Watt about the 95% certainty?

It has been suggested that the forthcoming IPCC AR5 report will claim that it is at least 95% percent likely that human activities – chiefly the burning of fossil fuels – are the main cause of warming since the 1950. Watts Up With That (WUWT) has a new post called when somebody hits you with that new IPCC is 95% certain talking point, show them this. What Anthony wants them to be shown are the two figures below, one showing HADCRUT3 temperature anomalies for 1895-1946 and the other showing temperature anomalies for 1957-2008.

HADCRUT3 temperature anomalies for 1895-1946 and 1957-2008 (credit : Anthony Watts, WUWT)

HADCRUT3 temperature anomalies for 1895-1946 and 1957-2008 (credit : Anthony Watts, WUWT)

Anthony is suggesting that the warming during the first half of the 20th century is essentially the same as in the second, so how can we claim that the latter warming was almost exclusively anthropogenic. Firstly, they’re not quite the same. The surface temperatures rose faster from 1957-2008 than from 1895-1946, but that’s not all the relevant. A much more appropriate figure is probably the one below. It shows surface temperatures (red), and solar insolation (blue) since 1880. Solar insolation (the amount of solar energy hitting the earth per second) rose during the first half of the 20th century and did indeed dominate the warming (although anthropic influences also existed). However, since about 1960 it’s been dropping while CO2 concentrations and surface temperatures have continued to rise.

Surface temperatures and solar insolation since 1880 (credit : Skeptical science)

Surface temperatures (red) and solar insolation (blue) since 1880 (credit : Skeptical Science)

So, in a simple sense it is difficult to explain the warming since 1950 as anything other than anthropogenic given that solar insolation has been dropping. There is, however, another issue. Focusing only on surface temperatures when determining if global warming is happening or not, is simplistic. Global warming is about rising energy, not only about rising surface temperatures. Currently we’re accruing energy at a rate of about 1022 J per year (mainly going into the oceans). However, surface temperatures are higher than they’ve been for most of the last 1000 years and are close to being as high as they’ve been for the last 10000 years. Solar insolation is dropping, temperatures are almost as high as they’ve been for thousands of years, so how can we be undergoing some kind of natural global warming? We should be starting to see cooling if the only influences were natural. The only explanation currently available is that the continued warming (i.e., increase in total energy) is because we continue to add CO2 to the atmosphere and it is acting as a greenhouse gas and trapping outgoing long-wavelength radiation.

So, if you want to mislead people about global warming, then show them the two figures Anthony suggests. If you actually want people to understand the current scientific view on global warming, then show them more than simply those two figures. The point is that it is extremely difficult to explain our current warming as anything other than anthropogenic. I suspect that most scientists are certain that it is all anthropogenic. The only reason that it is being framed as 95% likely is simply because scientists tend to acknowledge that maybe there’s something we don’t yet understand or something that we haven’t yet considered. In this case, I suspect that that is highly unlikely.

This entry was posted in Anthony Watts, Climate change, Global warming, Watts Up With That and tagged , , , , , , . Bookmark the permalink.

45 Responses to Watt about the 95% certainty?

  1. Tom Curtis says:

    The absurd, and common among deniers, fallacy in Watt’s presentation is the assumption that the warming from 1895-1946 is entirely natural. This is based on the IPCC’s claim that the early twentieth century warming lies within the 95% confidence interval for warming as a result of natural forcings alone. Of course, if later warming is not natural, or if anthropogenic forcings have the same warming effect as natural forcings, then inevitably the early twentieth century warming has an anthropogenic component. As it happens, the change in CO2 forcing from 1895-1946 is 0.27 W/m^2, ie, three times the 0.09 W/m^2 change in solar forcing over the same period. The transient response to the CO2 forcing would be 0.145 C or about a third of the total warming between the 1890s and 1940s. The net anthropogenic warming in that period may be less than that, due to aerosols, but the notion that the entire warming is natural is not tenable. Absent that absurdity, however, and the implicit argument that the late twentieth century is indistinguishable from wholly natural warming and hence may itself be natural simply does not hold water.

    Of course, the fact that the argument depends on carefully with holding information (as you point out), shows it to be specious, and of dubious honesty.

  2. Sou says:

    There’s a Guidance Note for Lead Authors of the IPCC Fifth Assessment Report on Consistent Treatment of Uncertainties. Virtually certain is 99%-100%, extremely likely is an extra category and it’s 95%-100%. I would translate these two categories from science-speak to gal in the street speak, as anything above 95% is the same as absolutely certain! Can’t be more certain. You can bet your house, car and bank account on it. In fact, what we’re gambling with is our future.

    The extra 5% is maybe to give just the teeniest bit of wiggle room to make sure everyone’s on board and just in case of who knows what.

    To hang one’s hat on the 5% would be giving oneself much worse odds than Russian Roulette – by a factor of almost four. Even “very likely” at 90-100% is much worse odds than Russian Roulette.

    Click to access uncertainty-guidance-note.pdf

  3. Tom Curtis says:

    Sou, absolute certainty (literally, a certainty that cannot be dissolved) is one for which there is literally no chance of error. We are absolutely certain that 2 + 2 = 4; but about the conservation of energy, or the special and general theories of relativity, there must always remain some real doubt.

    Having said that, the doubt is not significant.

    Assuming an error on this part is like playing russian roulette with a Desert Eagle. We know the firing pin will strike the cartridge. We’re just praying the cartridge misfires.

  4. Yes, I agree. People seem to ignore that the anthropogenic component during the first half of the 20th century was not insignificant and may well have dominated (as you point out the anthropogenic forcing for the period 1895-1946 was 3 times the change in solar forcing). It’s simply that it has clearly been the dominant component in the latter half of the 20th century. I was going to say that it surprises me a little that this point isn’t always made as clearly as it could be, but I may have made this point poorly in my own post. I should probably have checked (as you have) the anthropogenic contribution to the forcing.

  5. Tom Curtis says:

    For what it is worth, the GISS forcings show a total difference between 1895 and 1946 of 0.166 W/m^2 for all anthropogenic forcings, and 0.137 for all natural forcings. That probably underestimates the impact of natural forcings in that volcanic forcings fluctuated violently over the first part of that period, and an appropriate running mean of volcanic forcings would probably better represent their actual effect.

    By way of comparison, between 1957 and 2008, anthropogenic forcings were 1.276 W/m^2 for anthro, and 0.262 for natural.

  6. Thanks, that is very good.

  7. A skeptic person would ask: Why did the models got that right? Or why can we reproduce it taking the known forcings as a basis? Or most obvious: Why did land temperatures rise so much faster today then they did earlier?

    A fake skeptic person would show the above image and ask dumb questions. After all, we all know that it is caused by leprechauns! Therefore I certainly agree with Chris Colose, who appropriately tweeted the following yesterday: ““95% confident humans contributing to global warming” meme is going to annoy me…physics constrains the number to be no less than 100%
    “Human” is certainly supposed to be “leprechauns” … stupid little mistake, not worth mentioning 😉

  8. And…

    Watt about .8 C hotter than 1880?
    Watt about warming 30 times faster since end of last ice age?

    Of course we could also as Watt about how to obscure facts and make fossil fuel companies look better.

  9. dana1981 says:

    The deniers are certain to latch onto the reduced lower equilibrium sensitivity limit (1.5°C for 2xCO2), assuming it makes the final report. From what I hear, the lower bound of the transient response will be 1.0°C.

    You can do some quick and dirty math, using the formula ∆T = λ∆F. The net anthropogenic forcing over the past century is somewhere in the ballpark of 1.7 W/m2 (will be interesting to see this value in AR5).

    So I’m glossing over the significant uncertainty in the net forcing (particularly from anthro aerosols) – like I said, this is quick and dirty. But if you want to know roughly how much warming humans have caused thus far, λ = 0.27 (using the transient lower bound of 1°C), ∆F = 1.7, and thus ∆T(anthro-rough-minimum) = 0.46°C, out of 0.8°C observed surface warming. So there you go, even using the likely lower minimum transient response, you get AGW > 50%. While I’m glossing over the large uncertainty in the forcing, I’m also looking at the forcings and warming over the past century-plus, while the anthro contribution is even larger over the past 60 years. Hence the confident AGW > 50% statement is quite sound, even by the deniers’ own low-sensitivity argument.

  10. wow what a load of …. Your plagiarized figure shows the same temperature rise profile as the one you’re criticizing. LOL!

  11. Shub, what are you talking about? Which plagiarized figure? Do you actually read anything before you criticise it?

  12. Oh and by the way, I wasn’t suggesting that there was anything wrong with the temperature rise in the figure I was criticising. I was suggesting that looking at the temperature alone is insufficient to tell you what has caused the temperature to rise. Maybe you should really give this all a bit more thought before LOLing.

  13. Yes, I agree. I think I responded to Chris’s tweet. I can see why they’ve said “at least 95%”. As you say, given our current understanding it has to be no less than 100%. I assume that they haven’t said 100% because they’re being cautious, not because they think there really is a 5% chance that it could be something non-anthropogenic.

  14. You say you are criticizing WUWT. Here we are, with you posting a figure from skepticalscience as an ‘argument’ against WUWT. If I needed to compare them both, I can visit the respective websites directly.

    So your implication of your figure is that we know rising solar irradiance caused the 1910-1940’s warming? Total irradiance in the second half of the 20th doesn’t fall below 1940’s value. Another LOL.

    Temperatures high than thousands of years? Where did you get that from? Even the IPCC doesn’t say that.

    How did humans manage to emit the exact amount of CO2 required to produce a global rise of temperature in exactly the same way solar output managed to produce, 40 years back in time?

  15. “Why did the models got that right? ”

    Yeah right. ‘Success’ of models predicated on reproducing known changes. And you call yourself a scientist.

  16. If you wanted to visit both sites, you’re welcome to do so. That’s no argument for me not including these figures. If you’re going to accuse someone of plagiarism at least look up what the term means. I said “higher than for most of the last 1000 years” (hockey stick studies) and I said “close to being as high as they’ve been for the last 10000 years” (Marcott et al.). At least criticise what I write, not what you think I write.

    “Solar irridiance doesn’t fall below the 1940s value”. So what? Surface temperatures are 0.5oC higher now than they were then. How is that possible if CO2 isn’t providing a forcing.

    Humans didn’t emit the exact amount of CO2 required to produce a global temperature rise in exactly the same way as solar output managed 40 years back. The temperature changes in the two time intervals are not identical and CO2 (as others have mentioned) provided a significant fraction of the warming even in the first half of the 20th century. In the first half, however, solar forcing was making a significant contribution. In the second half it wasn’t (to the warming that is).

  17. Shub, do you not understand how science works? Sometimes you run models to try and understand what you’ve already observed. Simply making measurements doesn’t tell you why something is happening or has happened. Understanding why something has happened requires using models (of one form or another) to understand the data. This is not really something that there should be a great deal of disagreement about. Criticising scientists for running models to try and fit what’s been observed is absurd.

  18. This is getting worse. You mean humans emitted just the right amount of CO2 in concert with solar radiation to produce the same warming in both the first half and the second half of the 20th century?

    As an aside, I can’t look at the WUWT figure and tell which one is which. You should try showing the figure to a bunch of people you know in a blinded fashion. The same holds for skepticalscience’s figure even their bulldozer filtering. I can’t check their source as their website is down.

  19. No, that’s not what I’m saying. That they look the same is irrelevant and you probably know that (at least I hope you do). Unless there was some kind of significant acceleration in the warming, finding two time intervals that look the same (but aren’t quite) is easy and doesn’t signify anything.

    Do you seriously think that because the data looks the same that the cause should be regarded as the same. If someone was pushed down a flight of stairs would that mean every time someone fell down a flight of stairs they must have been pushed? Don’t bother answering that. I don’t really feel like extending this rather pointless discussion any further. Maybe just give it a little thought if you wish to do so.

    Anyway, the left-hand one is the most recent time interval, which you can tell because of the big ENSO spike which occured in 1998.

  20. This is getting even worse. Can you tell by looking at a man tumbling down a flight of stairs whether he was pushed or whether he tripped on his shoelaces and fell?

  21. My guess would be no, which was the point I was trying to make.

  22. You’re right about one thing though. This is definitely getting worse.

  23. “My guess would be no, which was the point I was trying to make.”

    If that’s point you are making, you are admitting that evidence for CO2’s causative role is circumstantial at best. CO2 is capable of pushing a man down stairs so he bobbles down the exact same way he did when fell by himself.

    Of course, it is getting worse too – things are nice when you have bunches of people agreeing.

  24. No, that’s not what I’m saying Shub. You really should stop trying to over-interpret what I’m saying, because you’re clearly not very good at it.

    For the very last time, I’m simply pointing out that only looking at one set of data, doesn’t necessarily tell you what is causing something (i.e., even if the data looks the same, what has caused the changes may not be the same in both cases). This is not a complicated concept and I’m not going to spend any more time discussing this with you.

  25. Tom Curtis says:

    Wotts has several times mentioned that increase in temperature over the two periods is not the same. That is not readily apparent from the two panels of the first graph, so I have produced an equivalent graph taking running 60 month means of the data directly overlaid to make the comparison easier. Doing so shows that the late twentieth century warming was approx 0.16 C greater (or 46% greater) as shown by comparing the mean of the first and last 60 months respectively.

    In preparing the figure, I noticed a couple of oddities in the WUWT graph. First, the first few months of the 1895 data have been dropped of, an omission which increases the visual similarities of the two curves. Second, the vertical distance between the lowest and highest points of the two curves are identical, whereas in the actual data the range is 6.8% greater in the 1957-2008 series. I only measured one difference, and it is possible that that odd coincidence of ranges in the WUWT graph is due to poor image reproduction. Never-the-less, this odd coincidence does falsely exaggerate the similarities of the warming rates.

  26. Tom Curtis says:

    Not the main point of this post, but the evidence of a reduction in OLR at 15 micro-Meters (ie, the main CO2 absorption/emission band) relative to that from the “atmospheric windows” is direct and incontrovertible. It is also very exactly predicted by lbl- radiation models. The former observation is physically inconsistent with CO2 not causing warming of the surface. And as the lbl models predict a further reduction of OLR in that band with increased CO2, the evidence for the enhanced greenhouse effect is so incontrovertible that even Lindzen and Spencer will have nothing to do with those who deny it. The physics of radiation, by the way, is non-chaotic and deterministic and macroscopic scales. Therefore, that is a straightforward prediction of a well confirmed physical theory, which has been further confirmed by observation (see main post at link).

    Given this, Shub’s claim that the causative role of CO2 is “circumstantial at best” merely indicates a massive disconnect between Shub’s mind and the actual physical evidence.

  27. Indeed, that’s why I find this whole debate about global warming so frustrating. As you say, there are direct measurements that are consistent with our understanding of how CO2 influences outgoing longwavelength radiation. This, together with a lot of other measurements and observations, effectively confirms the role that CO2 is playing in warming our climate. However, people end up focussing on the mismatch, since 1998, of GCMs and measured surface temperature anomalies. Not only are they still consistent at the 5-10% level, we don’t need to use GCMs to confirm ongoing global warming.

  28. Thanks Tom, that seems fairly clear. I’m sure Shub will now be satisfied that the two time intervals are not identical and hence that there is nothing particularly suspicious about this visual similarity.

  29. dana1981 says:

    For the record, the 1895-1946 trend was 0.078°C per decade. The trend since 1957 is 0.136°C per decade (75% larger). Those numbers are using GISTEMP, via the SkS trend tool.

    The results from NCDC and HadCRUT4 are similar, though the ratios are a bit smaller (in large part to their coverage bias).

    Watts’ comparison is a joke. He asks you to eyeball the data when calculating the trends is a trivial task. He uses HadCRUT3, which has a substantial known cool bias due to poor Arctic coverage. And he ends the second graph in 2008, which was an anomalously cool year.

    Frankly it’s pretty pathetic. And it’s offensive that anyone would compare the joke that is WUWT to SkS.

  30. dana1981 says:

    “evidence for CO2′s causative role is circumstantial at best.”

    That’s called denial of the greenhouse effect. Go join the Dragon Slayers.

  31. dana1981 says:

    As I noted above, the second (after removing the endpoint cherrypicks) is 75% larger in GISTEMP, roughly 40% larger in HadCRUT4 and NCDC. And it’s trivial to check those trends. The fact that Watts didn’t do so reveals how little he thinks of his readers’ intelligence. And Shub’s comments illustrate why.

  32. Dishonesty in action:

    I wrote:

    “If that’s point you are making, you are admitting that evidence for CO2′s causative role is circumstantial at best.”

    This sentence says: if you say ‘X’, then, ‘Y’ must be true

    Dana excerpts the above to:

    “‘Y’ must be true”

    This changes the meaning of what I said, to its exact opposite.

  33. More dishonesty:

    “Given this, Shub’s claim that the causative role of CO2 is “circumstantial at best” merely indicates a massive disconnect between Shub’s mind and the actual physical evidence.”

    Look at the above paragraph. Do you believe it represents what I wrote correctly?

    When supporters of climate orthodoxy run out of places to hide, they start seeing what doesn’t exist.

  34. Seriously, you’re asking me to defend you when most of our discussions have involved you mis-representing what I’m trying to say. Good grief, do you have no self-awareness?

    Okay, let’s see. I made an analogy of someone falling down a flight of stairs and was trying to illustrate that if you only observed the fall, you wouldn’t necessarily know why it happened. Your response was to say

    If that’s point you are making, you are admitting that evidence for CO2′s causative role is circumstantial at best.

    My interpretation of what you were suggesting was that this “confirmed” the suggestion made in the WUWT post that if CO2 didn’t play a big role in the first time interval in Figure 1 above, it couldn’t have played a big role in the second time interval. Although, I’ll admit that I’m somewhat confused as to how you got there, given that I’m unlikely to be trying to illustrate that, the only other conclusion I could draw was that you were trying to do so.

    If this is what you were suggesting, I see nothing wrong with what Dana has said. If not, why don’t you clarify, because I certainly interpreted it in a manner similar to the way Dana has interpreted it. That’s what these comments are for. A chance to provide clarity.

  35. Tom Curtis says:

    Shub continues playing dishonest games. The phrase “you are admitting that x” is only used when the person using it believes that x (or at least wants to be taken as believing that x). Thus, Shub’s use of the phrase implies that he believes that “evidence for CO2’s causative role is circumstantial at best”. That is how he was understood by all here, and that is how he was responded to.

    Having now found himself in an untenable position, he seeks to extract himself by weaseling out on what he actually said. That could be accepted as legitimate if accompanied by an apology for his poor choice of words. Absent that, however, then his choice of words must be taken as revealing his purported beliefs; and his attempt to extract himself from the implications of those purported beliefs be seen as yet another dodgy rhetorical trick.

  36. Tom, yes I’d missed that semantic subtlety. Maybe Shub will be willing to come and clarify as I suggested earlier, but it is hard to see how he can accuse others of being dishonest when their interpretation seems consistent with what he appeared to be saying. Always willing to be convinced otherwise though.

  37. wotts
    Is it really that difficult to understand sentences and follow a line of argument? When you point solely to a temperature record and say CO2 caused this, the support you offer for your position is indeed circumstantial. If you find yourself in that position, that is what it reduces to.

    Like you say, the comments sections is for clarification. Were you in that position? To me, it appeared that you were. Or it could be described so, in an reductio ad absurdum – which is what I did, ’cause it is quite evident to me that you love your CO2 hypothesis very much.

    If two patterns of change are similar, it is quite clear, that they need not be necessarily caused by the same thing. I think we both agree on that. The temperature record by itself contains no information that will distinguish a CO2 caused one from a non-CO2 caused one.

    The corollary, unavoidable given the circumstances of the present state of the climate debate, is simply how two time-frames, each significantly different from one another with respect to the CO2 regimes, have such similar patterns of temperature rise. There is every reason for them to be different from one another, given the primacy accorded to CO2, but yet it is their similarity that is striking.

    How do we explain this?

    Your post attempted an answer. It used solar influence as a variable that could account. But then, it spawns further questions. Solar activity is in a relative lull in the second half of the 20th compared to the first. CO2 is really high in the second, but was just starting to climb in the first. Given the sameness of the temperature rise in the two halves, it would mean these two factors were so constituted they balanced out so well and produced a net effect that was the same between the two halves.

    However you might adjust your starting points and tweak it, the problem doesn’t go away. The rate of warming is very similar between the two halves which means CO2 cannot have had a significant effect on the rate of warming.

  38. Shub, I agreed with most of what you wrote until I got to the end. I don’t understand how you can draw your final conclusion. I’ve got to say that I had intended not to respond to any more of your comments because, in my opinion, some of what you write is insulting and offensive, but – given that you’ve actually written quite a decent and thoughtful comment here – I’ll have a go.

    Firstly, as Tom and Dana have both pointed out, the temperature variation in the two intervals are not the same. They look similar, but they differ quite substantially.

    Secondly, in the first half of the 20th century, solar forcing increased by about 0.5 Wm-2. CO2 concentrations increased from about 290ppm to about 310 ppm. Using Δ F = 5.35 ln(C/Co) gives a change in forcing of about 0.35 Wm-2. So, the net forcing in the first half of the 20th century was close to 1 W m-2 (maybe a bit less, but close).

    In the second half of the 20th century (1957-2008 as per the figure above) you can determine the CO2 forcing as about 1.2 Wm-2. Solar forcing didn’t change much or dropped slightly (as you’ve already commented) so you end up with two time intervals with a change in forcing that is similar. The latter time interval has a slightly higher change in forcing (by 20 to 40%) than the earlier time interval, but – as Dana and Tom have pointed out – the change in temperature was slightly greater in the latter time interval than in the former.

    So, yes they’ve looked similar but that’s entirely consistent with what we would expect given the change in forcing in the two time intervals. There’s nothing suspicious here and just because they look similar is not an argument for CO2 not playing a significant role in the warming. If anything it’s the reverse. How can they be similar if CO2 is not playing a significant role?

  39. KR says:

    The two trends are similar, but not identical, with recent warming being larger and faster than early 20th century.

    The attributions of those two warming trends, on the other hand, are completely different. Natural forcings have about a 70% contribution (anthropogenic forcings ~30%) towards early 20th century warming. But for the recent period, natural forcings have instead a negative contribution (cooling influence), while anthropogenic factors have a greater than 100% contribution. If natural forcings had remained steady we would have seen more warming than observed by 0.1-0.2C.

    The graph comparison illustrates both why the ‘eyecrometer’ is not a reliable instrument, and the causal basis for considering these two periods in different lights.

    “The rate of warming is very similar between the two halves which means CO2 cannot have had a significant effect on the rate of warming.”

    Rates are not attributions, that statement is therefore a non sequitur fallacy.

  40. dana1981 says:

    Actually the change in solar forcing probably wasn’t that large. Recent estimates have the TSI increase at about 0.6 W/m2 (1900-1950), but then you have to account for albedo and spherical geometry, so the solar forcing was only about 0.1 W/m2. But there was also the volcanic forcing, and of course natural variability, on top of the GHG forcing.

  41. Yes, I’d missed that subtlety. Does reduce the solar forcing in the first interval. That would seem to make CO2 more significant in the first interval.

  42. Daniel says:

    It’s also worth noting that the superficial similarity in the shape of the two graphs is due to Watts particular choice of 50 year intervals (1895-1946 and 1957-2008). There’s no special reason to choose these intervals except they look the same
    If we change the start year by only 10 years the illusion disappears
    Of course cherry picking two graphs and then complaining that you can’t determine causality from them is a pretty pathetic argument anyway.

  43. Absolutely. I was trying to explain essentially that point to Shub earlier in the comments on this post, but it didn’t really get through.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.