59 thoughts on “AR5 WG1: Final Draft”

1. Lucia. the chapter 9 released today is identical to the June 7, 2013 version on which we were commenting/ It bears the identical header and footer and pagination.

2. “The total anthropogenic ERF over the industrial era is 2.3 (1.1 to 3.3) W m–2.3 It is certain that the total anthropogenic ERF is positive. Total anthropogenic ERF has increased more rapidly since 1970 than during prior decades. The total anthropogenic ERF estimate for 2011 is 44% higher compared to the AR4 RF estimate for the year 2005 due to reductions in estimated forcing due to aerosols but also to continued growth in greenhouse gas RF. [8.5.1; Figure 8.15; Figure 8.16]” Ch 8

   This blows a hole in some of the speculation about the hiatus. ERF is effective radiative forcing. This indicates, as it is a total, the effective forcing should have been higher. Note with a lowering of aerosol forcing and aerosol concentration, the temperature needs to accelerate even quicker at some point.

   “”The net forcing by WMGHGs other than CO2 shows a small increase since the AR4 estimate for the year 2005. A small growth in the CH4 concentration has increased its RF by 2% to an AR5 value of 0.48 (0.43 to 0.53) W m–2. RF of N2O has increased by 6% since AR4 and is now 0.17 (0.15 to 0.19) W m–2.””

   “The magnitude of the aerosol forcing is reduced since AR4. The RF due to aerosol-radiation-interactions, sometimes referred to as direct aerosol effect, is given a best estimate of –0.35 (–0.85 to +0.15) W m–2, and black carbon (BC) on snow and ice is 0.04 (0.02 to 0.09) W m–2.” Ch 8

   Ends some speculation as well.

3. Steve– Interesting. I thought June 7, 2013 was supposedly “leaked”! People are claiming ‘official final’ now.

4. I don’t follow why “it is more likely than not that internal climate variability in the near-term will enhance and not counteract the surface warming expected to arise from the increasing anthropogenic forcing”.

   Is there a definition of “near term”? If this doesn’t mean “through the next El Ni~o”, what is the basis for predicting warming due to internal variability??

5. SteveMcIntyre–
   There seems to also be a document listing changes. I saw no changes for chapter 9.

6. Why can’t they at least admit that SOME of the models were were wrong?

   Those models shouldn’t be simply “scaled down”, they should be eliminated.

7. Ray, Sadly, while some scaled down projections are shown, it appears that the ‘not scaled down’ ones are shown also. It’s a bit like having ones cake and eating it too. Some will wave in alarm at the ‘not scaled down ones’ as reason to fear and simultaneously say we can’t discount them because the observations still fall in the lower range of the ‘scaled down ones’. Just wait and see! 🙂

8. Re: BillC (Sep 30 11:32),

   I don’t follow why “it is more likely than not that internal climate variability in the near-term will enhance and not counteract the surface warming expected to arise from the increasing anthropogenic forcing”.

   I believe that falls into the category of SWAG. But the NAO is already headed for negative territory and the AMO index is due to start decreasing in the next decade. I would still say it’s a good bet that the trend for the next 15 years will be higher than the trend for the last 15 years, but it’s not a sure thing. I would need fairly high odds to take the other side of that bet.

9. Lucia, have you read the account of the plenary session here:
   http://www.iisd.ca/vol12/enb12581e.html. I’m intrigued by the assurance to the plenary by CLAs that the discrepancy between models and observations was “not statistically significant” and therefore should not be reported in the SPM. I plan to write on this further, but it’s right in your wheelhouse.

10. Lucia … Note: I do expect warming trends to increase. But predicting that the trends will turn up is not the same as predicting they will catch up to the AR5 trends (and note: the IPCC isn’t even necessarily predicting they will catch up to AR5 trends.)

    .
    That is a very key observation. However, if the trend turns up, and if it remains up, then at some point in the future, observable GMT will cross the AR5 modeling’s near-term lower boundary — just shifted rightward in time X number of years — at which point the IPCC will probably announce that the hiatus questions from AR5 have been resolved and that there is higher confidence in the IPCC’s long-term predictions.
    .
    Suppose the warming trend resumes before AR6, but the slope of its upward trend is less than what one would expect if one had simply shifted AR5’s predicted trend rightward six years. Is that possibility something that should be discussed in the near-term evaluations of AR5, or is it something that is a bridge best crossed only when we come to it six years from now?

11. http://www.iisd.ca/vol12/enb12581e.html

    On the explanation of the observed reduction in the surface warming trend over the period 1998-2012, Saudi Arabia strongly urged incorporating language from the Technical Summary on models overestimating the warming trend. The CLAs advised against including this statement in the SPM, noting that: the research is currently inconclusive; overestimation of the models is too small to explain the overall effect and not statistically significant;

    Bunk. And double bunk. The models are outside if we use 25 years. So, even ‘they’ think 15 years to short, why not just use 25?

    Anyway, I love the negotiations over the text

    On Thursday morning, Germany and the UK said that their objections were not noted the previous evening when a sentence on overestimates in some models introduced by Saudi Arabia was adopted. Saudi Arabia, supported by Sudan, expressed grave concerns in opening up agreed text, emphasizing that “we are in dangerous waters,” while Sudan added that opening up agreed text raises the issue of equal treatment of countries. No changes were made to the text.

    and it is difficult to pinpoint the role of changes in radiative forcing in causing the reduced warming trend, with Co-Chair Stocker referring to this issue as an “emerging science topic.”

    I think “emerging science topic” translates into “results we are reluctant to acknowledge”.

12. Beta–

    …is it something that is a bridge best crossed only when we come to it six years from now?

    You sure seem to like suggesting we discuss hypothetical questions that about what might be worth discussing if something very specific that has not yet happened and may never happen, eventually happens. I don’t find those questions interesting, but perhaps you’ll find someone somewhere who thinks it’s worth thinking about and discussing.

13. Stocker used the phrase “emerging science” again in his interview with the Christian Science Monitor. It is reprehensible that they relegate the reconciliation of models and observations to “emerging science”. The assessment of models is surely among the most important IPCC duties, if not the most important.

14. The language in the Technical Summary was:

    Most simulations of the historical period do not reproduce the observed reduction in global-mean surface warming trend over the last 10–15 years (see Box TS.3). There is medium confidence that the trend difference between models and observations during 1998–2012 is to a substantial degree caused by internal variability, with possible contributions from forcing inadequacies in models and some models overestimating the response to increasing greenhouse-gas forcing. Most, though not all, models overestimate the observed warming trend in the tropical troposphere over the last 30 years, and tend to underestimate the long-term lower-stratospheric cooling trend. {9.4.1; Box 9.2}

15. Re: Steve McIntyre (Sep 30 13:37),

    Most, though not all, models overestimate the observed warming trend in the tropical troposphere over the last 30 years, and tend to underestimate the long-term lower-stratospheric cooling trend. {9.4.1; Box 9.2}

    Well, that would certainly fit the hypothesis that aerosols are overestimated to allow higher climate sensitivity, I think. Of course that presumes that some aerosols other than volcanic make it into the stratosphere, leading to warming. There are other possibilities: underestimating the stratospheric CO2 concentration or not getting stratospheric water vapor right. I forget which way that goes.

16. Cross-posted from Judith Curry’s blog:
    Question for Nic Lewis (or anyone else competent): Using Bayesian statistics, how much does each additional single month of flat temperatures bring down estimates of climate sensitivity? How much CO2 reduction does it equal, in terms of the same temperature non-rise?

17. The Stratospheric Aerosol Optical Thickness does very little to support the IPCC’s position

    http://www.columbia.edu/~mhs119/StratAer/

    Note the scale in the 2001-present plot.
    The last decade has had lower aerosols that the preceding decades.

18. If the CMIP5 models were run with forcing estimates based primarily on AR4 (plus anticipated GHG concentration increases), and the estimate of negative aerosol forcing has been reduced since then, wouldn’t that increase the discrepancy between models and observations? Or did the scenarios reduce the aerosol forcing and are consistent with AR5?

19. Beta Blocker,
    “Suppose the warming trend resumes before AR6,”
    .
    You may suppose too much. I doubt there will be an AR6. The process has long since outlived its usefulness.

20. I wouldn’t be surprised if all chapters date to June 7. They issued a PDF of corrections dated Sept. 27 which are going to cause changes to various chapters to make the science match the SPM. They claim the changes will be minor. Take with a grain of salt and all that I expect.

    The top link on this page contains corrections to be made.
    http://www.climatechange2013.org/report/review-drafts/

21. HaroldW your comment is supported in Ch 8 as far as I can tell. It is a head ache reading though. The subject is apparently so delicate direct statements are avoided.

    j nospace pittman at scdotrrdotcom if you want those garlics. I have some I have to do something with.

22. These are the chapters where changes will be made.
    TS, 1, 2, 3, 4, 5, 6 … 11, 12, 13, 14

23. The final draft of the IPCC report reflects political, not scientific, considerations.
    This report is a sales tool, a prop to support AGW theater.

24. John F. Pittman (#119914)-
    From 8.3.4.3: “This yields an ERF due to aerosol-cloud interaction estimate of –0.45 W m–2 which is much smaller in magnitude than the –1.4 W m–2 median forcing value of the models summarized in Chapter 7, Figure 7.19 and is also smaller in magnitude than the AR4 estimates of –0.7 W m–2 for RF due to aerosol-cloud interaction.” I don’t see that they actually answered my question directly, but it seems an inescapable conclusion from the above statement.

    P.S. Have emailed re garlic.

25. Re: SteveF (Sep 30 18:01),

    Beta Blocker: “Suppose the warming trend resumes before AR6,”

    SteveF: You may suppose too much. I doubt there will be an AR6. The process has long since outlived its usefulness.

    .
    Of course there will be an AR6. And an AR7, an AR8, and … an AR(n).

    Sooner or later, a warming trend of some kind will resume, the only question about that is when and by how much.

    Lucia thinks it may be sooner, based on increasing GHG concentrations, etc.

    I myself speculate that it may be later, with a possible short-lived downward trend in between, based on the 300-year historical pattern of Central England Temperature.

    In any event, the IPCC will hang tough until GMST clearly resumes its upward trend — even if it takes 20 to 30 years before that happens — all the while pursuing a strategy of actively managing the public discourse over climate change with the goal of discounting the importance of the pause.

    Hundreds of thousands of people, possibly millions of people, owe their future livelihoods to promoting the issue of climate change. They will not give up without a fight.

26. The tyranny of the consensus allows skeptics to be reduced to Merchants of Doubt and challenge from within the climate community to be minimized as “emerging science”.

27. I wonder when they will hear the “child” say “he has no clothes!”
    (ie the models are wrong because they do not match the evidence.)

28. Beta

    Lucia thinks it may be sooner, based on increasing GHG concentrations, etc.

    I myself speculate that it may be later,

    Thanks for tell us what we all think. You think I think it’s sooner than what? And you think it’s later than what?

    In any event, the IPCC will hang tough until GMST clearly resumes its upward trend — even if it takes 20 to 30 years before that happens

    Your prediction… right? Hard to falsify your prediction because “hang tough” is not well defined. Also, the threshold for “upward trend” is not defined. And in any case, whatever the definition of “upward trend”, it might resume pretty soon. Or not.

29. DeWitt,

    They didn’t say that natural variability will be less of a cooling influence…they said it will be a warming influence, all by itself. In other words, not a return to a positive global trend counting anthropogenic + natural, but a trend along the lines of 1975-2002 or whatever. In the near term.

30. Lucia: “I do expect warming trends to increase.”
    Bob Tisdale has shown some sea surface graphs that show a slight decrease in recent years.
    Perhaps we are at a turning point.
    How about a betting sweep on “in which year will the MSM first raise the NeW Ice Age threat” I bet 2016.

31. Lucia: “I do expect warming trends to increase.”
    Bob Tisdale has shown some sea surface graphs that show a slight decrease in recent years.
    Perhaps we are at a turning point.
    How about a betting sweep on “in which year will the MSM first raise the New Ice Age threat” I bet 2016.

32. I’m not so certain the problem owes to “emerging science” but more to the so far unsuccessful search for a credible explanation which of course would not yet be science.

33. Richard

    How about a betting sweep on “in which year will the MSM first raise the NeW Ice Age threat”

    This is too ill-defined to permit a bet.

34. So they revised the projection figure to nicely hide the discrepancy in the projects.

    Dana N of course is taken in by this voodoo science.

    You can argue that making comparisons as show in the modified figure is a bad idea (I would). The truth of course is in the trends.

    But stealth changes of this magnitude is going to do nothing good for the reputation of the IPCC.

35. Yes. By showing ‘projections’ that project different numerical values (and using different baselines and so on) than those used by the authors when they made them, the projections “look” better.

    This is really shameful.

    Of course when comparing to observations, SkS has been shifting the projections up from the stated baseline (of Jan 1980-Dec 1999) for some time now. And it seems Tamino is ‘explaining’ why it is the ‘right’ thing to do. Yet, in reality, in all reports the baselines were stated when projections were made. Using anything different from the stated baseline is not comparing the projections to the data.

    I’ve criticized Monckton for shifting baselines. Am I supposed to say “Oh. Well, if the IPCC does it,that’s ok then?!”

    But it’s ridiculous. The method permits those making projections to project more warming at the time the projection is made. Then, later, they redefine the projections to reduce the mismatch. And somehow, one is supposed to say “Oh! Well… they were right then!”

    The AR5 projections appear to be based on 1986-?? (Don’t remember.) I’m going to have to see if shifting them to all match in 1990 will later shift them up or down as that option will no doubt be considered should the AR7 come around.

    (And we won’t even discuss the fictional uncertainty bands not shown in the previous projections. Even if they “learned” they ought to admit that these are not the original bands!)

36. lucia, I agree with this being shameful.

    One of the consequences of shifting the baseline appears to be to make the data less consistent with the models from 1950-1970. The data are running too hot. (John F Pittman raised this general issue on Steve Mc’s blog.)

37. For a hoot, see Figure TS.14. a new term in climate science seems to have been introduced: “indicative likely range”. A quick Google search produces exactly one hit for this phrase… the AR5 report. I wish I could get away with just making stuff up whenever I want to explain away an error.

38. Lucia, Carrick,

    Re: Shameful.
    .
    Please note that politicians are generally usually quite below caring about shame, and the report is perfectly consistent with that.

39. Lucia,

    I’ve criticized Monckton for shifting baselines. Am I supposed to say “Oh. Well, if the IPCC does it,that’s ok then?!”

    hmm. Not a Team player, are you. 😉

    /sarc. I think.

40. What happened to ‘the oceans have the missing heat’?

    Were they worried that people would want to know why the oceans won’t continue to absorb heat?

41. Lucia (or anyone else),

    Do you have any idea how the shaded areas (“the
    5–95% range (±1.64 standard deviation) across the distribution of individual models”) of figures like SPM.7 or 12.5 are derived? Are they the envelope of the collection of annual anomalies from the models for each year, or the envelope of the trends from individual runs (if the latter, are recalculated for each end year)?

    It seems to me, at first blush, that there is potentially a significant difference between those two possibilities.

    Thanks,

    -Chip

42. Chip–
    No. I’ve assumed they are the envelope of the collection of anomalies from each year. That’s the sort of thing Ed Hawkins has been showing here:
    http://www.climate-lab-book.ac.uk/2013/near-term-ar5/

    There probably would be a difference between the two.

43. Posted at Climate etc:

    {D}id you notice they threw Nuccitelli and Mann’s “How The Economist got it wrong”under the bus? 1.7C went from the very likely range to likely and the IPCC is claiming it is the same where Dana and Mike correctly point out that likely is “the generally accepted 2 to 4.5°C range.”

    “”Quantification of Climate System Responses: On equilibrium climate sensitivity, several delegations, including Australia, the Netherlands and others, noted that the message that the lower limit of the assessed “likely” range of climate sensitivity is less than the 2°C in the AR4 can be confusing to policy makers and suggested noting it is the same as in previous assessments. The CLAs explained that comparison to each of the previous IPCC assessments would be difficult, and new language was developed adding that the upper limit of the assessed range is the same as in AR4.””

    They developed a new language in which 2C is actually 1.7C now. Reminds me of the scene in the book 1984 where he realizes it is easier to make a new person (language) than disappear a real person (fact). A much better fit than we are at war with _____ and always have been at war with ____ that is quoted often.

44. On further thought, the difference between the two methods is larger over short time frames than it is for longer ones–with the spread in the trends being larger than the spread in anomalies.

    -Chip

45. Thanks, Lucia (#119944),

    I wonder what the plot would look like if it were plotted the other way…instead of the shaded areas diverging over time, they would converge, with huge uncertainty at the beginning.

    The way that is depicted now gives the impression that that there is more certainty in how much temperature change the models project between now and 2050 than between now and 2100, when, in fact, the situation is the opposite.

    -Chip

46. Lucia, I can’t define “hanging tough”, but I know it when I see it:

    http://www.climateaudit.files.wordpress.com/2013/09/figure-1-4-final-models-vs-observations.png

    Now that is hanging tough, any way you want to define it.

47. Chip,

    You might find these plots useful as a way to examine the range CMIP5 of model projections. Note that this is looking at the range of anomalies for each month, not the trends per se. The weather and model (structural) uncertainties are plotted on top of eachother; the dotted line represents their sum in quadrature. Structural/model uncertainty is estimated based on differences between lowess-smoothed runs removing most variability sub-30 years. Weather uncertainty is estimated based on the residuals between the smoothed and actual runs.

    http://i81.photobucket.com/albums/j237/hausfath/Modelandweathervariabilitycomps1880-2100v2_zpsd919ea5e.png

    http://i81.photobucket.com/albums/j237/hausfath/Modelandweatherannualvariabilitycomps1970-2020_zps34d5ad06.png

48. Zeke,

    Thanks. It is interesting to see the contribution from weather as well as model differences.
    To me, though, examining the envelope of the collection of model anomalies is not particularly useful is assessing model performance against observations.
    The trend is much more robust measure.
    For example, from your first figure, the observations could stay completely within the model envelope from 2030 to 2100 and yet have a zero (or perhaps even negative) trend–which I am almost sure would be a gross odds with all the climate models.
    -Chip

49. Zeke,

    IMO, the GCM’s and the IPCC are burned toast. In many ways, AR5 WG1 is laughable; the remainder of AR5 will only get worse. Maybe you should find another line of work. 😉

50. OK, I think I figured it out:

    This figure I made contains the range of trends for the RCP8.5 runs starting in 2006 and ending in the year on the x-axis (2015-2100):

    http://www.worldclimatereport.com/wp-images/AR5_CMIP5_RCP8.5_trends.jpg

    As I suspected, the range is wide at first, narrowing to mid-century and then broadening a bit as intermodal differences start to dominate model noise–just as Zeke’s figures indicate.

    -Chip

51. Chip,

    I agree that its an incomplete method of assessing any particular trajectory. However, it might be useful to examine subsets of models with structural behaviors more (or less) consistant with the longer-term temperature record and see if the recent pause falls outside the weather noise of those models.

    SteveF,

    GCMs certainly are far from perfect, and the multi-model mean is biased high. That doesn’t mean all GCMs are garbage. As far as the AR5 goes, the chapters I’ve read so far are much better (and more careful) than the equivalent AR4 chapters. That said, I’ve only really gotten though chapter 4 so far, as its quite a long and dense document.

52. Zeke,
    “GCMs certainly are far from perfect, and the multi-model mean is biased high.”
    .
    There is hope for you! 😉
    .
    But honestly, why do you (as a budding climate scientist) think the IPCC does not just acknowledge that (obvious to most all) reality? Try as I might, I really can’t understand their take on this. I mean, as scientists, it would seem that progress toward a clear understanding of the system ought to trump all else. Yet we get things like “indicative likely range” and graphical baseline changes/shifts from AR4 to AR5 instead of a straightforward discussion of ‘the pause’. Really pretty awful stuff.

53. The climate models which are the most accurate are the ones with the lowest sensitivity. Even these are too high.

    Simple enough.

    Now how do we get 3,000 climate scientists and 200 million followers and 50 governments and 90% of the media to recognize it.

    We need a miracle for the truth to be recognized.

    Strange enough.

54. Zeke, I am starting to think that GCM’s are really little better than simpler models based on energy balance. The problem I think is too much dissipation. That will damp the dynamics. There is some evidence (from responses at Real Climate) that when Hansen first came up with this idea that if you take a weather model that diverges from reality after about 48 hours and run in on a course grid with a large time step that the results is meaningful, that his proposal was criticized for exactly this reason, namely, that the individual trajectories would be very wrong, so what is the reason to think that if you integrate far enough, your results has any meaning? The only reasons to this day are just hand waving.

55. David Young–
    I think the idea of running GCMs predates Hansen, though I might be wrong.

    I’m not bothered by individual trajectories deviating from earth’s. Running ensembles and averaging is done in engineering and it can work even though no individual trajectory matches any individual trajectory in a ‘observed’ flow and never will even though we can even run multiple experiments to get many observed flows.

    The real issue is that you need to incorporate more physics into a climate model.

    For example: The trajectory in a weather model will ‘go wrong’ in a short amount of time due to initial conditions. This is a challenge because one might wish to predict weather for the next year, but the deviation from earth’s trajectory arises long before we need to worry about whether we’ve correctly modeled the effects of evapo-transpiration at the soil later, heat transfer at the surface of the ocean and certainly circulation in the deep ocean. In many cases, one can neglect these sort of difficult to model features entirely when predicting weather over a day, week or even month. But for climate, these things become very, very important as they will affect the heat balance. When predicting small changes in the temperature of the earth (1K relative to 293K or possibly 1K relative to a range of 80K from pole to equator), these things become even more important. (And all of these difficulties arise before we even consider ‘uncertain’ information like the precise volcanic aerosol loadings over the full sky and well before we consider truly ‘unknown unknowns’.)

    So climate is difficult to predict even if we could deal with numerical issues like excess numerical dissipation of either heat or mechanical energy, effects of grid resolution and so on.

56. Yes, Lucia, I agree. Even if we “fixed” all the numerical issues, I doubt if long term performance of GCM’s would improve much. You are looking for a very small effect that will on any finite computer be smaller than the numerical errors. But I’m not sure more physics will do much good. I tend to favor simpler models that incorporate observational based feedbacks. There are a number of examples of this in CFD where making a model more complex and including more physics can actually make it less accurate. The problem is that you also may introduce more parameters that are virtually impossible to constrain meaningful with observations. An example is Reynolds’ stress models vs. eddy viscosity models (which are simpler). Another is boundary layer methods vs. RANS methods. The boundary layer methods often are better because we know a lot about boundary layers.

57. David Young (Comment #119964),
    “Even if we “fixed” all the numerical issues, I doubt if long term performance of GCM’s would improve much.”
    .
    Sure. But the GCM’s pretty much all diverge from the Earth in the same direction (too warm), at least when many runs are averaged. Seems to me that the errors in the existing model physics and/or in existing model parametrizations are most likely to be limiting performance.

58. SteveF, I suspect that different parameters might yield lower forecasts. But the codes are so complex, I’m not sure anyone knows for sure. I like simpler models where you can constrain them meaningfully by data.

59. David Young (Comment #119957)
    October 1st, 2013 at 8:46 pm
    … Zeke, I am starting to think that GCM’s are really little better than simpler models based on energy balance. The problem I think is too much dissipation. That will damp the dynamics. …
    —–
    FWIW… I agree with you David

Comments are closed.