17 September 2013

Global Temperature Trends and the IPCC

As the excitement builds about the release forthcoming IPCC report (snore), debate is underway on how to interpret previous IPCC predictions for the evolution of global surface temperature trends. The debate has been super-charged by a recent article in The Daily Mail by David Rose, leading the usual suspects to say the usual things. Such debates involve exegeses of generally inscrutable IPCC statements filtered through the imperfect process of media (social and mainstream) reporting, colored by agendas.

In this post I pass on the exegeses and have a look at the actual numbers to address several questions and raise a few of my own.

How have the IPCC's out of sample predictions for the evolution of global average surface temperature fared against observations?

I first addressed this question in a running series of blog posts at Prometheus, back in the day. That exercise resulted in a correspondence published in Nature Climate Change in 2008 (here in PDF). Here is a quick update of that analysis.

The graph at the top of this post updates Figure 1a from Pielke (2008) through 2012. I show only the NASA GISS observational dataset (with data from the KNMI Climate Explorer), as it is the "warmest" of the four datasets.

The data shows clearly that the observations are running cooler than the out-of-sample predictions of the IPCC from each of its past 4 reports.

How much cooler?

If we simply compare rates of projected increase (1990 to 2012) the answers with respect to each previous IPCC report are:

  • 47% = NASA GISS linear trend slope as percentage of IPCC 1990
  • 91% = NASA GISS linear trend slope as percentage of IPCC 1995
  • 80% = NASA GISS linear trend slope as percentage of IPCC 2001
  • 80% = NASA GISS linear trend slope as percentage of IPCC 2007

With different rates of increase, the absolute difference in observations vs. projections will be a function of the period being looked at. Note that a comparison with the other 3 surface datasets, not shown here but which appear in Pielke (2008), would lead to larger discrepancies.

What quantitative conclusions does this exercise lead to?

1. The observations of global average surface warming are about half that predicted in the first IPCC report from 1990. Over the past 25 years, projections of rates of future surface temperature increase have clearly come down dramatically.
2. Subsequent IPCC reports reduced their projections, but global average temperature observations are still running lower than that projected in 1995, 2001 and 2007.

Are the lower observed temperatures significantly different than the projections?

Fortunately, there is a just-published peer-reviewed paper in Nature Climate Change which takes up this question, and concludes:
Recent observed global warming is significantly less than that simulated by climate models.
This won't be a surprise to anyone who has followed the ongoing, high-quality discussions of the subject by bloggers, such as Lucia Liljegren.

Does the inconsistency between observations and models have much significance for climate policy?
Not really. The fact that some enthusiasts have over-egged the climate prediction pudding does not take away from the core understandings of climate science, namely that humans influence the climate system, via greenhouse gas emissions and other means, and such influences carry with them some risks.

Of course, the over-egging has set the stage for the discrepancy to be of political significance, as the credibility of the IPCC and its champions is what is really under siege by its critics. Had the IPCC more faithfully represented uncertainties and had its public representatives been less strident and less arrogant, then the fact that we cannot actually predict the short-term evolution of the climate system would have been expected rather than treated as a scientific failure of some sort by the IPCC.

What should scientists do?

Back when I paid more attention to such things I offered the following as a (tongue-in-cheek) suggestion to those wanting to desperately prove that models were in fact consistent with the observations, which is offered without explanation.
More seriously, rather than engaging in proxy wars over media reporting and the short-term PR spin associated with it -- which may in fact just make things worse -- it would be in the long-term interests of the climate science community to take a step back and consider the role of their spokespeople (official or otherwise) in aiding and abetting the skeptics, deniers and other nefarious evil-doers.

A difficult question for the climate science community is, how is it that this broad community of researchers -- full of bright and thoughtful people -- allowed intolerant activists who make false claims to certainty to become the public face of the field? 

It is a question with continuing relevance.

65 comments:

  1. Hi Roger
    thanks for this insightful post.

    You wrote:
    "such influences carry with them some risks"
    This is the whole problem of course, this is all we can say right now and doesn't sound very alarming.
    I know you're not a big fan of the precautionary principle, but your phrase sounds like you are using it here.
    In your excellent book The Climate Fix you presented decarbonisation as a win-win situation. This was one of the few parts that I found hard to understand and I remember that in an email exchange I had Richard Tol on my side.

    So maybe you could elaborate on this a bit more in a future post.

    Marcel

    ReplyDelete
  2. Ah, the old Willis trick.
    ----------------------------
    The trick Eschenbach used was to use a single year for the baseline instead of the thirty year average that is normally used. Yes, it’s another version of the disingenous baseline game that produced all those bogus “global warming ended in 1998″ claims. Given the year to year variability of climate, by choosing the right year to use as a baseline you can manufacture almost any result you want.
    ----------------------------

    ReplyDelete
  3. -2-Eli Rabbit

    Thanks for the comment, but you are incorrect.

    All numbers in the graph at the top of this post and in Pielke (2008) reflect a 1980-1999 baseline.

    Thanks!

    ReplyDelete
  4. Roger -
    You posed the question "How have the IPCC's out of sample predictions for the evolution of global average surface temperature fared against observations?" But you answered it by taking trends since 1990, which includes for the later predictions, historical data. AR4, for example, should be considered to be out-of-sample only since 2000 or 2001. Over this shorter interval, the ratio of actual-to-predicted is less than in your table.

    Also, I think you have overstated the FAR's prediction. From memory, their mid-range business-as-usual prediction showed a temperature rise of about 0.2 K from 1990 to 2000 and about 0.3 K/decade thereafter, so 1990 to 2012 would be an increase of about 0.55 K, not 0.75K as in your diagram. If that's correct, then the FAR prediction isn't quite as far off.

    ReplyDelete
  5. -4-HowardW

    Thanks ... a few replies:

    1. I am comparing the linear rates of change, which are not sensitive to time scale of prediction.

    2. For FAR I am using 0.35 C/decade, taken from the IPCC 1990 WG1 report, as discussed here:
    http://cstpr.colorado.edu/prometheus/?p=4309

    Thanks!

    ReplyDelete
  6. The system is remarkably stable. However, it is not defined by a single number (i.e. "normal"). Our skill to forecast is limited, and our skill to predict is severely limited. Instead of preparing for catastrophic change, which cannot be reasonably accomplished, perhaps we should focus on something more reasonable. We should consider that science is distinguished from philosophy by virtue of it being constrained to a limited frame of reference. The reason being is that our awareness is limited; our access is limited; and our grasp is limited. This should be a lesson to people who claim an insight that exceeds a limited frame, especially when it is derived from inference, including: pattern matching from limited, circumstantial evidence; improper or insufficient characterization of a system; or an unwieldy composition.

    That said, the prognosticators of calamity may yet be proven right, but it will likely be an event that occurs in the intersection of disparate events, and will be of limited duration. Our efforts to mitigate risk should focus on likely events, which will vary in time and space.

    ReplyDelete
  7. Hi Roger,

    Glad you gave a shout-out to Lucia; she really has done a lot of very good work on models vs reality.

    For sure rising GHG's, and a host of other human caused changes, have and will continue to influence Earth's climate. But the size of that influence in the future matters in formulating sensible public policy, including the kinds of 'no-regrets' policies you have advocated in the past. I agree that the "over-egged" projections of warming have hurt the credibility of climate science, but I think you may draw too fine a distinction between 'intolerant activists' and climate scientists; often in public debates they seem to be the very same people. Worse than hurting the credibility of the field, those 'intolerant activists', have advocated public policies which are bound to be strongly opposed by many. They make good, sensible policies the enemy of the politically impossible policies they desire. The result is either no policy at all, or foolish policy (Solyndra, etc.) which only wastes money and has no possibility of making any difference.

    I hope you are right that there are a lot of 'bright and thoughtful' researchers in the field, but more than that, I hope those people will force a change the way that 'the science' has been used by intolerant activists, and the way it has been communicated to the public. So far, I see no evidence of that.

    I suspect that kind of change will have to await the departure of many of the leading lights in the field, which could take a long time. Of course, continued divergence between models and reality may hasten those departures.

    ReplyDelete
  8. If you compare the small amount of manmade CO2 in the atmosphere from 1910 to 1945 and the large temperature increase versus the huge amount of CO2 produced after 1998 versus the non-existant warming, one must conclude CO2 is irrelevant.

    ReplyDelete
  9. Ross McKitrick had trhis essay on related themes in the Financial Post:

    http://opinion.financialpost.com/2013/09/16/ipcc-models-getting-mushy/

    Some remarkable quotes from the IPCC AR5 draft there.

    ReplyDelete
  10. -5- Roger
    I looked up your reference -- the 1992 supplement is now available online here, by the way. Going back to the cited figure Ax.3, the rise from 1990-2012 is a little over 0.5 K. 0.35 K/decade may be OK as a century-long estimate, but because the curve accelerates, it overstates the slope at earlier times.

    ReplyDelete
  11. -10-HowardW

    Thanks much ... yes, and if you follow the discussion on Prometheus you'll see that my first efforts were criticized for assuming that the short-term curves were in fact meaningful, when the IPCC was not in the business of issuing decadal forecasts. So what you see there is indeed the long-term average trend.

    I'm not sure how one would derive a decadal forecast from the 1990 IPCC, but if one were developed and came out at 0.5 C, everything I've written here would still stand (and note that if that were the case, it would imply a much higher rate of warming from 2013-2100 than implied by the FAR in any case.)

    Thanks!

    ReplyDelete
  12. Ross McKintrick probably overstates the case a bit, but things are really not looking good for the modeling community or the IPCC. The more interesting question for me is if people in the field can bring themselves to change the projections of warming so that reality is at least included in the uncertainty range, or will they "go down in a blase of glory", insisting that catastrophic warming is virtually certain. I suspect they will find the path to glory tempting; they have been 'saving the world from catastrophe' for decades, and that paradigm may be the most difficult for practitioners in the field to give up.

    ReplyDelete
  13. Hi Roger,

    You brought up David Rose in your post.

    It is odd that you do not call out Rose on his disinformation and distortion The Daily Mail. Instead you simply link to his disinformation with no caveats whatsoever.

    Could you please unequivocally state whether or not you agree with the assertions made by Rose. If you like I could provide some specific examples.

    Thanks!

    ReplyDelete
  14. Hi Roger,

    You cited the recent commentary Fyfe et al. (2013). However, you neglected a key observation that they highlighted, namely (my bolding):

    "Recent observed global warming is significantly less than that simulated by climate models. This difference might be explained by some combination of errors in external forcing, model response and internal climate variability."

    Additionally, you also failed to inform your readers that the same models successfully simulated the rate of warming between 1900 and 2012.

    You are cherry picking Roger, and in doing so are misrepresenting Fyfe et al's research.

    ReplyDelete
  15. -13-Albatross

    You are welcome to visit and comment, but your repeated demands and accusations are past tiresome.

    I have no interest in playing games, sorry. Because you have repeatedly abused commenting privileges here, for a while I'll limit you to one post per day, please use it wisely, thanks!

    ReplyDelete
  16. Funny thing is, a few years ago, we (including Lucia) wrote that some of mismatch between models and observations during the past 10-15 years quite possibly rested with the climate models themselves. The reviewers were horrified and our paper was rejected. Now, 3 years later, the IPCC admits to the same. I guess it matter who says it:

    http://www.cato.org/blog/peer-reviewed-or-not-ipcc-accepts-our-conclusion

    -Chip

    ReplyDelete
  17. Roger,

    I believe temperatures are colder than the GISS data set you cite. Please see: http://stevengoddard.wordpress.com/data-tampering-at-ushcngiss/

    Something very similar happened in the transition from HADCRUT 3 to 4: The distant past got colder and the recent past got warmer.

    Mike

    ReplyDelete
  18. Chip #17,
    Reality is harsh; it makes people admit things they would otherwise never admit, and accept things they would never otherwise accept, no matter how distasteful. This is what is happening now in climate science.

    You and Lucia were technically correct, of course, and that was obvious to anyone who actually looked at the content rationally. But I don't think it is surprising your paper was rejected: the politics of climate science at the time pretty much demanded it.

    Take solace in the fact that you were right and the reviewers and editor in hindsight appear both unprincipled and factually wrong.

    ReplyDelete
  19. I struggle with the whole debate. The atmosphere is a massive open, non-linear system and we are treating it like we would a test tube. A particular point is that the magical "Climate sensitivity" is a constant !

    This clear mistake underlies the whole rotting mess that is climate change. By looking at behaviours in the tropics it's clear to see, as the temperature rises the negative feedbacks rise rapidly in the form of tropical storms. Climate sensitivity falls as temperature rises.

    Not that even the average can be correct, taking into account direct radiation to space and other known negative feedbacks, there must be positive feedbacks acting to both counteract the negative feedbacks, and then multiply the initial pertubation by a factor of 3 (IPCC). This implies the positive feedback sources amount to a gain of 15 or a loop gain of 0.95 or more. The physical mechanisms cannot be time correlated, they are vectors consisting of magnitude and lag. Simple scalar models cannot possibly reliably deliver a result in a system with multipe feedback mechanisms with temperature dependant, and otherwise chaotic magnitudes and lags. Such high positive feedbacks would lead to an oscillating climate over relatively short timeframes. Of course this doesn't happen

    ReplyDelete
  20. From this statistically-challenged person's perspective, it would seem that despite the best efforts of the IPCC/UNFCCC and its activist cheerleaders in the green NGO sector, as well as the MSM, the IPCC/UNFCCC can no longer ... uh ... hide the decline!

    This, perhaps combined with the aftermath of the election results in Australia - not to mention the increasing backlash against the foolhardy, premature and costly implementation of so-called "green" solutions (biofuels, wind, solar) - must be giving the powers that be in the IPCC/UNFCCC pantheons considerable "pause" for something or other. One would hope it would be "thought" for a change, but there are no guarantees!

    Nonetheless, this may well be the "reasoning" behind the recent resuscitation of the somewhat dormant Momentum for Change PR "initiative", as I had observed in a recent post [http://hro001.wordpress.com/2013/09/12/hiding-the-decline-of-the-ipccunfccc/]

    This may also be the reason that they have decided (somewhat late in the game, IMHO) to parachute into the lead authorship ranks of WGII, a philosopher with green credentials.

    One suspects he will be sufficiently infused with the tenets of the Precautionary Principle to colour the "findings" of WGII.

    They will be counting on time and distance from whatever the outcome of WG1-12 might be next week, so that whatever "the science" says, just won't matter any more. [She says, somewhat skeptically!]

    ReplyDelete
  21. Albatross -14- How can you say that the climate models successfully simulated the rate of warming between 1900 and 2012? The whole point of this post and other recent articles is that they have not successfully simulated the warming up to 2012. Since the late 90's they have diverged from reality by quite a bit. There may be reasonable explanations for that, but it is not true that they have "successfully simulated the rate of warming between 1900 and 2012".

    ReplyDelete
  22. -22 Mike-

    I've heard this sort of argument a lot. The basic idea seems to be that the models are giving us the signal and these decadal excursions are just natural noise. So it's basically an exercise in question begging.

    I think this argument is pretty brazen considering how chaotic and complex the climate system is. But that's climate science in the 21st century!

    ReplyDelete
  23. I remember around 1990 the prevailing wisdom was lets see what happens in 20 years and make a judgement then.

    ReplyDelete
  24. ===]]] Had the IPCC more faithfully represented uncertainties and had its public representatives been less strident and less arrogant, then the fact that we cannot actually predict the short-term evolution of the climate system would have been expected rather than treated as a scientific failure of some sort by the IPCC. [[[===

    Counterfactuals bring along with them, Roger, a high bar for evidence.

    Unless you have some divine insight (if so, please do tell), you might consider the uncertainties associated with determining what would have been had things been different.

    My guess is that the differences resulting from different actions on the part of the IPCC would have been far less categorical than what you describe. For example, we have seen how some "skeptics" have habitually distorted the uncertainties that climate scientists have described, for the purpose of rhetorical expediency. Maybe you should allow some room in your unqualified depiction of what would have been for the continuation of that quite evident pattern?

    But in reality, neither of know what would have been if things have been different, now do we? Perhaps you might consider rewriting your statement to allow for the uncertainties?

    ReplyDelete
  25. debate is underway on how to interpret previous IPCC predictions for the evolution of global surface temperature trends.

    The IPCC does not make predictions, it makes projections. Surely you understand the difference.

    ReplyDelete
  26. Roger Pielke Sr:

    "...my colleagues and I have shown that global average surface-temperature changes are not particularly useful for assessing the broad range of human influences on climate."

    http://davidappell.blogspot.com/2013/09/does-bjorn-lomborg-really-understand.html

    ReplyDelete
  27. "A difficult question for the climate science community is, how is it that this broad community of researchers -- full of bright and thoughtful people -- allowed intolerant activists who make false claims to certainty to become the public face of the field? "

    And that, my friend, is the right question.

    But, Roger, I think you perfectly well know the answer to that question.

    It fits a preconceived political agenda, provides a shield to hide behind in the hostile outside world, and is accepted because they believe in the "knowledge to policy" myth, i.e. (consensus) knowledge will almost automatically lead to policy. As long as scientists are sure enough about something, politicians will follow suit and install the preferred solution. Criticizing the consensus thus undermines the path toward the preferred solution, or so the thought is.

    Deficit model and all that ...

    ReplyDelete
  28. -26-David Appell

    "The IPCC does not make predictions, it makes projections. Surely you understand the difference."

    Sure I do. Though what I don't understand is why you think your comment is in any way relevant to this post ;-)

    -27-David Appell

    Roger Pielke Jr.

    "Does the inconsistency between observations and models have much significance for climate policy?

    Not really."

    Thanks!

    ReplyDelete
  29. Received for posting from Joanne Nova:

    "The quote from the first page of the Executive Summary of the Summary for Policy Makers, FAR 1990:1

    “Based on current model results, we predict:

    Under the IPCC Business-as-Usual (Scenario A) emissions of greenhouse gases, a rate of increase of global mean temperature during the next century of about 0.3C per decade (with an uncertainty range of 0.2°C – 0.5°C)” [IPCC FAR summary]

    For anyone who wants to argue the finer points of the 1990 IPCC predictions: Here's exactly what the IPCC predicted (http://joannenova.com.au/2012/12/the-ipcc-was-wrong-england-and-the-abc-mislead-australians/) and why we know they were wrong , here is my response to SkepticalScience's tricks (http://joannenova.com.au/2012/05/the-ipcc-1990-far-predictions-were-wrong/) to defend the IPCC. Here are the questions real journalists could ask (http://joannenova.com.au/2012/12/the-ipcc-was-wrong-england-and-the-abc-mislead-australians/) our "climate experts" about the 1990 report. Here are the major flaws in the Frame and Stone defense. (http://joannenova.com.au/2012/12/the-ipcc-was-not-right-frame-stone-ignore-main-ipcc-predictions/)

    There is no weasel-room here: even if emissions are stabilized at 1990 levels temperatures should rise by 0.2C per decade for the first few decades."

    ReplyDelete
  30. "The IPCC does not make predictions, it makes projections. Surely you understand the difference."

    David Appell (#26),

    Sure we understand the difference. A projection is a prediction with sufficient ambiguity that it can't be falsified. In other words, it's not science but useful for science-based rhetorical arguments.

    ReplyDelete
  31. So....Joanne Nova (#30) has a quote from the IPCC saying that "we predict...."

    And David Appell (#26) says that the IPCC doesn't make predictions.

    Hmmm....game and set to Joanne Nova. Does David need to improve his fact checking?

    Or perhaps the IPCC has, lately, given up on making predictions, given how badly they have fared?

    ReplyDelete
  32. Mike @22,

    You are mistaken. My comment was based on the caption for Fig. 2 in Fyfe et al. (my bolding),

    "...the horizontal red lines show the observed 1900–2012 trends averaged over 100 realizations. Black crosshatchings are the 95% uncertainty ranges for simulated 1900–2012 ensemble mean trends. Note that the observed and simulated long-term trends are very similar to one another, and are smaller than the short-term trends”

    From the paper's science brief at PCIC,
    "The authors also find that the same models successfully simulate the rate of warming over the 1900-2012 period."

    It is unimpressive that Roger, who presumably would have read Fyfe et al. before blogging about it, didn't correct you. Then again, he seems interested in only speaking out about problems (real or perceived) in comments made by people who have the temerity to challenge him.

    Roger,
    There are problems with your simplistic comparison between the rates of warming reported in the various ARs and the time series of observations. Tamino addressed this and came up with a more rigorous and fair comparison (note the ranges), see here.

    You claim "Had the IPCC more faithfully represented uncertainties...", putting aside that that is not an entirely true nor fare statement, your simplistic critique of the surface temperature trends allows for no uncertainties whatsoever.

    Just one final observation for today. So far you have linked to McKitrick's latest disinformation and apparently found no problems whatsoever, same with the disinformation of David Rose and that of Jo Nova. For good measure Pat Michael's side kick dropped by, and again not a single word from you challenging their claims. This is beginning to look like crowd sourcing exercise for Climate Depot.

    ReplyDelete
  33. Roger... your 'difficult question' is not really all that difficult to answer. Let us assume that the climate science community really is full of 'bright and thoughtful' individuals. (I have no doubt that it is.) Typically, bright and thoughtful people with a penchant for science, have less than an average interest in the irrational worlds of media and politics, but they still have a great interest in eating and providing for their families.

    The other group you mention in your question is 'intolerant activists who make false claims to certainty...'. I find this description fits very well with the behavior of politicians on any side of an issue.

    So why would any 'bright and thoughtful' scientist defer the position of spokesperson to the political in their ranks? Because they don't want the job, the political among them do and, most importantly, the political spokesperson is far better at acquiring funding from politicians! Far better!

    If my income was dependent on government funding, which, in turn, was dependent on the perception of public crisis, I would defer to the colleague who was most effective at communicating the crisis and the need for more study to the politicians. I would reason that my spokespersons claims were not out of the realm of possibility, and that my work was important enough to be funded, even if a little exaggeration was required to make it happen.

    Given that AGW science is almost entirely funded by government and entirely dependent on the perception of crisis, it is entirely natural for the AGW spokespeople to be political (intolerant activists making false claims to certainty).

    Honestly, Roger, could you imagine this unfolding any other way? Do you really think a purely scientific voice, complete with caution and uncertainty, would be tolerated by the AGW science community, resulting in a significant percentage of them being out work? How have such deservedly cautious scientists, who even flirt with publicity, been treated by their peers?

    ReplyDelete
  34. Albatross@33

    So, over time, the models have been tweaked with unobserved fudge factors and speculation to be more in line with observations, which have also been tweaked (for whatever reason) and are now more in line with the models. Consequently, you can now argue that the models are showing some 'skill', although it is not a strong argument when considering all that tweaking!

    Still, that has nothing to do with the fact that the models have shown an increasing lack of skill over the last 15 years, and are, therefore, increasingly useless and increasingly wrong, which I believe is the point of this posting and Mike@22.

    ReplyDelete
  35. -33-Albatross

    Thanks for the comment ... with respect to the data presented in this post, it is consistent with that presented in the draft of the upcoming IPCC, see the relevant figure here:

    http://wpmedia.opinion.financialpost.com/2013/09/0916graphic.jpg?w=620&h=507

    Thanks!

    ReplyDelete
  36. Albatross -33- I guess it's all in what you consider to be successful. I agree with Jim Clarke (35) that the past agreement depends on a lot of tweaking (for example, effects of aerosols). That's why predictions of the future are so important; you find out whether your tweaks continue to work or not.

    ReplyDelete
  37. How can you compare an observed dataset showing more than 0.2 degrees of annual variability and intra-annual variability of even more than that to linear trends showing a rise of 0.4 degrees over 22 years (~0.02 degrees/year). It does not appear that there is enough data to say whether the models were right or wrong based on what you've shown in this entry.

    ReplyDelete
  38. -38-Jacob B

    Thanks ... I'm pretty sure that the words "right or wrong" don't appear in this post.

    For a rigorous treatment of model-observation differences, including metrics of uncertainty, please see:

    http://www.nature.com/nclimate/journal/v3/n9/full/nclimate1972.html

    Thanks!

    ReplyDelete
  39. wrt that wpmedia graph, it mirrors an incorrect figure in an early IPCC draft which made the classic willis mistake, e.g. starting all the trends at a relatively hot year 1990. See hotwhopperfor details, as to your cherry pick, remember how important it is for your argument to include 1998 (the big El Nino year.)

    ReplyDelete
  40. Roger @36,

    No, thank you! Specifically, thank you for going on the record that you accept the version that McKitrick used in his ongoing disinformation campaign in the media.

    As for your graphic, no is not consistent with the IPCC figure. I'll say it again, that graphic has time-varying ranges for the observations and model output. Those uncertainty data are key, and your simplistic graphic with fixed rates of warming does not include any ranges. A big difference. However, I know you well enough by now to expect you to not concede that as being a problem ;)


    Roger's followers,

    I encourage readers here to read Reichler and Kim (2008) on the continuing improvement in climate model performance. Their model performance index (taken as the mean over all climate variables and with smaller scores indicating greater skill) has decreased from 3.5 in 1995 to 0.8 in circa 2013 (the latter number is from the latest issue of Nature).

    The fake skeptics favourite past time of "attacking" the models (when they are not attacking climate scientists) is a red herring, and is now being recycled again before the release of AR5. There was concern about AGW long before the first numerical climate model was developed. Additionally, scientists have identified multiple independent lines of evidence that point to a climate sensitivity close to +3 C for a doubling of CO2 and we could easily increase CO2 levels fourfold over their preindustrial values.

    Regardless, contrary to claims being made here (and presumably endorsed Roger Pielke Jnr because he is reticent to correct fake skeptics now matter how egregious their errors) their are some concrete reasons for the models improving, some examples:

    Smaller horizontal grid spacing and more vertical levels
    Improved microphysics (something Curry has worked on, see Morrison et al. 2005, JAS)
    Addition of atmospheric chemistry and aerosol processes
    Improved representation of surface processes
    Fully coupled ocean-atmosphere models
    Ensemble techniques

    But despite these facts, fake skeptics have repeatedly demonstrated that they will believe what they want to believe, even in the face of facts that run counter to their beliefs.

    ReplyDelete
  41. Thanks for this post, and for this bit in particular:

    "Instead of preparing for catastrophic change, which cannot be reasonably accomplished, perhaps we should focus on something more reasonable."

    In my work in NYC, I frequently must deal with projections that ASSUME the Rapid Ice Melt scenario will come to pass by 2050. When I point out that such an unlikely event would be a game-changing catastrophe, to say the least, and is not relevant to our more limited planning efforts, I am ignored. The planning continues, and as a result, accomplishes little of value to mitigate the already awful state of preparedness for flooding.

    ReplyDelete
  42. -41 Albatross-,

    RK2008 looks like an interesting paper, but doesn't seem to say anything about the models' predictive skills. They only seem to be looking at how well the models match the past (1979-1999). OK, I guess the earliest models got a little bit of prediction in there.

    Also, the aggregated statistics mean that crappy temperature prediction is not inconsistent (ha!) with an improved fit of the aggregate score. I mean, I guess it's good that they're getting better at simulating different aspects of wind, but then you shouldn't use this score as evidence that they can tell temperature changes due to CO2 from a hole in the ground.

    I'm not sure who you mean by "fake skeptic," but the term seems to fit one who trusts that the models are predicting much of anything in the face of such evidently poor predictive skill.

    ReplyDelete
  43. Albatross@41

    Over 500 years ago, astronomers had a model of the solar system that included epicycles and the Earth at the center. On your 'model performance index' this model would have scored very close to zero or nearly perfect, yet the model was fundamentally wrong in its basic assumptions!

    Climate models suffer the same defect as the Earth Centered Solar System model; they are fundamentally wrong in their basic assumptions. (On top of that, they are not nearly as accurate as the Earth Centered Solar System model.)

    I recognize that there have been improvements in such things as grid size, crunching power, atmospheric chemistry, ocean-atmospheric coupling, and so on, but all of that 'improvement' cannot change the fact that the fundamental assumptions are wrong. All the improvements do is generate a much more sophisticated and expensive WRONG answers.

    For your edification, here are the assumptions and why they are quite obviously incorrect: 1.) Total feedbacks to increasing CO2 are significantly positive, and 2.) Natural variability is almost entirely limited to solar irradiance and volcanoes. I can say that both of these assumptions are false, because there is no evidence for either one of them in the historic climate record. There has never been runaway warming, even when CO2 was much higher,suggesting the feedbacks are at most, weakly positive and most likely negative, proving the first assumption wrong. The last 15 years alone prove the second assumption is wrong, as there has been no warming with increasing CO2, while volcanoes and irradiance have stayed fairly constant. There must be 'something else'! Really, all of climate history proves the second assumption wrong. There is clearly a lot of climate chance in the historical record that is not tied to volcanoes, irradiance or atmospheric CO2. The assumptions are wrong, so the models continue to get better at doing the wrong thing.

    Also, Reichler and Kim (2008) give the models a high score by 'taking the mean of all climate variables'. Does anyone really care about the mean of 'all climate variables', most of which are probably 'gimmies' and have little to do with how hot the models say it is going to get? No. We only care about how hot it is going to get, and on that, the models are looking pretty awful!

    Finally, I am a skeptic of an AGW crisis. Why would you insist that I, or anyone else, is a 'fake' skeptic. Is that so you can discard our words without thought or consideration? What are you so afraid of?



    ReplyDelete
  44. +1 for @JimClarke

    Roger, I am still bothered by the assumptions of the basic physics of greenhouse gases and climate models.

    As the earth rotates, parts that are in the sun in the daytime are in space at night. Anything that warms more in the sun cools more at night. Therefore, no changes in material properties will change mean temperature of anything subject to radiative heating and cooling. That means GHGs and anything that changes albedo (ice coverage, etc) will not alter mean temperature of the heating/cooling cycle. What goes up must come down - everything that enhances radiative heating also enhances radiative cooling. Maxwell's demon does not exist. One can easily prove or disprove my statements with simple physics experiments. I have not found anyone has done any.

    As far as I can figure, there are two things which keep the earth warm. 1-a molten core as internal heat source, 2-clouds which reflect outgoing heat from the core.

    ReplyDelete
  45. wrt that wpmedia graph, it mirrors an incorrect figure in an early IPCC draft which made the classic willis mistake, e.g. starting all the trends at a relatively hot year 1990

    Roger discusses only trends -- which are unaffected by starting point. He never talks about the absolute difference from projection to data.

    It is an annoying tactic to point out irrelevant "mistakes" in order to obscure the point in issue. Please stick to whether the projected trends of the IPCC are correct or not.

    ReplyDelete
  46. -46-Mark

    "It is an annoying tactic to point out irrelevant "mistakes" in order to obscure the point in issue."

    Thanks for pointing this out;-)

    ReplyDelete
  47. So, short-term failures don't invalidate the models. What would?

    Crickets.

    ReplyDelete
  48. I will second the comment by Mark (#46). Or third, I guess, since Roger already seconded. I would also like to point out that Albatross (#41) is similarly off-topic. I will part ways with Jim Clarke and say that I have no doubt that the models are getting better. But that, by itself, says nothing about whether they are yet good enough. If you want us to make multi-trillion-dollar investments based on computer models, they'd better be pretty damn good!

    ReplyDelete
  49. Mike@49

    I agree with you that the models are getting better, but they will never be 'good enough' until the basic assumptions about CO2 are altered to fit reality. The models are like a paraplegic with a broken arm. In time, the arm will heal and may become very strong and able. At which point, we can say that the paraplegic has become a lot better, but he is still no closer to walking. There are aspects of the models that are continually getting better, but the debate is about the models ability to accurately predict future global temperatures. In this regard, they are permanently disabled by the CO2 assumptions, and no amount of strength in other areas will correct that inherent disability.

    I hope this makes my point a little clearer.

    ReplyDelete
  50. Jim -50- I think that's too strong. I think the problem is mainly a blind spot that causes people to be more sure of themselves than they should about certain things.

    ReplyDelete
  51. Jim Clarke said:
    I agree with you that the models are getting better, but they will never be 'good enough' until the basic assumptions about CO2 are altered to fit reality.

    WHAT assumptions?

    From what I can tell, CO2's effect is calculated based on its absorption spectrum -- an easy measurement.

    ReplyDelete
  52. J. Clarke wrote:
    For your edification, here are the assumptions and why they are quite obviously incorrect: 1.) Total feedbacks to increasing CO2 are significantly positive, and

    THis isn't an assumption, it is a calcultion -- in agreement with paleoclimate observations.

    2.) Natural variability is almost entirely limited to solar irradiance and volcanoes.

    While it does seem that the big models have trouble reproducing short-term variability... how does natural variability alter temperature in the long-term?

    ReplyDelete
  53. 35. Jim Clarke said...
    So, over time, the models have been tweaked with unobserved fudge factors and speculation to be more in line with observations

    Such as what? Specifically? What physics equations, what lines in the code, where?

    I honestly would like to know the answer, not just vagure notes about what you think modelers are doing.

    ReplyDelete
  54. Roger et al.,

    It has been, shall we say, intriguing watching the back-and-forth banter ;) Roger's latest post rails against the term "consistent with" calling it a canard.

    I'm afraid that you guys mistaken and confused about McKitrick's outdated figure (that figure was "leaked" almost a year ago now). Let me revisit the timeline.

    1) I noted that Roger's simplistic attempt to compare the model output with global temperatures falls short (please don't ask me explain why again, but it has to do with including representing the ranges). This met with evasion by Roger.

    2) At #36 Roger attempts to retort by linking to the outdated (and incorrect) IPCC graphic used by McKitrick in his ongoing disinformation campaign and vendetta against climate science. Roger claims that:

    "...with respect to the data presented in this post, it is consistent with that presented in the draft of the upcoming IPCC."

    No, the data in Roger's post are not consistent with that outdated draft figure-- Roger's simple figure has its own issues. The outdated draft figure has a noteworthy problem that is being exploited by McKitrick to misinform. Yet Roger is happy to state that his simple representation of the data in his post are consistent with the incorrect (outdated) IPCC figure used by McKitrick to mislead.

    3) Eli tries to help again @40 by noting that the outdated figure Roger linked to @36 has a very real problem that does affect the results and/or the interpretation of the figure.

    4) Roger's supporters are having none of that factual stuff. Mark jumps to Roger's defence claiming that noting the problem with the incorrect draft version of the graphic is just an attempt by Eli to obscure the issue by point out "irrelevant "mistakes". Incorrect.

    5) Roger @47 is then quick to thank Mike for pointing out this. Dear me, what a tangled web you have woven.

    Folks, it is Roger who trotted out (i.e., linked to) that outdated and erroneous graphic used by McKitrick to misinform. It is Roger who stated that his graphic "is consistent with" the the erroneous figure. The errors in the graphic are not irrelevant, yet Roger was keen to associate/link his figure to that troubled figure, going so far as to claim that they were consistent with each other.

    Roger declares on his own blog that using the term "consistent with" is a canard.
    So at the same time Roger is eager to associate himself with the figure used by McKitrick to misinform, by saying his data "are consistent with" Roger is in fact (by his own admission) claiming that doing so is a canard.

    He really seems to be tying himself in knots in his attempt to obfuscate and have a self-serving dig at climate science (whilst also providing some fodder for the fake skeptics). The reality is that his whole sad endeavour by Roger is a canard.

    I do not expect subsequent posts to fix Roger's tangled web, it will likely just be more of the same canards.

    Oh, and thanks Roger!

    PS: Mike and Jim, Prof. Pierrehumbert's book "Principles of Planetary Climate" will address many of your questions and misunderstandings.

    ReplyDelete
  55. David Appell - 52

    Yes, it is easy to calculate the impact of doubling atmospheric CO2, all else being equal. Science tells us it will be near 1 C. That is the 'settled' science. That is also not a crisis, but largely a beneficial warming.

    David Appell - 53

    "1.) Total feedbacks to increasing CO2 are significantly positive, and

    THis isn't an assumption, it is a calcultion -- in agreement with paleoclimate observations."

    David, the calculation ASSUMES that the observed 20th Century warming was almost entirely due to increasing CO2. The assumption is built into the calculation. Calling it a scientific 'calculation' does not rid us of the assumption. The last 15 years are rapidly invalidating the assumption/calculation.

    "While it does seem that the big models have trouble reproducing short-term variability... how does natural variability alter temperature in the long-term?"

    What and excellent and important question; absolutely vital in answering what impact humans are having...don't you think? Unfortunately, those who try to study it are underfunded and disenfranchised by the AGW climate community, because their results continually show that half or more of the 20th Century warming was likely natural, which reduces climate sensitivity to CO2 and dramatically reduces the threat of a climate crisis. More and more recent papers are conceding this.

    As to the fudge factors and speculations: Aerosols. Peer reviewed science on aerosols is all over the place. Take your (cherry) pick. We are still debating the impacts of aerosols, and we have very little information on the composition and distribution of them in the middle 20th Century. In order for the models to appear to have skill, modelers gave aerosols very specific values and impacts, which is speculation at best. One thing we do know about those aerosols is that they were almost entirely in the Northern Hemisphere and cannot explain the Southern Hemisphere cooling that took place at the exact same time. Did the thing that cooled the Southern Hemisphere also cool the North, reducing the arbitrary impact modelers have assigned aerosols? That would be an assumption with more supporting evidence than the ones the IPCC adopts.

    Your request that I produce the lines of code is pathetic. Do you really need such strawmen to defend your position?

    ReplyDelete
  56. Albatross-55

    Do you really think that anything you just wrote would convince anyone that the models are not diverging from reality? Do you really think that the addition of error bars (regions) would make us think that the models are more skillful than we currently do? Do you really think that your twisting semantics, ad hominem attacks and condescending attitude will sway anyone to your point of view?

    ReplyDelete
  57. 41-Albatross:

    "Additionally, scientists have identified multiple independent lines of evidence that point to a climate sensitivity close to +3 C for a doubling of CO2 and we could easily increase CO2 levels fourfold over their preindustrial values."

    If CO2 levels were four times pre-industrial values, that would be ~1120 ppm. The current value is ~400 ppm. That's a 720 ppm difference. Over the decade from 2003 to 2012, CO2 increased by about 2.1 ppm per year.

    And the current increase in temperature above pre-industrial values is about 1 degree Celsius. So the world is still ~5 degrees away from a 6 degree increase. Currently, temperature is rising by less than 0.2 degrees Celsius per decade (less than 2 degrees Celsius per century).

    My question is: when do you think atmospheric CO2 levels will hit 1120 ppm, and when do you think the temperature increase will hit 6 degrees Celsius above pre-industrial levels? Not this century, right? How far into the 22nd century do you think this will happen? Or do you not expect 1120 ppm and 6 degrees Celsius above pre-industrial levels to be reached until the 23rd century?

    Do you really think that humans of the 22nd (or even 23rd!) century will not be able to keep the CO2 level at whatever level--perhaps 350 ppm--that they consider to be optimal?

    ReplyDelete
  58. Has there been thought/study/etc. given to the possibility of what, if any, role that regulations, emission reductions, etc. over the past 20 years may have contributed to the decreased projections? Or is that unlikely to have been a factor here?

    ReplyDelete
  59. "Has there been thought/study/etc. given to the possibility of what, if any, role that regulations, emission reductions, etc. over the past 20 years may have contributed to the decreased projections? Or is that unlikely to have been a factor here?"

    That's unlikely to have been a factor. Here are the the amounts of global CO2 emissions growth for various decades:

    1960-1970 = 58 percent

    1970-1980 = 31 percent

    1980-1990 = 15 percent

    1990-2000 = 10 percent

    2000-2010 = 33 percent

    So decadal CO2 emission growth rates had been slowing to the year 2000, but accelerated dramatically in the decade from 2000 to 2010. Much of that acceleration was from growth in emissions of emerging economies...China in particular.

    ReplyDelete
  60. Hi MIke, Mark, Jim,

    You should really be directing most of your questions and arguments at Roger because I doubt very much that Roger shares some of your misguided opinions and beliefs. Roger might also have more success addressing some of your misconceptions. Besides are you not interested/curious to note what his position is regarding your arguments?

    FWIW, here is a link to the Representative Concentration Pathways framework.

    ReplyDelete
  61. Hi "Albatross",

    You write, "You should really be directing most of your questions and arguments at Roger because I doubt very much that Roger shares some of your misguided opinions and beliefs."

    Let me first clear up a misguided opinion or belief that you have. That opinon or belief is that I have any beliefs. I don't. I do have opinions, but they change if new data warrants that they change.

    "Roger might also have more success addressing some of your misconceptions."

    You continue, "Besides are you not interested/curious to note what his position is regarding your arguments?"

    No, not particularly. You were the one who made the assertions in comment #41 that:

    "Additionally, scientists have identified multiple independent lines of evidence that point to a climate sensitivity close to +3 C for a doubling of CO2 and we could easily increase CO2 levels fourfold over their preindustrial values."

    I asked you several questions in comment #58 about that statement. Why can't you or won't you answer my questions? It doesn't seem to me that they are particularly difficult, since they are simply asking what your opinions (or beliefs) are.

    ReplyDelete
  62. Albatross - You don't seem to pay any attention at all to what people say to you, so it is pointless to spend any more time on replies.

    ReplyDelete
  63. Guys,

    An important point of clarification. Unlike you I have provided links to support what I have said. What I have written about equilibrium climate sensitivity or future CO2 levels is factual and from the scientific literature. I provided a link to the RCPs and a text. It has had nothing to do with beliefs.

    Here is a paper written by Kiehl in the journal Science in 2011 titled "Lessons from Earth's past". There is also a lot more where that came from in the peer-reviewed scientific literature.

    It is disturbing that Roger (a self proclaimed "honest broker") is quite happy to be complicit in the distribution of misinformation on his blog (e.g., by Nova and McKitrick et al.) and to permit misinformation by fake skeptics on his blog to go unchallenged. Not to mention the times Roger actively disseminates misinformation and half truths on his blog.

    So much for scientific integrity and the pursuit of truth Roger ;)

    Thanks!

    ReplyDelete
  64. "Albatross"

    I have twice asked you a simple questions regarding assertions you have made. You wrote, ""Additionally, scientists have identified multiple independent lines of evidence that point to a climate sensitivity close to +3 C for a doubling of CO2 and we could easily increase CO2 levels fourfold over their preindustrial values."

    I asked, "...when do you think atmospheric CO2 levels will hit 1120 ppm, and when do you think the temperature increase will hit 6 degrees Celsius above pre-industrial levels? Not this century, right? How far into the 22nd century do you think this will happen?"

    What are your answers to those simple questions?

    ReplyDelete