less than 5% signifies “no wind” there was at this time no wind over an area up to 1,000 km wide extending from Gibraltar at least to the northern tip of Denmark and probably as far north as the White Sea:

Figure 9:  Map of percent of maximum wind generation, February 2013

During this period the wind was clearly not blowing “somewhere else”, and there are other periods like it.

Combining wind generation from the nine countries has also not smoothed out the spikes. The final product looks just as spiky as the data from Spain we began with; the spikes have just shifted position:

Figure 10: Spain wind generation vs. combined generation in all nine countries, 2013 (scales adjusted for visual similarity)

Obviously combining wind generation in Western Europe is not going to provide the “reliable electricity” its backers claim it will. Integrating European wind into a European grid will in fact pose just as many problems as integrating UK wind into the UK grid or Scottish wind into the Scottish grid, but on a larger scale. We will take a brief look at this issue before concluding.

Integrating the combined wind output from the nine countries into a European grid  would not have posed any insurmountable difficulties in 2013 because wind was still a minor player, supplying only 8.8% of demand:

Figure 11: Wind generation vs. demand, nine countries combined

But integration becomes progressively more problematic at higher levels of wind penetration. I simulated higher levels by factoring up 2013 wind generation with the results shown on Figure 12, which plots the percentage of demand supplied by wind in the nine countries in each hourly period. Twenty percent wind penetration looks as if it might be achievable; forty percent doesn’t.

Figure 12:  Percent of hourly demand supplied by wind at different levels of wind penetration using 2013 data

Finally, many thanks to Hubert Flocard, who recently performed a parallel study and graciously gave Energy Matters permission to re-invent the wheel, plus a hat tip to Hugh Sharman for bringing Hubert’s work to our attention.

This entry was posted in Energy, Political commentary and tagged , , , , , , , . Bookmark the permalink.

86 Responses to Wind Blowing Nowhere

  1. shaszeldine says:

    Very nice compilation indeed.
    Its hard to avoid the conclusion that 100% storage, or generation backup is necessary.
    I archived European weather maps from 1st July 2009, where I think exactly the same blocking high happened. Less wind generation then, similar weather. Lasted several days.

    • Willem Post says:


      Your set of graphs should be required reading for all Kool-Aid drinking wind energy promoters.

      Sharman, Bach, Flocard, etc., are the European pioneers who have put together such graphs during the past 5-10 years.

      I seriously doubt the European grid could handle 20% wind, even after major changes to the European grid, which would require many billions of euros and decades to implement.

      The upshot: After having littered the landscape with tens of thousands of those 400 – 500 ft tall wind turbines, AND made the grid changes for grid adequacy, AND made the generating capacity changes for capacity adequacy, to then find there are significant periods with no, or minimal wind, would mean almost ALL other generators would need to be kept in good repair ready to run, staffed and fueled to provide the required energy to service the demand. Yikes!

      I recommend much increased energy efficiency, starting with energy surplus buildings that also charge vehicles. It is nearly invisible and much better for the environment.

    • Leo Smith says:

      Its hard to avoid the conclusion that 100% nuclear or fossil and no wind or solar is necessary 😉

      • davekimble3 says:

        You also have to match the supply to the demand in real time – load-matching. Nuclear is the WORST technology for its ability to load-match, then coal, then gas, with hydro being best.

        When talking hydro, you have to look at both the power capacity (in GW) and the energy capacity (in GW.seconds). The energy capacity (volume of water in the dam) also has a replenishment time. When armed with that dtata, then you can work out how much wind+hydro can reliably produce, with the dips in wind filled in with hydro.

        Optimising further, rather than running only hydro for load-matching, there is some break-even point where you would also “fire up” gas to assist in load-matching – perhaps when the meteorologists recognise low wind conditions developing. So a proportion of gas-fired generation needs to be kept on “hot standby”, and a proportion on “cool standby”.

        A simplistic single sentenence is just not adequate to address this complex issue.

        • clivebest says:

          Having more supply than demand from nuclear is a luxury rather than a problem. Having excess power would provide new services such as lighting motorways at night, charging electric cars reliably overnight and providing local hot water heating.

        • Willem Post says:


          Norway is 95% hydro. Has been load matching with hydro for a century, plus, the last three decades, has helped Denmark to do its load matching.

        • Louis says:

          When armed with that dtata, then you can work out how much wind+hydro can reliably produce, with the dips in wind filled in with hydro.

          … and there can be volatility in the power available from Hydro that lasts for a lot longer than the daily fluctuations in wind and solar which I would think means that if you have Hydro in the mix to balance daily/hourly fluctuations in wind and solar you need something in the mix to balance fluctuation in Hydro that can span years rather than days.
          Venezuela being a case in point where rolling blackouts due to water shortages undermining Hydro generation pushed forward the idea of wind and nuclear as backup which then set off the arguments about who is allowed to develop nuclear and who is not.

          This link from 2010 sums it up, sadly involving the word ‘green’ … never mind

          … and this link from 2014 suggests that drought, like the wind not blowing is difficult to predict and rectify.

          • Louis: So many things get blamed on the climate these days that I automatically get suspicious when someone blames something else on it – particularly when it involves droughts.

            I just checked some of the GHCN rainfall records for Venezuela. 2009 was a dry year in some places but not in others, and since 2009 rainfall has been about average. So the current water shortage in Venezuela isn’t caused by drought. It’s almost certainly a result of the water system slowly falling apart under the Chavez/Maduro regime, like everything else in Venezuela.

            But you’re right about the hydro

          • davekimble3 says:

            > slowly falling apart under the Chavez/Maduro regime, like everything else in Venezuela.

            Not to mention the financial/trade blockade and demonisation/ridicule of leaders used by the US against regimes it intends to change, like North Korea, Cuba, Zimbabwe, Argentina.

          • Louis says:

            So many things get blamed on the climate these days that I automatically get suspicious when someone blames something else on it – particularly when it involves droughts.

            I wasn’t suggesting the blame lay with climate change, and I don’t think climate change is mentioned in either of the articles I linked to.
            I guess my point is that the bulk of populist renewable power requires the power generator .. wind, sun, wave, water to be present at the generating infrastructure. Once the capture facility is built it generally can’t be moved and if there is a change in the input .. no wind, no water, no sun, no waves then the power generation falls/fails.
            As there are approximately 196 countries (give or take Taiwan) and only 31 are currently ‘allowed’ nuclear it seems that fossil fuel power generation, being the only large scale option where the input fuel is acceptably transporteable to the generating plant, remains the only viable backup for about 74% of countries. (give or take Taiwan).

      • davekimble3 says:

        You also have to match the supply to the demand in real time – load-matching. Nuclear is the WORST technology for its ability to load-match, then coal, then gas, with hydro being best.

        When talking hydro, you have to look at both the power capacity (in GW) and the energy capacity (in GW.seconds). The energy capacity (volume of water in the dam) also has a replenishment time. When armed with that dtata, then you can work out how much wind+hydro can reliably produce, with the dips in wind filled in with hydro.

        Optimising further, rather than running only hydro for load-matching, there is some break-even point where you would also use gas to assist in load-matching – perhaps when the meteorologists recognise low wind conditions imminent. You need a proportion of gas-fired on “hot standby” and a proportion on “cool standby”, which is expensive idle capacity most of the time.

        A single glib sentence is clearly not sufficient for this complex problem.

      • roberto says:

        Well… I agree with you…but seems that nuclear is gonna have a harder time!… look at this!…

        … crazy, uh? What do you think?

        How about the UK suing Germany or Italy for having promised 50 Eurocents/kWh to those who installed PV back in 2009?

        Then some people dont’ understand why there are more and more people who don’t trust this EU BS????


  2. John Reid says:

    What are your views on the recently announced Sadoway liquid metal battery for grid-energy storage? Such a gadget could go some of the way to smoothing out the spikes but my guess is that at US$500 per installed kilowatt-hour of storage it would not be economically viable. How does that figure compare with, say, hydro storage?

    • Leo Smith says:

      economics 101.

      Wind more expensive than nuclear.
      Solar much more expensive than nuclear.
      Wind & solar much less reliable than nuclear.

      Therefore wind or solar plus storage plus fat wires to take peak flows much much much more expensive than nuclear.

      Conclusion: adding complexity and cost to an already complex and costly and totally unnecessary technology wont make it anything but even more costly and useless.

      ‘Throwing good money after bad’ is a well known way of expressing this.

      • Willem Post says:

        Thank you for clearly debunking fantasy thinking.

      • John Reid says:

        The issue of storage has always been the Achilles heel of alternatives. Sadoway was on the (Australian) ABC News the other night spruiking his invention. He was not asked this key question and we were left with the impression that this problem has been solved. I don’t believe it has been solved.

      • scopri says:

        If nuclear is so cheap, why does Hinkley needs a subsidy? Maybe it’s because the construction alone will probably cost more than £24bn.
        Maybe there is another reason, who knows?
        But if something needs a subsidy, it’s probably not so cheap, as people may think.
        (Or to brag about how cheap it is).

        • Louis says:

          If the French and Chinese are putting up the money for construction as private investors isn’t it more like a loan with interest due on top of repayment rather than a subsidy.

          Austria to bring lawsuit against Hinkley .

          Austria’s complaint hinges on giving subsidy to nuclear power: “Hinkley Point is a bad precedent, because guaranteed feed-in tariffs were previously reserved for renewable forms of energy.”

        • A C Osborn says:

          It is simple they demanded the same subsidies as Wind so that they are entering a level playing field.
          Subsidies are now being offered to Gas generators to keep their stations open for the same reason.
          If you totally skew the market those being totally put upon say Sod this and leave the market, something the UK goverment can’t afford to allow.

  3. Graeme No.3 says:

    Confirmation, if needed, from

    P. Miskelly (Australia)
    Reprinted from ENERGY & ENVIRONMENT VOLUME 23 No. 8 2012

    In an examination of the 5-minute time-averaged wind farm operational data for 21 large wind farms connected to the eastern Australian grid- geographically the largest, most widely dispersed, single interconnected grid in the world (AER, [1])- this paper challenges…. the opinion that “the wind is always blowing somewhere”.

    See also along these lines
    talks of South Australia home of the most wind capacity in Australia (also the home of Australia’s most stupid politicians and worst drivers, but I digress).

    “on July 21 registered wind generation in the National Electricity Market fell to a low of 23 megawatts. The minimum average for any two-hour period was 24MW. That is the power production figure for all wind turbines in South Australia, Victoria, NSW and Tasmania, which have a combined capacity of 3300MW.
    The crash in supply from wind was repeated the following day”.

    • Willem Post says:


      Miskelly is the Australian who blew the lid of the blowing somewhere fantasy regarding Australia’s south-eastern grid.

      All we need to do now is the get the US DOE and its minions to finally fess up to the obvious.

    • TonyfromOz says:

      I hope you don’t mind if I mention this here, and it’s probably been mentioned before, but let’s look at what happens when the wind dies off to virtually nothing, and here I’m using a specific instance in South Australia, and for some perspective it entails a bit of work on the part of readers. There were a number of factors at play here, all resulting in what have been the near perfect storm, blackouts, and you wouldn’t have read this anywhere. They wouldn’t dare mention it.

      Here’s the first link

      Now this is for Australian wind plant performance, the older version of the Miskelly daily charts, and this is for 3rd June 2013.

      Now, under the first two images there are listed all the operational wind plants at that time. So, untick all the boxes at left and the box at right which says ‘All’. So now, all you have showing are those South Australian wind plants and the thick black line on the lower chart is the total power output for all that wind power.

      Now, scroll down a little to the third chart there, total power consumption, not generation, but what is actually being consumed, here, a typical Winter Load Curve with the two peaks in the AM and PM measured in hours across the day from Midnight and back to Midnight.

      Note when everyone gets out of bed for the morning, say around 5.30AM, and take an average of power consumption across the day, around 1650MW average.

      Now, scroll back up and look at the output from wind for that same time, and this is for a Nameplate Capacity of 1200MW, and the average for that same period of time is around 35MW, which is only 2.1% of actual consumption.

      Okay, now here’s the second link and this is to the power costs for Australia for the Month of June 2013.

      Note the wording above the cost chart there where it says that the Peak Power period is from 7AM to 10PM, the period of virtual zero generation from Wind.

      This is the wholesale cost of electricity, in other words, what the retailers buy their power for from the generation entities and then onsell it to consumers at their retail price. Note the cost of power for SA for that peak period for that day 03Jun2013, and see that it’s $866.41 per MWH, and extrapolated down, that’s virtually three times what those retailers can sell it for at retail.

      I know that here I have picked just this one huge day, but look down the list there and there are a number of days where power costs well up over $100 per MWH, and the same applied for the two previous Months as well.

      So then, what happened here?

      The Government decided to close the only coal fired power plant in operation in the State at the End of Summer, and that plant then initiated a huge maintenance task and some major work to ensure the extended operation of the plant.

      So, now the State of SA had to rely on the wind power, safe in their perceived knowledge that with so much wind power, then they would be able to do just that.

      This is also at a time when the Federal Government had in place its Tax on CO2 emissions.

      So with no wind, and no coal fired power, then all those tiny little plants which only operate on a short time basis, perhaps a couple of hours a day now had to run for not just that, but for almost continuous times over days, and weeks. They had calculated their CO2 tax on those couple of hours a day, and had now completely and utterly blown their budget, and were now in huge deficit for the year, incurring staggering costs with overrun of their CO2 emissions, along with sourcing all that extra supply of Natural Gas which they had to buy at a premium, not just on a short basis, but now for almost 3 Months.

      So, quietly, right at the end of that Month, that coal fired power plant finally ran back up, and began delivering its 24/7/365 power, and power costs almost returned to what they were before.

      So, there consequences when the wind stops blowing.

      I can’t say if those retailers were forced to absorb those huge costs for almost three Months, and no one will ever say, but you can bet politicians were on the phone telling those plants to keep generating power no matter what the cost.

      Wind power is expensive when it actually is delivering power and even more expensive when it isn’t delivering power, a cost directly attributable to wind power.

      I apologise for leaving such a long comment here, but this is exactly the same situation.


      • William says:

        Tony, I’m surprised at what you said here: “$866.41 per MWH, … that’s virtually three times what those retailers can sell it for at retail.” So dividing $866 by virtually 3 gives about $300.

        From your data, average wholesale electricity cost is about $100/MWH. And that is being sold retail for $300. Nice markup! If I were a SA resident, that is what I’d be upset about, not worrying about the poor retailers. I’d have solar panels on my roof faster than you can say “rip-off”.

  4. Euan Mearns says:

    Roger, amazing set of charts. Many years ago now I made a compilation of Ireland, Germany and Denmark. It took only a few minutes to work out that the argument about the “wind always blowing somewhere” was false. Of course it is true that the wind is always blowing somewhere, but for practical and economic purposes that is irrelevant. The renewables enthusiasts seem incapable of understanding simple facts and seem prepared to perpetuate miss-information for either financial gain or ideological advancement.

    Interesting exchange with Hubert via email where we both agreed that Grid Operators will promote renewables since this advances their own business plan. The National Grid is very much in favour of building power lines and inter-connectors everywhere. Its a good example of where capitalism is totally broken. Its like a doctor going out and breaking legs so that he can get some extra business.

  5. Sam Taylor says:


    One of the assumptions behing arguing for interconnectors all over the place seems be that since the wind over large geograpic regions is uncorrelated (I would assume the correlation coefficient between say spain and finland is low) then this will smooth out the regional variations. Interesting to see that this is not the case.

    The red line in figure 12 is the most worring for me. Assuming we keep building wind farms, how the hell do we balance something like that.

    Would you happen to have a link to the Flocard work?

    • Leo Smith says:

      The real case for inter-connectors is to access useful hydro power from further way. And also good nuclear base load.

      A rational pan European grid would be overwhelmingly nuclear and hydro, with nuclear providing the baseload grunt and hydro doing the demand following.

      With interconnectors supplying the transfer capabilities.

      In the end the case for an inter-connector is simple: that it allows you to reach an instantaneously cheaper source of electricity at less cost than building a power station locally.

      • Willem Post says:


        Some of France’s nuclear reactors have DESIGNED-IN some load-following capability.

        • Roberto says:

          For the record… ALL of the presently operating PWR reactors in France, plus the EPR under construction in flamanville, have load-following capabilities.
          EDF normally uses only few of them depending on the status of the fuel, and the need of the moment (day, season).

        • Leo Smith says:

          Where did I say they didn’t?

    • Sam: This graphic shows the correlation coefficient between individual pairs of countries (total 36) plotted against distance between the countries measured from midpoint to midpoint:

      Wind generation in Western Europe becomes substantially uncorrelated (R<0.3, R^2<0.1) beyond about 1,000km and becomes asymptotic to zero somewhere around ~4,000km, although there aren't enough points to say exactly where (the rightmost point is Spain and Finland).

      Wind power is stochastic - i.e. effectively noise - and when you add noise to noise what you get is more noise.

      • Willem Post says:

        You are very clever. My hat is off. What further proof, entirely based in emperical data, is needed?

      • Roberto says:

        Roger, on your last sentence… ON PRINCIPLE adding noise too noise increasing the number of noise-creating sources REDUCES the overall noise amplitude… that’s a direct consequence of the structure of the noise of windmills’ power generation.
        In PRACTICE, it’s another story, though… as you’ve rightfully noticed.


        • Roberto:

          On principle you’re right.

          In practice you’re right too 🙂

        • I think the reason theory doesn’t work in practice here is that wind spikes/troughs (noise) are quite highly correlated between adjacent countries. There probably would be a reduction in the amplitude of the spikes/troughs if you connected Spain with Finland and Ireland with the Czech Republic and missed out the countries in between, but that isn’t a very practical proposition.

      • Leo Smith says:

        wind power is not geographically stochastic.

        • roberto says:

          Agreed!… but if you look at some early works on this subject, like the ones by Mark Jacobson and his never ending platoon of students.

          Another more recent example:

          “The stochastic nature of wind and solar complicates the
          treatment of system reliability in grid integration studies.”

          …. in “A Monte Carlo approach to generator portfolio planning and carbon emissionsassessments of systems with large penetrations of variable renewables”, Elaine K. Hart, Mark Z. Jacobson, Renewable Energy 36 (2011) 2278e2286.

          This is a more recent one…

          “Correlation and statistical characteristics of aggregate wind power in large transcontinental systems”, Henry Louie
          Wind Energy (Impact Factor: 1.44). 06/2014; 17(6). DOI: 10.1002/we.1597

          “ABSTRACT Studies have shown that the unpredictability and variability of wind power is reduced in systems with large numbers of geographically diverse wind plants. These effects are caused by the decreased correlation of power output between wind plants as their separation and diversity in terrain increases. One way that system operators have increased geographic diversity is by enlarging balancing areas through the physical or administrative connection of adjacent systems. This strategy can be extended from the regional level to the transcontinental level. As such, it is important to study the correlation and statistical characteristics of aggregate wind power between large, distant systems. This paper analyzes multi-year historical data from four North American system operators—Bonneville Power Administration, Electric Reliability Council of Texas, Midwest Independent Transmission System Operator and PJM—to see how effective transcontinental interconnection of systems is at enabling wind plant integration. The effects of separation and timescale on correlations of instantaneous and hourly variations are analyzed. The analysis is complemented by a study of a hypothetical transcontinental connection of the systems across yearly, monthly, daily and hourly timescales. The results show that correlations between large systems exhibit similar characteristics as the correlations between individual wind plants, but are somewhat larger in magnitude. The transcontinental system exhibits a close to normal distribution of power output and decreased variability, but there is still appreciable and statistically significant correlation at the longer timescales driven by seasonal and diurnal forcing, as well as synoptic weather systems.”

          –> HYPOTHETICAL TRANSCONTINENTAL CONNECTIONS! <– … the sky is NOT the limit for green guys!… unbelievable!

          You can see that they've assumed just that, and then all the rest follows… and EWEA and the like have all the interest to keep on spreading the myth that "wind always blows somewhere"… which is simply what comes out of the south end of a north facing bull, as we all know here.


      • Andrew says:

        Indeed, the graphic shows trivial correlation between sites >1000km apart. The same pattern was shown for UK by Sinden []

        With respect, these observations are at odds with your argument that there is too high a correlation between countries’ wind speeds for wind power to be ‘reliable’ in a system inter-connected over large scale.

        The graphs in your post show that times can be found when the wind is blowing nowhere, and other times when it is blowing everywhere. But these are coincidences which will always occur in any poorly-correlated paired data. Stepping back from specific occurrences, the evidence in the graphic above and in Sinden [it’s his Fig 5] shows that low correlation over sufficient spatial scale allows some dampening of deviations in supply.

        Actually, it’s not low correlation we want for wind to have any use as a reliable energy source over large scales: it’s strong negative correlation, which is apparently not on offer

        Sinden is widely cited, eg: and

    • Willem Post says:


      Regarding balancing, see my above comment. The balancing part is only one part of a three part problem.

  6. dilaseuq says:

    If the graphs at different scales don’t smooth out it implies that the process has a fractal dimension (same shape at different scales)

  7. roberto says:

    On the same subject:

    “The variability of interconnected wind plants”, Energy Policy, Volume 38, Issue 8, August 2010, Pages 4400-4410, Warren Katzenstein, Emily Fertig, Jay Apt


    We present the first frequency-dependent analyses of the geographic smoothing of wind power’s variability, analyzing the interconnected measured output of 20 wind plants in Texas. Reductions in variability occur at frequencies corresponding to times shorter than ∼24 h and are quantified by measuring the departure from a Kolmogorov spectrum. At a frequency of 2.8×10−4 Hz (corresponding to 1 h), an 87% reduction of the variability of a single wind plant is obtained by interconnecting 4 wind plants.
    Interconnecting the remaining 16 wind plants produces only an additional 8% reduction.
    We use step change analyses and correlation coefficients to compare our results with previous studies, finding that wind power ramps up faster than it ramps down for each of the step change intervals analyzed and that correlation between the power output of wind plants 200 km away is half that of co-located wind plants.
    To examine variability at very low frequencies, we estimate yearly wind energy production in the Great Plains region of the United States from automated wind observations at airports covering 36 years.
    –> The estimated wind power has significant inter-annual variability and the severity of wind drought years is estimated to be about half that observed nationally for hydroelectric power. <–"

    • Roberto: I didn’t look into frequency spectra but I did do some work with coefficients of variation and ramp rates. The coefficient of variation decreased from 0.54 to 0.47 after I added the other eight countries to Spain – a minor improvement. The maximum ramp rate, however, increased from 53 to 64 MW/minute.

      I also ran another case that distributed the wind generation between the nine countries in proportion to their land areas rather than concentrating it in a few of them. It didn’t change the numbers significantly.

  8. roberto says:

    …and this one is for UK wind lovers… interesting conclusion…

    “The progressive inefficiency of replacing renewable obligation certificates with contracts-for-differences in the UK electricity market”, Energy Policy, In Press, Corrected Proof, Available online 14 January 2015
    Derek Bunn, Tim Yusupov


    This paper looks at the emerging risk/return profile for new renewable assets as a conventional wholesale electricity market progressively decarbonises. Using a detailed fundamental model of price formation risks, under increasing replacement of fossil fuel facilities with onshore and offshore wind, we show that the risk return profile becomes less attractive over time, and may therefore need sustained and possibly increasing policy support. Furthermore, we show that green certificate trading may become progressively more attractive as a supplementary support to wholesale prices, compared to fixed feed-in-tariffs. This is because the increasingly negative correlation between renewable output and wholesale prices reduces its revenue risk compared to fixed feed-in tariffs, if other factors remain constant, and thereby improves conventional financial performance risk metrics.
    —> In particular, this suggests that the recent energy policy change in Britain to move away from green certificates and into contracts-for-differences may have been ill-founded. <–"

  9. clivebest says:

    The IEA defines energy security as:“Energy security refers to the uninterrupted availability of energy sources at an affordable price.”

    Moving from fossil fuels to wind power threatens energy security because it increases the likelihood of a shortfall in supply. One example of this was December 11th 2012 at 5pm when with peak UK demand was 56 GW wind output was essentially zero (<0.1GW). Since then wind capacity has increased by 4 GW but what extra energy security has that brought us ? Twice this week we have again seen essentially zero wind output (0.4GW) at peak load providing less than 1% of UK power needs.

    This post provides good evidence that extending the analysis to include other European countries does not help either. Germany was producing very little wind power this week either. I once did an exercise to scale up UK wind power by a factor of 10. So instead of investing £100 billion in wind energy we invested £1 trillion ! This is what the grid would look like

    This madness must stop.

    • roberto says:

      Interesting graph, Clive… but how could 40,000 turbines cost 1 trillion pounds?… isn’t it approximately a factor of 10 too high?


      • clivebest says:

        Ed Davey’s energy statement 2014:

        Record investments of £45 billion in electricity generation and networks since 2010 have put us on target to meet our future low carbon power requirements.

        Indeed, in four years under this Coalition, the available evidence suggests we have surpassed the total electricity investment in the whole of the previous decade under the last Government.

        So in total 45*2 = £90 billion by Nov 2014 So it will soon be £100 billion, the bulk of which is on wind farms, their interconnect and their subsidies.

        So let’s be conservative than and say £500 billion needed to increase exisiting wind capacity by a factor of 10.

        Sizewell B cost just £2 billion and regularly outperforms all wind farms combined.

      • Leo Smith says:

        Turbine Capex ~ £1m/MW onshore, £3m/MW offshore.

        so 100GW capacity (25GW average flow) is probably around £200bn.

        Plus cost of balancing etc,

        £200bn will buy you 40GW of longer lifed, more reliable nuclear power.

    • Todd says:

      At least some of this kind of information is gradually making its way into mainstream media “Badly located renewable power plants cost Europe $100 billion: Davos report”

  10. matthew_ says:

    Thank you for the graphs.
    While figure 10 may appear equally spiky, I believe you would see some changes if you plot both as a power probablity distribution. Probably changes along the lines of what is shown here URL: (Archived by WebCite® at Could you make those plots?
    The graphs show that this does not remove all periods of very low production.
    In my mind this does not mean that wind should not be a part of a risk balanced electric generation portfolio because it still gives a fixed cost, home grown power source.
    The risk of zero wind is real and quantifiable. The risk of not getting adequate gas at an affordable price from another country is real but not so quantifiable.

    • Willem Post says:


      “The risk of not getting adequate gas at an affordable price from another country is real but not so quantifiable.”

      This is not valid.

      Russia is capable of supplying Europe with gas at reasonable, competitive prices for many more decades in the future. If not Europe, then East Asia. It is Europe’s choice.

      Russia has PROVEN it for decades in the past, despite being hindered by dysfunctional Ukraine.

    • Roger Andrews says:

      Matthew: You can add as much wind power as you like but you’re still going to need gas turbines to fill demand when the wind doesn’t blow. So adding wind doesn’t reduce your dependence on secure gas supplies.

      I can’t make power spectra plots, or at least not easily. I don’t think they would tell the readership much anyway. Actual generation data are much more instructive. Also see my response to Roberto above.

      • matthew_ says:

        Adding wind reduces the volumes involved in the gas dependence equation. Lower volumes means that it’s easier to build sufficient storage, have alternative suppliers, etc.
        The number of gas turbines needed will increase somewhat (if gas is used for balancing), which is also stated by EWEA.
        Integration costs and additional infrastructure needed look reasonable up to about 20-30% of annual energy being supplied by wind. Above that we should take a new look at the integration costs.
        We certainly should not be dismantling reserve generation. Capacity markets need to be implemented so that we keep enough gas turbines available for the days when we need them.

  11. Owen says:

    I done a similar study here but over a smaller period for the UK and Irish systems :

  12. James Philip says:

    Wind would work fine in one application and that is industry which can be rapidly started and stopped.
    It is clearly incompatible with always on industry or governmental services or domestic use and any attempt to shoehorn it into such uses is bound to fail.
    That doesn’t mean that you shouldn’t persue wind but rather that it should be pursued with an industrial policy that can make use of the wind generated. And the acceptance that very little of the wind electricity can be used in homes, offices or services.

    • Leo Smith says:

      Green advocates always use the word ‘can’

      The real world is actually about the phrase ‘cost effective to’

      Wind plant is a large capex that delivers 25% of its peak capability on average, and so too will anything coupled to it.

      Those are bad numbers for utilisation of any big capital plant.

      Once you have built an aluminium smelting plant, you want to run it 24×7, not stop and start it – which adds wear and tear for no profit as well.

      I – and it seems increasingly the public – are sick and tired of renewable apologists who claim that throwing good money after bad will solve the problems of renewable energy,.

      Renewable energy so called of the intermittent kind is a total financial and economic disaster, and it contributes almost nothing to carbon reduction.

      It can never be improved to eliminate the fundamental flaws in it, because they are in fact fundamental to it.

      Wind is diffuse, has low specific energy content, and is liable to large geographical low values.

      Which means to extract energy you need a lot of windmills over a very large area and you STILL may get periods of extremely low output when a continental sized anti-cyclones sits down on them. And your energy on average will be lower – far lower =- than the peak value, wasting expensive infrastructure.

      No amount of extra technology can change that. Wind is useless enough before you add storage or ICTs, and it simply gets worse with them added,

      • davekimble3 says:

        The point was that the industries hanging off the wind farm should be of the sort that can be turned on and off easily. Aluminium smelters are the exact opposite of this, so don’t prove anything. Water purification by reverse osmosis would be well-suited. We are only talking about having to shut down quickly for perhaps 1 or 2% of the time.

        • Leo Smith says:

          Even the draining of the Anglian fens – one of the least demanding exercises, time wise – had replaced every windmill within decades of the first steam pumps being installed.

          The problem is pure economics in the end. The only thing that can be relied upon to not be needed during times of high wind electricity, is electricity, so apart from pumped storage you can never guarantee a need will exist that surplus electricity will be handy for.

          And the cost of the plant so involved will always be higher than for 24×7 plant.

          People seem to fail to grasp just how deep the intermittency dagger goes into the heart of renewable energy. Every single thing subjected to intermittent flow, must needs be oversized by the inverse of the capacity factor.

          The grid as a transmissions system runs at about 60% capacity factor. It is oversized already to cope with winter peaks.

          Stick solar power on any part and the solar links need to be 10:1 oversized, because the average contribution of solar is 10% of its peak capability.

          Wind is about 25%.

          All this intermittency leads to extra capital expense and environmental impact.

          On anything that has to deal with it.

          It doesn’t matter whether its a link to somewhere else, a factory that sometimes operates, or a pumped storage unit or other store, its bigger than it needs to be because intermittent power needs to be utilised at peaks and isn’t there the rest of the time.

          The renewable lobby has hand waved intermittency away by mumbling ‘easily fixed with storage’ (or some other random technology).

          It isn’t easily fixed, and indeed it cannot be fixed at all, it can only be offset by spending extremely large capital sums – of the same order as the generating kit that is deployed – and then only partially.,

          Up till now the cost of so doing has been borne by the counter-generators – the gas power stations in the main. And they are now closing as uneconomic.

          Leaving us dangerously exposed.

          Meanwhile the renewable lobby continues to look at the costs of renewable energy solely in terms of nameplate capacity, not actual average delivery, and totally ignore the costs of providing dispatch, as their members don’t have to do it.

          If they were a commercial company, Renewable UK would be in court for fraud and misrepresentation to their shareholders. As a political lobby merely funded by Big Green they can lie to their bank balances’ content, and nothing can be done.

        • Louis says:

          Water purification by reverse osmosis would be well-suited.
          You can’t let the membrane dry out in an RO system so a critical civic system like RO desalination would not be suitable for the fluctuations of wind power.

      • James Philip says:

        Firstly I don’t personally think there is too much that can be done about CO2 emissions and so that is not part of my equation.
        The kind of industries I had in mind were a machines for making nails, screws, etc. Simple generic metal bashing, woodworking, food milling. However that does not change the reality that if you want to power such industries with wind you will need 4 times the machinery to do it.
        The difference in price between always on electricity and intermittent electricity must justify the cost of acquiring the additional machines.
        If it does go with wind if it doesn’t don’t.

    • Nador says:

      “Wind would work fine in one application and that is industry which can be rapidly started and stopped.”
      But the question is whether there are many industries like that. There probably are not. This hypothetical industry needs to be energy intensive, should have relatively low capital and human employment costs and also be able to ramp up/down fast. I do not think there is particularly many of them. Should there be such an industry it would already run at night when electricity is cheap. As night time spot prices are generally not more than half of the daytime ones (I do not know the exact value), it is safe to assume that any additional industry that could use intermittent energy requires even cheaper electricity than half the normal price [probably much cheaper]. So unless the intermittent energy costs a fraction of the dispatchable sources, it is not worth having it.

      • James Philip says:

        I think the problem lies with concept the normal price. Make everyone pay the spot price with prepay. Then you send a price signal that it maybe worth switching off your machines at certain times or that its worth paying the extra for nuclear. Different consumers will come to different conclusions, but the effect should be the same toward stability.

      • James Philip says:

        Just correcting two errors in above post.
        ‘the concept’, not ‘concept’ in first sentence.
        And ‘toward stability in price’ instead of ‘toward stability’ in last.

  13. Kaj Luukko says:

    Thank you for an interesting post!

    A few years ago I made a similar estimation by using wind data from weather stations around Europe. I took data for the whole year 2009. I plotted the mean value of wind from all station as m/s (yellow). According to the wind data I calculated how much power a grid would have generated as a whole in those wind conditions (red). The profile looks like this.

    The result is very much alike yours. A google-translation from Finnish to English is not very good, but the idea may be understandable.


  14. Ted Trainer says:

    I don’t think the comments or the original article point out that even if the wind is “always blowing somewhere” at a point in time, for wind to be able to meet a large fraction of demand at that time you would have to have enough turbines located at that somewhere … and tomorrow when the wind is blowing somewhere else you would have to have the same amount located at that somewhere else … which leads to the impossible implication of needing enough turbines at each and all of the locations where the wind might be blowing at a point in time.

    The other major point is that Europe in particular gets anti-cyclonic effects, whereby the whole continent can be pretty cloudy, calm and cold for days at a time.

    • Leo Smith says:

      My calculation was that for the UK in isolation, the number of turbines needed would result in an electricity price of £9.99p a unit.

      I simply took Gridwatch’s lowest recorded wind output and divided it into Gridwatch’s highest ever demand and multiplied the resultant capacity by a million a megawatt, and wrote that down over 20 years of electricity.

      One of the prime motivations for building gridwatch was to get the data to assess the statement ‘the wind is always blowing somewhere’ and its relevance and validity.

  15. Jacob says:

    There are here daily data about Spanish generation and production.
    To sum it up: They have some 23 GW of wind capacity installed.
    Rarely do they get 10-12GW – 40-50% of installed capacity.
    Some of the time they get 6-8 GW production – 25-30% of capacity.
    Quite often they get only 3-4 GW – 12-16%
    And sometimes less than 1 GW – less than 4%. And all values in between.

    Also remarkable are the hourly jumps – on some days you get most of the day 6-8 GW, but on some 2-3 hours there is a “hole” of only 1GW.

    They seem to do a lot of balancing with hydro, which makes sense, since water availability is limited, so wind saves water. They also use gas for balancing.
    Nuclear and coal are more steady – at 25-30% of production each.
    Wind renders between 1-2% and 20-25% of production, depending on wind availability.

    • Jacob, wind in Spain has represented in 2014 exactly 21.3% of the total national electric demand (258.5 TWh), but Red Electrica Española (the ree you have quoted) has been able to manage peaks of up to 60% of the national demand.

      Considering Spain is almost an electric island (with probably less connection links than even UK with the continent) due to the opposition of France and its peculiar energy policy, Spain can only exchange very little electricity with Portugal, Morocco and France. The Spanish networrk is helping, as much as interconnections permit, the Portuguese smaller network to buffer/backup its considerable penetration of wind power also.
      But this, on the other hand, proves that much more than 20% of wind energy could fill the European demand, if so decided, without big stability issues.

      Another issue is that best wind fields are already taken and now promoters struggle to get onshore fields with more than 2,000 hours/year, unles they go to offshore (also limited), which is another order of magnitude in costs and long term maintenace, due to heavy corrosion in a salty environment.

      And yes, you are right, hydro, in this mountanous country is an important buffer to back up wind and solar. Hydro represented in 2014 a 18.4% of the total yearly national demand. Also combined cycle gas fired plants are helping a lot, because sometimes this country badly needs the scarce water, specially in half of the country Southwards.

      That is what is making the gas fired plants a total disaster. They went form representing a 25% of the total national electric demand in 2010 to 10% of the total national demand in 2014. This means that gas fired plants 5 to 10 years old, originally designed to work 5,500 hours /year minimum, are now working 870 hours a year. Amortization is horrible and stress higher than originally designed, because the high number of prewarmings and post-cooiings required, as in Spain, renewables always enter first into the network by law (at least, until very recently)

      • Hubert Flocard says:

        I have made an analysis of the correlation in Spain between gas plant production and wind production. The anicorrelation is almost perfect meaning that in Spain gas plants are the crutches of wind energy.
        You are right that engineers can do everything even balancing wind peaks as happen in Spain (which only recently moved its connections to France – and North Europe – from 1 to 2 GW. The question are : at what cost ? Is it beneficial to the consumer ? Is there a chance that producers with dispatchable plants will accept working as crutches to subsidized intermittent production ?

        • Euan Mearns says:

          Hubert, glad to have you posting here. Both Hugh and Leo post regularly. This post form November 2013.

        • The debate whether producers with dispatchable plants will accept working as crutches to intermittent production is what is at stake in Spain since three or four years, with no clear solution.

          Gas fired plant promoters built their modern plants in early this century from almost nothing (in combined cycle) to 27 GW of installed power. There were three main reasons:

          First the compliance of the Kyoto Protocol in a country that was heavily coal dependent in generation and was afraid it would be penalized, after signing a Treaty that these days it seems everybody has forgotten.

          The second reason was that Spain was growing steadily in a 3-4% these years and politicians always believe in permanent growth (not only Spanish politicians, by the way)

          The third and important reason is that Spain is connected by two gas pipelines with Algeria and has the third most important NGL fleet and 7 regasifying ports, only after Japan and South Korea and had signed, years before, long term supply agreements with the most important exporters, which worked very well until the crisis unfolded.

          (N.B. Spain handles now about 16 Bcm of gas /year by pipeline but another 15 Bcm by NGL tankers The agreements were signed mainly, after the coup d’Etat in Algeria left this country aparently quiet for Europeans, but the uncertainties raised again with the Arab Spring revolts. Curiously enough, it seems that the rest of Central and northern Europe gas supplies are now even more uncertain than those coming from Northern Africa, due to the conflict with Ukraine and the clashes between the EU and Russia and the EU is now thinking to use Spain as a hub for alternative gas to Northern Europe.

          What it was frozen by French for electricity exchanges with Spain for years, seems now to be speeded up for gas, for geopolitical reasons. The big energy problem here is that huge energy infrastructures take at least one decade to be built and need many decades of regular operation to amortize the investments and the political and geostrategic events these days are changing much faster than this. See the North Stream project, as an example)

          Now, Spain has fallen 6.1% in electric demand since 2010 and 7.7% since the beginning of the crisis in 2008. This has automatically given an artificial boost to the penetration of renewables, even there are no significant new installations of wind or solar (except of the 2 GW of CSP), since 2010, basically due to drastic cuts on the premium tariffs, even to already existing installations, thus breaching the most elementary rule of law. And as the renewables enter first, their coverage of the national demand has increased, detrimental to gas fired plants.

          In summary dispatchable plants (and specifically combined cycle gas fired plants) may perfectly work as backup for renewables if there is a national agreement considering the combined costs of creating this stability as a whole. I have participated in meetings on the subject and of course, there were no strategic agreements. Only the big electric oligopolies got some payments to the gas fired plant promoters to keep working as ‘crutches’ for renewables, but these expenses are not clearly recognized by renewable promoters as a burden they have to acknowledge as theirs, as well.

          (N.B2. Nuclear plants are not good as load-followers. French claim they can regulate them, but there are important constraints. Of course, less backup is needed, for instance between days and nights if a big portion of the home, commercial or industrial heating is also promoted to be electric and the differences between peaks and valleys are previously moderated by a subsidized coverage of the night valleys)

          Similar issue is the hydro-wind compound in El Hierro island in the Canary archipelago. They are having now 100% electricity from wind, whose surplus is stored by pumping up water to a high reservoir, to then turbine when tehre is no wind. In the days renewables had excellent preium tariffs, we were discussing for years who would have to pay the costs of implementing the complementary pump up system, not only there, but also in other bigger islands, as pump up was not among the generations or carrier systems having premium tariffs. That made the clear difference between a go/no go project.

          What we need is a more holistic view to know the complete costs incurred when renewables have to be deployed but with backup systems to provide 100% stability in a grid.

  16. Florian Schoepp says:

    Very good article and comments. Apart from all of the above, I would like to add this link:!119462/ (in German only, sorry)
    It basically says that the vast majority of German wind farms do not show any profit at all, despite the heavy subsidies. 37% of them (as of 2013) even had a negative Cash-Flow!

    So we have:
    1) high subsidies for loss-generating (no tax revenue) investments
    2) no CO2 savings, since conventional power plants have to be kept “alive”
    3) privatised profits for manufacturers and developers and socialized losses/high electricity prices for the 99%
    4) less safer electricity supply

  17. Hubert Flocard says:

    I read this from davekimble3
    “You also have to match the supply to the demand in real time – load-matching. Nuclear is the WORST technology for its ability to load-match, then coal, then gas, with hydro being best.”

    I am surprised to read this (not too much in fact) as the flexibility of nuclear plant is a known fact from engineers. I believe this wrong statement which from being repated over and over has acquired the satus of truth is easily disproved by looking at the way French plants are being used for instance to balance wing gusts. I can provide many graphs which on the contrary prove the flexibility of nuclear plant. In many ways, on this point, it can compete with coal and gas fired plants of the last generation.(although not with hydroelectricity). Of course, one has to consider the three levels of balancing (secons to 1minute, 1minute to 5, 5 to 15 and beyond). I can provide numbers on the contribution of nuclear, hydro and fuel to the balancing in France

    In case you would think that this a miracle associated with French plant ( which are just local adaptations of US PWR technology) such as the French miracle associated with Bordeaux wine I can send a text to the same effect produced by E.ON (German firm) nuclear engineers who had to answer to the same sort of unjustified critics.

    I suspect that the main reason for the wrong statement about rigidity of nuclear plants comes from a confusion between technics and economy. In a moment when balancing is needed, the last mean to call to help is the cheapest, that is nuclear most often. So most producers strives to keep their nuclear plants working at a baseload level because that is the most profitable way to handle the situation when other means are available.

    Of course when the share of nuclear energy is large (in France 75 %) nuclear can’t avoid to get involved in the balancing. The public data that can be downloaded from the RTE website (the french grid) which show how each reactor (several reactors per plant) is operated with a good time resolution (1/2h) . Spikes up and down often by several hundreds of MW are often visible. Variations at the second levels (seconds) up to 2% of the reactor nominal power are constantly used by RTE to provide the fine balancing required by the grid.

    The post rightly says that sometimes there are moments when there is wind nowhere. I would like to point oiut that until conventionnal means are destroyed, that is not the main problem. Indeed before we starting asking the gods (Eole, Helios) to produce our electricity the engineers knew how to handle the balancing associated with the fluctuations of the consumption.

    The new problem is that sometimes the gods decide to work too much and everywhere also. Then one gets to face the problem of having “a reserve of dispatchable power to shut down” (unless of course one accepts to shut down wind and solar and hopefuly not pay for the energy that has NOT be produced as is the case in some countries). There have been instance when en summer time a large number of plants has been stopped a wind gust over Germany let as only available option to have water being discharged from dams without the associated electricity production. When a dam is full (end of spring) and neverthelles gets continuously some water from a stream it has to let this additional water flow down permanently. Until these events this flow of water was used to produce electricity.In these events one had to destroy available renewable energy to balance a foreign wind production.

    Another correspondant states rightfully that the sum of even uncorrelated random variable is still a random variable. Of course, as the post and the his diagram also show, wind production is still correlated over large distances.

    For the sum of uncorrelated random variables the central limit theorem says what you should expect at best : randomness according to modified distribution of probabilities predicted by the theorem. We performed such a study by drawing numbers using distribution from small areas as a basis (typically Denmark onshore and offshore) and showed that the european probability distribution of powers is not too far from the distribution predicted by the central limit theorem assuming that over Europe there is no more than six to eight independent production zones.

    The only parts which the simplest theory of probability does not well reproduce are the tails of the distribution (very small or very large total production) which occur more frequently in reality than should appear with six or seven independent production zones. Of course this corresponds to the fact that when extreme events occur (North Atlantic storm or large cold Anticyclone) they sweep over all of Northern Europe so that the level of correlation is enhanced for the exteme events*.

    The amount of correlation between the european wind productions can only grow in the next decade. Indeed, Spain which is the only country (with Portugal) to have a really uncorrelated wind regime has stopped its wind turbine construction. The same is true for the south east of France because of the potential damage to tourism which thre has locally an economic value. The rest of France behaves like most of Northern Europe. Both in France and in Europe, this is where new wind capacity (onshore and offshore) is going to be added. Thus even the modest smoothing that comes from uncorrelated productions is going to decrease.

    * this analysis is presented in a document unfortunately written for our expected readership thus in french.

    • Hubert: Thank you for your input.

      I mentioned the fact that French nuclear plants are used for load-following in this recent post

      But as you say, a lot of people still assume that nuclear plants can only be used as baseload. It’s true that they’re more economic running in baseload mode, but if they can follow load as well they become the obvious choice for decarbonizing electricity generation.

      I think we agree that there’s no way of smoothing out wind generation by combining stochastic output from different countries no matter how far apart they are. To do this you have to link with a country where wind output is anticorrelated, but I don’t think such a country exists.

  18. Pingback: It Don’t Take Sherlock to Know; When the Wind Don’t Blow, The Power Don’t Flow – STOP THESE THINGS

  19. Pingback: The Futility and Ridiculousness of the Windscam! | the Original "Mothers Against Wind Turbines™"

Comments are closed.