RealClimate logo

Make your own forecasts of future energy, carbon emissions, and climate

Filed under: — david @ 22 June 2005

Over the years, IPCC has issued numerous scenarios describing the trajectory of civilization and what they may mean for CO2 emissions and the like. The most famous of these is the “Business-as-Usual” scenario, also called IS92A, although this has been supplanted somewhat by the SRES familiy of storylines that have been discussed here often.

While the different storylines and assumptions can be a little confusing, the ingredients for making such a forecast can be fairly simple, and I have coded them up into an interactive web site which can be used to explore the world of possibilities. The prediction is based on an idea called the Kaya identity, using numbers published by Hoffert et al. in Nature 1998 [Hoffert et al., 1998]. You could just read the excellent Hoffert et al. paper, but you might also enjoy playing with your own “live” forecasting model, located here.

The Kaya identity states that

C emission = population * GDP/capita * watts/GDP * C emission/watt.

Population is easy to forecast a few years in advance, and impossible to forecast a century in advance. In general, the rate of population growth on earth is slowing, and the hope is that population will level off at some figure in the coming century. The first parameter in the Kaya model is then the asymptotic leveling-off population.

The second term in the equation is the GDP per capita. Adjusted to 1990 U.S. dollars, this term has risen by about 1.6% per year over the past century. You may adjust the growth rate to whatever you like, and you will see how well your growth rate hindcasts the past as well as what it predicts for the future.

The third term is called the energy intensity. It is the number of watts of energy required to produce a dollar of GDP. The energy intensity reflects energy efficiency, and as such has been declining at a historical rate of about 1% per year. The energy intensity also reflects the character of industry, light versus heavy and the balance between manufacturing and services in making up the GDP.

The fourth term is the carbon efficiency of energy production. Coal releases more carbon per energy yield than oil or gas do, and nuclear or renewable energy sources release almost no carbon at all. Historically the carbon efficiency has been decreasing by 0.3% per year.

Change those parameters and push the button. The top two plots in the resulting output will be the predicted carbon emissions to 2100, and the resulting pCOs atmosphere using the ISAM carbon cycle intermediate complexity model. (Yes, the ISAM model was run just for you. Change the inputs and you’ll get a different answer.) What if we decide that this climate trajectory is not acceptable, and we decide to replace some of the carbon-based energy with some new carbon-free energy source? Let’s assume we just replace coal, because oil and gas will be gone soon anyway. The third plot then shows how much carbon-free energy we will require in the coming century if we are to stabilize CO2 at some level. The lower the level, the more carbon-free energy. The third plot shows curves for the IPCC 350, 450, 550, 650, and 750 ppm stabilization scenarios. The result is typically several tens of terawatts of carbon-free energy will be required by the end of the century. For comparison, global energy production today is about 13 terawatts.

Where could this energy come from? Pacala and Socolow [2004] propose a solution (only good to 2050, not 2100) based on a combination of technologies and conservation strategies. There are also a few radical ideas such as solar cells on the moon , beaming energy back to space by microwaves, and high-altitude windmills , flying like kites in the jet stream, tethered on conducting cables. Carbon sequestration is also an option; saline aquifers within the earth are permeable enough for injected CO2 to spread out away from the injection point, yet large enough to flush the bulk of the fossil fuel carbon into.

The bottom line is that the change in the world’s energy infrastructure that would be required, to limit the CO2 concentration in the atmosphere, is not small. A few Toyota Priuses are not going to do it, nor is the Kyoto Protocol by itself even close to solving the problem. Conservation helps, but the historical rate of improvement in energy efficiency is already built into the forecast. There is some scope for trade-offs over time, cuts in emissions now versus cuts later. Ultimately, however, tens of terawatts is a lot of carbon-free energy.

Hoffert, M.I., K. Caldeira, A.K. Jain, and E.F. Haites, Energy implications of future stabilization of atmospheric CO2 content, Nature, 395, 881-884, 1998.

Pacala, S. and R. Socolow, Stabilization wedges: Solving the climate problem for the next 50 years with current technologies. Science 305: 968-972.

22 Responses to “Make your own forecasts of future energy, carbon emissions, and climate”

  1. 1
    Dunc says:

    “nuclear […] energy sources release almost no carbon at all”

    If only this were true… However, reactor fuel doesn’t grow on trees – it needs to be dug out of the ground, refined, and transported very long distances. Also, nuclear reactors are extremely energy-intensive both to build and decommission, plus all the waste has to be transported, processed and stored.

    The only decent full-lifecycle analysis of nuclear power I have seen indicates that in the best possible case, nuclear power emits (or rather, requires the emission of) roughly one-third as much CO2 as the equivalent gas-fired plant. However, this assumes the use of the very highest ore grades available, and these ores are in relatively short supply. As ore quality falls, CO2 emissions rise rapidly.


    Of course, there are similar lifecycle concerns for renewable sources, but they are much less severe.

    [Response:Wow, this was terrifically interesting. The study you cite looks thorough and credible, although I could wish for something like this published in the peer-reviewed literature. Thanks, David]

  2. 2
    Thor Olson says:

    Might I mention that the cost of implementing Dr. Criswells Solar Cells on the Moon can be heavily influenced by robotics. The successful development of electrical acceleration to get things off of the moon (mass driver) would potentially reduce the costs further. The moon has about a 3 second communication delay meaning that you potentially could develop the moon with people sitting in Houston or elsewhere.

    It takes a long time to ‘grow’ a lunar manufacturing base so don’t assume that high energy microwaves would be necessarily beamed to earth on startup. Like all controversial technologies, it would be best to prototype this at a remote location like the middle of the Pacific.

    Even if you never beam a watt of energy back to earth, it is still a really nice asset to have terrawatts of power on the moon as sort of an insurance policy. Somewhere in the future if the climate gets bad and/or the ice caps melt, where else are you going to find enough power to put humpty dumpty back together again?

  3. 3
    Arthur Smith says:

    Dunc – that’s a typical energy payback issue: of course every industrial product requires energy in its production, and so emits CO2 via the carbon intensity we currently have. But there’s nothing intrinsically in nuclear power that results in CO2 releases – if all the energy required were supplied from nuclear power plants, there would be no CO2 involved at all.

  4. 4
    JC says:

    Color me dense, but I don’t quite get what the calculator is telling me regarding carbon-free energy requirements. You say that’s the amount of power that will need to be generated through carbon-free means in order to reach stabilization at a given CO2 ppm by 2100. But if we play with the numbers a bit, reducing the plateau population to 8.5 (the lower end of the revised UN medium scenario) and increase the W/$ efficiency growth to 2% (significantly less than what California managed during the 2000-2001 enron crunch, and a bit less than what the US as a whole managed during the 1979-1981 oil shock), the results are confusing. The 650ppm point, for example, would need -4 terawatts of carbon-free energy… what does *that* mean?

    [Response:That means that pCO2 is actually less than the 650 stabilization trajectory, without any new carbon free energy. You can see in the plots that CO2 emission start falling immediately under this scenario. David]

  5. 5
    Aaron says:

    “The energy intensity reflects energy efficiency, and as such has been declining at a historical rate of about 1% per year.”

    Does this figure reflect a single country, or the entire world?


  6. 6
    Glen Raphael says:

    Is there any reason to expect the global population to ever “level out”?

    My guess is that global population will start shrinking as soon as it stops growing. The birthrate is below replacement levels now in the first world; as the economy grows that will be the case in the third world too, so a century or two hence, the population could easily be less than it is now. But when I plug in a “level out” below our current population it makes for an awfully funky curve…

  7. 7
    JC says:

    If I understand you correctly, then what the graph shows is the answer to “…of the amount of energy produced, how much needs to be carbon-free?” If so, then I think I see the missing element. The one other line on that graph that would provide useful data would be “change to total energy output,” integrating the increase in energy required for the growth in GDP/capita and population along with the increased efficiency of production. That way, it would be visible at a glance how much of the carbon-free production is from the new energy sources and how much is from replacement of existing energy sources.

    [Response:This is a good suggestion. I’ve modified the model web page to do this. Thanks!]

  8. 8
    Alain Henry says:

    We’re closer to the field of policy analysis than to that of climatology here!
    Simplified approaches like this one are useful. However, it is assumed here that policies will not change in the future, and leaves the impression that there’s not much we can do about it. The question is not only “How much needs to be carbon free?”, which focus of the fourth factor in the right term of the equation. We can also focus policies on the three other terms. A full policy analysis must thus ask four questions: What can we do about
    – demography ? (education, development aid, …)
    – the economy? (what growth rate do we want ? in which country ?
    – energy efficiency and energy use ? (housing insulation, consumption patterns, transport, industry, energy sector, etc.)
    – carbon intensity of energy ? (carbon free energy, etc.)

    Policies can (and I believe should) be developed in these four areas

  9. 9
    Max says:

    I don’t like this meter out of many reasons. First, world population growth is not constant and while the first world has a reduction of population, it is also starting to affect the second and third world, now.
    Also, I see this falling of energy prices which is a false idea, because neither the green energy productions are cheaper, nor is nuclear power going to be cheap in the long run. Also, in the next 10 years we will have the cost-effect of the Kyoto-Protocol that opens an imaginary market with new (not so imaginary) costs imposed on energy suppliers.

    Also, CO2 is not only caused by people and power plants, but also by individual traffic (cars, planes, ships etc.). Not to mention that the hybrid cars will create new kind of gases en masse.
    So, this program is a rough estimation and I think will not help much to further the real implications of CO2 reduction versus climate adaption.

    [Response:The Watts/$ is not a price of energy, but the amount of energy it takes to fuel $1 of GDP. More like an energy price for money than a money price for energy. Also, CO2 from transportation is part of the extrapolation. If we started to travel less or drive more efficient cars, that would be reflected in the Watts/$ energy intensity. David]

  10. 10
    Henry Molvar says:

    Re #1 As a chemist and materials engineer I have always been opposed to the excessive burning of fossil fuels for energy generation. I have viewed this as the unnecessary destruction of valuable raw materials which will eventually be sorely needed to produce the lubricants, engineering plastics, coatings, adhesives, etc. that our modern world is becoming increasingly dependent upon. Adding the issue of GW through CO2 emissions greatly magnifies this problem.

    Therefore any means of reducing the burning of fossil fuels, whether it be through conservation or alternative sources such as solar, wind, nuclear and others is a huge benefit to our generation and future generations.

    Nuclear power is potentially a savior and a curse. However I believe that all the safety and environmental issues associated with nuclear are soluble through enlightened planning, research and development, modern engineering, and political farsightedness.

    Sure it can be demonstrated that building and operating nuclear power plants presently leads to significant CO2 emissions in our fossil fuel economy. But the same can be demonstrated for almost everything that humans do, right down to heating water for a cup of tea.

    But why let that distract us from the big picture: the less dependent we are on fossil fuels the fewer the CO2 emissions and the more raw materials we leave for future generations

  11. 11
    Dunc says:

    I simply don’t believe that it’s feasible to ramp up nuclear power to anything like the levels that would be necessary to make it self-sustaining without massive fossil fuel inputs.

    Firstly, most of the mining activities take place in remote locations in the 3rd world – hardly the most likely place to set up a couple of terawatts of carbon-free power to run your mines and refineries. Then there’s the massive transport infrastructure required – OK, in theory you can run it on nuclear-backed hydrogen, but that technology is at least 20 years off full-scale deployment. And of course, the other really major issue is time. It takes between 10 and 20 years to get a single nuclear power station up and running. How long will it take to build the thousands necessary to support even our current energy use, never mind the inevitable, massive, additional requirements from such an energy-intensive technology? How much will that cost, given that we seem to be on the threshold of a massive and permanent increase in oil prices?

    Finally, it must be remembered that uranium, like oil, is a strictly non-renewable resource, and subject to exactly the same production dynamics. Only the initial, easy production is cheap. Even if there were enough uranium in the world to supply our energy requirements for more than 3 years (which there isn’t, as far as anyone knows), once you get past the depletion midpoint the economics get silly very quickly.

    The idea of using nuclear power to make nuclear power possible is a serious bootstraping problem. Not necessarily insoluble, but highly unlikely to be solved in the necessary timescales – ie, before we run out of easily accessible uranium anyway.

  12. 12
    Max says:

    And we have to see when oil is getting short in supply the prize will rise until the production of oil will be so expensive that no one wants it anymore. Until then, the industry will have developed new methods, because enterprises have to make profit. I don’t think that nuclear power is an option, because there are not so many places where we can stuff all the waste into. On the other side, the conservation of ressources is of no use at all, because ressources have no value by itself and to ration them will only bring further problems up.
    Energy is the backbone of the human society and thus we will need as much energy as necessary to find new ways to get energy.

    I think it is similar to the ressource oil:
    A lot of people thought that there’d be a limit to oil ressources until the year 2000 and then we won’t have anymore oil. However, we have more ressources than before and there has been produced more than they had predicted there is. This is why the limit of ressources is so hard to define, because human ingenuity often proves that limits are not so strict as we believed some years ago.

    If you compare the industry of a few decades ago and today, the reduction of dangerous gases that pollute the environment has been done neatly. Today, private enterprises are working more pollution-free than ever before, but still CO2 is rising, if I am correct. So, what can be additional reasons? Human influence, but I don’t hope that you want to suggest that there are too many humans.
    Private traffic, perhaps, but this is also a cost-effect-calculation that is not easy to do, because it has many side-effects that affect everything we do.

    Perhaps then, we should take Mann truly serious and think that there is a climate change that comes faster and we should start thinking a lot to invent cheap methods to reduce the problematic molecules’s output.

  13. 13
    Steve Bloom says:

    Re #1: Thanks so much for that link. Regarding David’s response, I wonder exactly what peer-reviewed publication would be appropriate for something like this? The subject matter certainly crosses a lot of fields.

    Re #11: That three-year figure may be low, but even if it’s thirty years the inevitable consequence of a decision to rely substantially on nuclear power will be a shift to breeder technology. Contrary to the impression left on the study’s main page, such technology is available right now — it’s just that it tends to create a horrible mess (mainly in the reprocessing). Here in the US, taxpayers are still paying dearly for cleaning up our part of that mess (which is a result of producing the Pu needed for nuclear weapons); the former Soviet Union’s portion is even worse. All of the negative consequences of a Pu-based energy economy (much larger-scale than that needed for weapons) are painful to contemplate. Of course, the nuclear industry will happily take us down that path as long as someone else is willing to pay for it.

  14. 14

    Simplistic models postulating constant year-over-year decreases in energy consumption are worthless; there are some functions which can conceivably have their energy requirements reduced to zero (e.g. space heating can be eliminated by sufficient insulation) but others have hard physical floors (lighting, formation of chemical compounds) or are already near the limits of diminishing returns (freight transport by electrified rail).

    I’d give an exercise like this to a high-school science class, or maybe bright middle-schoolers; to be useful it needs to be broken down by category of use. It could make an interesting assignment with a spreadsheet.

  15. 15
    Andrew Dodds says:

    Re #1: This study invariably crops up whenever Nuclear energy is debated as a low CO2 emission source of baseload electricity. I would strongly advise serious analysis of the assumptions behind it.

    Firstly, the study is based on only U-235 being used; this would come as a surprise to those reactor operators who already rely on some of the U-238 becoming Pu-239 and fissioning during normal reactor operation. Secondly, the study assumes that coal burning is used for every stage of extraction and enrichment – it is not explained why this has to be so. Thirdly, it is simpy assumed that decomissioning will take three times as much energy as the already-inflated estimate for construction. At every stage, the highest possible energy costing is used, together with the lowest possible estimate of resources.

    Processes such as fuel breeding or obtaning uranium from seawater – both perfectly feasable but currently uneconomic – are dismissed out of hand. The impression I get is that the authors started with a conclusion and went looking for data to fit. Certainly there is no sense of what happens to the CO2 emissions of NG fired plants as you go to more distant natural gas sources and/or LNG; this is apparently a free lunch.

    As far as the ‘bootstrapping problem’ goes, the french are already doing this, so I’m not quite sure what the problem is. Luckily, reprocessed plutionium has the incorrect mix of isotopes for a nuclear weapon – anyone who can separate Pu-239 and Pu-240 could in any case just enrich normal uranium to weapons grade, which is easier.

    The world’s biggest suppliers of Uranium are currently Australia and Canada; these are not third world countries (as in #11).

    Nuclear power is not a perfect energy source, but compared to the continued burning of coal – which is having a strong effect on the environment right now, or large scale reductions in overall energy use – something that is not politically feasable – it does have considerable advantages.

  16. 16
    paul gosling says:


    It is interesting that the lifecycle analysis of nuclear power you refer to makes its comparison with gas fired power stations. Gas is widely recognized as the cleanest of the fossil fuels, and also the one which will run out first (in the UK we will soon be importing most of our gas by tanker and pipeline). It requires almost no processing, is easy to transport and produces no waste. However, most electricity is generated by coal with gas only representing about 15%. Coal requires similar mining to uranium, with many of the environmental problems which the authors only equate with uranium mining. It is much more energy intensive to transport than either uranium or gas and produces a significant quantity of toxic waste ash. I wonder how nuclear power would look compared to coal?

  17. 17
    Dunc says:

    Paul, Andrew – thanks for your comments. I’ve never really been entirely sure about the provenance of that study, so it’s always good to get other people’s reactions.

    Andrew – As to the source of the uranium – yes, you’re right, “third world nations” was incorrect. However, both Canada and Australia are (a) bloody massive, and (b) mostly empty. I can still see definite practical problems with running the mining infrastructure from electricity in these sorts of locations.

    On the subject of coal, would I be correct in thinking that you’re commenting from the US? Over here in Blighty, NG displaced coal from the mid-nineties (Paul, you seem to be under the impression that we’re still using the generating mix we were in about 1995 – although I can’t blame you, as current figures seem almost impossible to come by). There’s still a fair bit of coal in the generating mix, but gas is where we’re betting the farm. From next year, something like a third of our NG is going to be imported in tankers from Qatar, which is obviously going to change matters somewhat.

    My impression of both fuel breeding and uranium extraction from seawater is that these technologies are only currently theoretical. Has anyone even done a small-scale pilot of either technology? (Not a lab eperiment, but and actual pilot implementation?) If not, then their feasibilty is entirely unproven.

    Finally, I’m not sure that coal mining is really like uranium mining. With coal, you mainly just dig it out of the ground with big powered shovels. Uranium mining tends to involve a lot more nasty, expensive, energy-and-chemical-intensive processes to extract and refine the ore once you’ve dug it out, with the resultant massive heaps of heavily contaminated tailings.

    Finally, neither of you seem to be addressing the waste issue… As a Scotsman, living within a couple of hundred miles of several of the preffered waste disposal sites and in the country that’s already playing host to one of the worst nuclear waste displosal attempts in history (the big hole up at Dounray that exploded a while back), this is an issue that concerns me.

    I’m very interested in you opinions. But have we hijacked this thread too much already? ;) If you’d like to discuss these issues offline, email me at dunc [at] gregorach [dot] com. :)

  18. 18
    paul gosling says:


    On a global scale coal accounts for 40% of electricity generation, natural gas 15% and nuclear 16%. And as natural gas is going to be the first fossil fuel to run out and China and India are building their growth on coal and the USA is turning more to coal and away from oil these proportions are going to change.

    The waste issue was not what this tread was about, however, the waste from fossil fuels is going to cost a lot more to deal with than the waste from nuclear power.

  19. 19
    Andrew Dodds says:


    I’m not sure what size has to do with the feasability of electrifying mining operations; nor any issues of tailings. I do get the impression that when it comes to nuclear, there are people who try any possible argument against without doing the research. That may sound aggressive, but considering the number of times I’ve seen (and had to deal with) the same approach being used to argue against global warming, evolution, vaccination or just about anything where science comes into conflict with people’s a priori beliefs, I think it’s justified.

    As far as Uranium from seawater goes, it’s been demonstrated in situ, so it would simply be a matter of scaling. Estimated costs are quite a bit higher than current mined uranium, but Uranium can increase in price by a lot with little impact on generation costs. Breeder reactors have about 40 reactor-years of assorted operation. The feasability of either comes down mostly to the cost of uranium.

    I am British, I live about 20 miles downwind from a nuclear plant. Personally, I find our switch to NG as almost criminal in stupidity; it’s drained the North sea in short order, thus making us energy importers. The switch to long distance imports means we lose any GHG savings – although they show up on other people’s emissions, the atmosphere does not care.

    As far as waste goes, recycle the transuranics and turn the fission products into nuclear batteries. Anything remaining can be liquified and pumped under salt-sealed old gas fields; these are already leak-tested on a timascale of millions of years. Burying the stuff as currently proposed is both wasteful and very hard.

  20. 20
    Dunc says:

    Well, I’m in the Scottish Green Party, and part of their Energy Policy Working Group… The consensus there is very much against nuclear, but I’m trying to remain open-minded about it. However, if I were to try and argue in favour of nuclear in that venue, I’d need very good answers to all these questions, for the very reasons Andrew cites.

    You’ve both given me lots to think about. Thank you! :)

  21. 21

    Andrew Dodds:  I am slowly going through the van Leeuwen/Smith paper and I am finding exactly the same things that you are.  I find it particularly amusing that they postulate an energy penalty for “disposal” of depleted uranium, when the quantity already in the ocean would argue for taking the excess as oxide and just dumping it in barrels from any convenient ship.

    FWIW, you’re not going to turn gamma-emitters into nuclear batteries.  There’s a reason that the spent fuel is treated as waste; it’s not capable of giving off enough energy to make it worthwhile to capture it.  I suppose you could use thermovoltaics to self-power monitor and alarm systems on storage casks, but that’s going to be about its limit.

  22. 22

    Nuclear Power and Climate Change: Is Our Choice Glow or Cook?
    Nuclear energy, which not so long ago seemed an obvious dead-end to most people concerned about the environment, has seen its fortunes rise a bit…