[Warning: This post has lots of numbers in it and isn’t short. But I hope this will be a “Eureka! I finally get it” post for those who read it to the end, as I’m going to unlock the final mystery of the wedges.]
I’d like to thank (!) Roger Pielke for his post, “Joe Romm’s Fuzzy Math.” Not because his analysis is correct — it isn’t. Not because of the tone — he says my climate solution is “fantastically delusional,” which better not be the case or our children and the next 50 generations are screwed. No, I’m thanking him because explaining why he isn’t correct will illuminate two key points in the climate solutions debate:
- Princeton’s Socolow and Pacala make an important but surprising assumption in their wedges analysis (here) that few people realize. Anyone who wants to come up with their own 14 wedges (as opposed to accepting my solution laid out in Part 2), must understand what they did.
- When you understand what Princeton did, then you’ll understand why Pielke’s critique is fundamentally wrong, and then I think you can understand at a more intuitive level why his Nature article is wrong, too.
THE WEDGES’ FUZZY MATH
Let’s start with what looks to be a major analytical mistake in the Princeton analysis. Recall that one wedge is a climate strategy that ultimately saves 1 billion tons of carbon a year or 1 GtC/year.
If you look at the original paper (here), they identify a typical wedge (#9 on their list) as “Nuclear power for coal power.” That “would require 700 GW of nuclear power with the same 90% capacity factor assumed for the coal plants” [which means the plant delivers power 90% of the time, or 8760 hours/yr x 0.9 = ~7900 hours/yr.]
But wait, you say, everybody knows a typical coal plant has a carbon intensity of 290 kilograms per Megawatt-hour. Or at least everyone who reads page 9 of the online Supplemental material (available here with subscription), much of which is now on their website, wedge by wedge, here. As they explain in the “efficient baseload coal plant” wedge (here), citing the IEA’s World Energy Outlook, 2002:
Year 2000 carbon in and electricity out for coal-based power plants were, respectively, 1712 MtC/y and 5989 TWh/y, resulting in a carbon intensity of 290 gC/kWh.
I have always preferred to use tons per MW-hr, so 290 gC/kWh = 290 kgC/MWh = 0.32 tC/MWh.
[For those who prefer CO2, that is ~1.2 tons/MWh for a typical coal plant — which, by the way, is a handy number to keep in your head for back-of-the-envelope calculations.]
But wait. If the 700 GW (= 700,000 MW) of nukes are replacing 700,000 MW of coal running 7900 hours a year and spewing out 0.32 tons of carbon per MWh, then this wedge yields 700,000 MW x 7900 hrs x 0.32 tC/MWh = 1.77 billion tons of carbon.
Oops. Isn’t a wedge 1 billion tons of carbon? This would seem to be a big mistake. Ah, but you didn’t read the fine print:
Princeton clearly states it is assuming that, in 2054, the nuclear plants are displacing “efficient baseload capacity.” And they clearly state (well, “clearly” if you look hard enough, in this case in Option 5) that is a coal plant whose “efficiency has improved to 50%.” Now, if you read the supplemental online material, which nobody I have ever met has except Socolow himself, then you know from page 14 that such a plant “has an average carbon intensity of coal plants in 2054 of 185 gC/kWh” = 0.20 tC/MWh. Avoiding or shutting down those coal plants gets you roughly the traditional wedge: ~1 billions tons of carbon saved by the 700 GW of nuclear.
WHY DID PRINCETON MAKE THIS “DANGEROUS ASSUMPTION”
Why did Princeton do this? Well, they don’t say exactly — and as I noted in Part 1, they don’t have an analytically rigorous baseline. So to make up for that flaw, they presumably wanted to incorporate some of the normal technological improvements that you would expect over 50 years. Indeed, if they kept technology frozen, people would no doubt have accused them of overstating the challenge. But yes, this does mean that they assumed that Business as Usual (BAU) included improvements in decarbonization — the same “dangerous assumption” Pielke criticizes the IPCC for. As we will see, however, this assumption isn’t all that dangerous, and ironically, it has the reverse impact on the wedges analysis than you (or Pielke) might think.
[NOTE: Stay with me readers, we are close to the crux of the entire debate that unlocks the final mystery of the wedges.]
Princeton says “Our BAU simply continues the 1.5% annual carbon emissions growth of the past 30 years.” That is fundamentally why they assume coal plants will be 50% efficient in 2054 when they are only about 30% efficient today.
But in fact, since 2000, carbon emissions have grown much closer to 3% per year. Does that mean we need twice as many wedges? Does this mean Pielke is right that I am assuming 16 (!) free wedges? Is he right when he states (here):
… the only way that Joe Romm’s ambitious solution even comes close to the mark is by assuming a significant spontaneous decarbonization of the global economy.
To repeat the key point, it is not me who makes the spontaneous decarbonization assumption, it is Princeton. As I have exlained, the wedges already assume “a significant spontaneous decarbonization of the global economy” when they make the 2054 reference baseload coal plant for the wedges 50% efficient, or 0.20 tC/MWh.
What happens to the analysis if that decarbonization doesn’t occur for, say, the reason that a certain huge Asian country starts building an unimaginable amount of traditional coal plants rather than, say, superefficient, partially decarbonized coal plants. The number of wedges doesn’t change at all!
Why? Because now, when the world builds a “wedge” of nuclear plants or concentrated solar thermal plants, it is avoiding 700 GW of relatively inefficient, un-decarbonized coal plants. And that “wedge” is worth 1.77 billion tons of carbon.
In other words, the wedge in a world without “significant spontaneous decarbonization of the global economy” is worth 1.77 wedges from the world Princeton imagined where steady decarbonization has been assumed.
In other words, in a coal-mad recarbonizing world, you can get a 1 GtC wedge with a lot fewer nukes, CSP plants, efficiency, and so on.
REDOING (briefly) THE SOLUTION TO GLOBAL WARMING IN A WORLD THAT AVOIDS SIGNIFICANT SPONTANEOUS DECARBONIZATION
Imagine we’re in a world of 8 GtC that has gone mad for old coal, so emissions are rising 3% per year. Assume no “significant spontaneous decarbonization of the global economy” (i.e. we stay mad for old coal for decades). That means the BAU projection for 2050 is about 28 GtC. But we need to be at 4 GtC in 2050 for the 450 ppm path. So we need to cut 24 GtC off of BAU, which is 24/1.77 = 14 wedges!!
That’s right. If we use Princeton’s BAU assumption of steady decarbonization of coal plants, which returns us to their BAU assumption of 1.5% annual increase in carbon emissions, we need 14 wedges to get down to 4 GtC in 2050. And if we use Pielke’s assumption that does not allow “significant spontaneous decarbonization,” which means that BAU is 3% annual increase in carbon emissions, we still need the same exact 14 wedges to get down to 4 GtC in 2050. In Princeton’s world, those 14 wedges are worth 1 GtC/yr each in 2050 because they replace efficient, futuristic, decarbonized coal plants. In Pielke’s world, those 14 wedges are worth 1.77 GtC/yr each in 2050 because they replace traditional, inefficient, undecarbonized coal plants.
I really hope that is clear.
And that I think is another way to explain why the IPCC’s assumption of spontaneous decarbonization is not “dangerous,” since it doesn’t affect the outcome of their analysis. If we posit a “frozen-technology” scenario as Pielke does for his baseline in the Nature article, then all of the low carbon technologies the IPCC believes we need to deploy to stabilize at 450 ppm are all suddenly much more effective — since they are replacing 2000-era technology, and not future technology that has spontaneously decarbonized.
And I hope that is now clear, too.
I am not, however, going to accuse Pielke of fuzzy math in his post because, really, anybody could have made that mistake. I myself had to read the wedges paper and its supplemental material several times — and talk to Socolow a couple of times — over the last few years to figure out what they were really doing. I really think Socolow and Pacala should have spelled this out much more clearly up front. Pielke concludes:
Joe’s fuzzy math explains exactly why innovation must be at the core of any approach to mitigation that has a chance of succeeding.
Since my math isn’t fuzzy, you won’t be surprised that I believe my analysis here is yet more grist for my argument that we don’t need major technologies breakthroughs (at least in the first half of the century) to stabilize at 450 ppm. And as for my solution being “fantastically delusional,” well, I suppose that is for others to judge … I have always said it is not close to being politically acceptable today. But as we will see in Part 3, which discusses the possible role of breakthrough technologies in getting to 450 ppm, if my solution is fantastically delusional, then I do not think there are words to describe Pielke’s strategy.
[Note: Yes, I did everything in this post in terms of carbon even though I said I wouldn’t (here) because Princeton does everything in carbon, and I wanted to use the numbers they use in order to minimize the possibility of inconsistencies or mistakes. As an aside, I can’t tell whether what they label “tons” are or are not metric tons, which should be labeled “tonnes” — if anyone knows for certain different please let me know. A ton = 907 kg = .907 metric tons. It is only a 10% difference — and as is probably clear by now, Princeton’s analysis is simply not precise to better than 10% to 20%, if that, so it doesn’t really matter.]