Comment on unproven models

A recent comment, here, questions the AGW prediction of polar amplification. He cites a paper by Polyakov et al that he claims shows temperatures in the arctic were warmer than they are now earlier in the 20th century.

[Update: paper is here[PDF]]

I don’t have access to the paper or time to research it well, does anyone else have any comments? My initial impression is that it is about ocean temperatures in one region of the north Atlantic, and I do not trust the numbers he quotes which came from CO2 Science, a site that habitually misrepresents the papers it highlights. But that is hardly enough to base a respectful reply upon.

Regarding the antarctic, I would just point out that the latest studies do indicate some mild warming, not cooling, and sea ice formation is a complicated process, so it is difficult to base firm conclusions on very minor trends. But regardless, models do not anticipate a strong amplification effect there until later. The antarctic is isolated climatically by the circumpolar current in the southern ocean so it will be some time before a dramatic rise is expected on that continent as a whole.

Here is his comment:

I have a few problems with the models:

First, the models call for polar amplification.

Polyakov shows that the Arctic warming trend is only slightly higher than the Northern Hemisphere trend. He also shows that the Arctic is cooler now than it has been in the 20th century.

Polyakov et al (2003) The composite temperature record shows that since 1875 the Arctic has warmed by 1.2°C, so that over the entire record the warming trend was 0.094°C decade−1, with stronger spring- and wintertime warming. The Arctic temperature trend for the twentieth century (0.05°C decade−1) was close to the Northern Hemispheric trend (0.06°C decade−1). The oscillatory behavior of Arctic trends results from incomplete sampling of the large-amplitude LFO. For example, the Arctic temperature was higher in the 1930s-40s than in recent decades, and hence a trend calculated for the period 1920 to the present actually shows cooling. Enhancement of computed trends in recent decades can be partially attributed to the current positive LFO phase.

Serreze realized this problem with the climate models and attempted to show that ice loss would actually start the Polar amplification.

Serreze et al (2006) Rises in surface air temperature (SAT) in response to increasing concentrations of greenhouse gases (GHGs) are expected to be amplified in northern high latitudes, with warming most pronounced over the Arctic Ocean owing to the loss of sea ice. Observations document recent warming, but an enhanced Arctic Ocean signal is not readily evident. This disparity, combined with varying model projections of SAT change, and large variability in observed SAT over the 20th century, may lead one to question the concept of Arctic amplification.

Then we look at the Antarctic. Once again studies show that the Antarctic was the warmest in the 1930-1940s. Further, the Antarctic has been cooling for the last 40 years.

So there is currently no polar amplification. The only way to show one is to cherry pick your start date, as was done for the Antarctic to show warming.

I do not have the time to answer Vernon thoroughly so I appeal to any readers who might be inclined to help out here.

Thanks in advance!

82 thoughts on “Comment on unproven models

  1. Crakar –

    Very well. I apologize for misinterpreting your intentions. I hope, given your previous comments, that you understand why I would misinterpret what your point was.

    Like

  2. I think at times all (myself included) need to step back and take a deep breathe.

    I am pretty thick skinned so i dont mind if you or anyone else for that matter disagrees with me (re you above posts, muppet etc).

    I do not intend to deliberately mislead the debate by claiming things as fact when i know they are not.

    The term “beggars cant be choosers” meant we cant see the entire 30 year record because the sat was broke but we can still see the current ice levels.

    I hope this clears up a few things.

    Like

  3. Yet a few more studies that have identified climate factors that the models fail to capture.

    Bony et al (2005)
    http://www.met.sjsu.edu/~tesfai/RESULTS/Journals/how%20well%20do%20we%20understand%20and%20evaluate%20climate%20change%20feedback%20processes.pdf
    Based on recent studies, no substantive evidence suggests that the weak relative humidity response of climate models, and thus the large magnitude of the water vapor–lapse rate feedback under climate change, are an artifact of climate models. However, as the water vapor feedback represents the strongest positive feedback of the climate system, uncertainties about how small relative humidity changes should be, or how accurate the magnitude of the correlation between humidity and temperature should be, can matter for the spread in water vapor–lapse rate feedback and for the magnitude of climate sensitivity.

    However, recent comparisons of the observed and simulated variations of water vapor and relative humidity in the current climate reveal biases in GCMs (sections 3c, 3d, and 3e), and there is still a nonnegligible spread in the model estimates of the water vapor–lapse rate feedback under climate change (Fig. 1). This spread is likely to result from intermodel differences in the meridional patterns of surface warming and in the magnitude (albeit small) of relative humidity changes.

    Bennhold et al (2008)
    http://66.102.1.104/scholar?hl=en&lr=&scoring=r&q=cache:-QFJipqUrooJ:earth.geology.yale.edu/~sherwood/BennholdS08.pdf+cloud+negative+climate+feedback
    As a primary result of this work, we have documented the mean biases between models and products derived from satellite observations, namely, upper-tropospheric humidity, precipitable water, and net cloud radiative forcing, which are too high, too low, and too strong, respectively, in all models. In addition, we have documented several significant and systematic discrepancies between modeled and observed relation-ships among cloud- and moisture-related variables.

    STEPHENS (2005)
    http://irina.eas.gatech.edu/EAS_spring2006/Stephens2005.pdf
    While GCM climate and NWP models represent the most complete description of all the interactions between the processes that establish the main cloud feedbacks, the weak link in the use of these models lies in the cloud parameterization imbedded in them. Aspects of these parameterizations remain worrisome containing levels of empiricism and assumptions that are hard to evaluate with current global observations. For example, the relationship between convection, cirrus anvil clouds, and SST is a recurring theme in many feedback hypotheses (section 5) yet the connections between convection and cirrus in parameterization schemes is highly uncertain, in many cases empirical, and difficult to evaluate with observations.

    Ringer et al (2006)
    http://luv.dkrz.de/publications_2006/pub_362_402.pdf
    Consistent with Senior and Mitchell [1993] we find that in any given model the sign of global mean net cloud feedback, and of its shortwave and longwave components, may differ between ±2K and 2 CO2 experiments. The relative strength of the cloud feedbacks across GCMs is also likely to be different. However, the most important conclusion of the original Cess et al. [1990] study, namely that the variation of the total climate feedback across an ensemble of GCMs depends primarily on the variation in the cloud feedback, holds in both cases.

    Williams et al (2007)
    http://cosis.net/abstracts/EGU2007/01305/EGU2007-J-01305.pdf
    It is shown here that evaluation and subsequent improvement in the simulation of the present-day regime properties has the potential to reduce the variance of the global cloud response, and hence climate sensitivity, amongst GCMs. For the ensemble of models considered in this study, the use of observations of the mean present-day cloud regimes suggests a potential reduction in the range of climate sensitivity of almost a third.

    have fun reading.

    Like

  4. Sorry for the delay in posting but I was busy with RL for the last month+.

    Anyway, yet another nail in the coffin of GCM modles being unproven. None of the models show why there would be 20-30 periods of warming or cooling that is not related to CO2 or volcanic activity.

    Basically, as was in a current RC post, there is current cooling and warming is not expected to resume till 2020. Please point to any GCM’s that predicted that?

    Like

  5. None of the models show why there would be 20-30 periods of warming or cooling that is not related to CO2 or volcanic activity.

    Models generate noisy up-and-down results just like the real world, imposed on a trend of increasing temperature.

    You’ve been told this approximately 527,987 times in the past, Vernon.

    Basically, as was in a current RC post, there is current cooling and warming is not expected to resume till 2020.

    June, according to NOAA and GISTemp, was the 2nd warmest on record. The piece posted on RC was by Swanson, who was explaining a controversial paper he co-wrote.

    Note that his prediction and analysis aren’t accepted by the climate science community. For one thing, they only hold for one dataset, HadCrut. For another, the supposed “current cooling” isn’t statistically significant and indeed is only “cooling” if you carefully cherry-pick starting points over the last decade. Other starting points yield a rising trend. Neither is a statistically valid trend.

    Please point to any GCM’s that predicted that?

    Basic understanding of statistics will tell you that a noisy signal imposed on a rising trend will yield a curve where short-term downward portions would appear.

    Digustingly, someone had to actually write a paper in a journal to point out with graphics why this is true, and Tamino did the same with his “wiggles” post over at Open Mind.

    Disgustingly, because anyone whose studied statistics for even a few hours knows this.

    Anyone who’s ever gambled knows this.

    The obvious only needs to be pointed out to denialists whose ignorance is unbounded.

    Like

  6. yet another nail in the coffin of GCM modles being unproven.

    Oh, and I’m sure you meant to say “proven”, not “unproven”. Regardless, no one claims GCMs are “proven”. They’re *useful*. They’re not complete; no “modle” will ever be complete.

    However, research is always being done to make them better.

    For instance, this paper you reference above:

    It is shown here that evaluation and subsequent improvement in the simulation of the present-day regime properties has the potential to reduce the variance of the global cloud response, and hence climate sensitivity, amongst GCMs. For the ensemble of models considered in this study, the use of observations of the mean present-day cloud regimes suggests a potential reduction in the range of climate sensitivity of almost a third.

    have fun reading.

    You’re right, it *is* fun reading, the abstract, at least.

    They claim to be on to something that will allow them to tighten up the range of climate sensitivity by 1/3 by improving modeling of cloud-related phenomena, an area that all modelers openly state is the most poorly modeled bits of climate.

    In other words, rather than the current range of climate sensitivity computed by model of say 2-4.5C, they say their work might enable modelers to tighten up that range around the 3.0C mark, to say about 2.5-4C.

    This constitutes “victory” for denialists, in your eyes?

    Or is this just another case of Vernon reading a paper and totally misunderstanding what the authors are saying?

    Like

  7. dhogaza,

    The point of this discussion is not what is the climate sensitivity to CO2 but rather, have the GCM models been shown to be accurate enough to base future policy on. My point is that they have not been.

    If we can get proven GCM models that accurately project future climate, then they could be used as a basis for policy decisions. GCM’s have not reached that point yet.

    That is the point of this discussion.

    Like

  8. Vernon –

    The point of this discussion is not what is the climate sensitivity to CO2 but rather, have the GCM models been shown to be accurate enough to base future policy on. My point is that they have not been.

    You have stated this numerous times, and yet cannot show anything that demonstrates that the models are way off base. What data do you have to show that the models make poor representations of the climate?

    Most people with any knowledge in the field would say that to ‘valid’ a model, you compare the model results to observations. Given that, using the 20th century to test models is an appropriate dataset, because we know both the outputs and the inputs. We can see that the models do a decent job of reproducing 20th century temperatures.

    Also see here: http://scienceblogs.com/illconsidered/2006/03/models-are-unproven.php

    Sure there are parameters or factors going into the model that we [well, not me, but climate scientists in general] don’t know well enough to reproduce very accurately. But we don’t need perfection, not every little thing affects the general output of the model. No model is perfect, and we sacrifice accuracy for simplicity where we can (the law of diminishing returns applies to models as well as anything else, so we can, at some point, make this trade without losing too much accuracy).

    As an example, we can safely ignore relativistic effects when modeling earth-bound, human scale mechanical systems. Strictly speaking, the model is ‘incorrect.’ But, we don’t need to add that complexity, because we get good results without it.

    But, this isn’t new information to you. You know all this, and have heard it all before. So I have a question for you, Vernon. What test do you propose would determine the models’ validity? What would Dr. Vernon, Climatologist, do to test his own climate model before releasing the results for scrutiny? What would Dr. Vernon need to see before he was comfortable basing future policy on his model?

    Like

  9. Adam,

    Why don’t you try to answer peer reviewed studies with something other than opinion pieces. I have presented peer reviewed studies that address the problems with the models. You have presented opinion and links to other peoples opinions. Why not try to use current peer reviewed science.

    So, peer reviewed studies vs your opinion.

    Sorry, but I have seen nothing that shows the were the models are good enough to use as a basis for policy. I have listed many studies that have found problems. RC has posted a study that shows that there is major dynamics and processes that the models do not capture because we have not figured out what they are or how they work.

    I am presenting studies, arm waving is not equivalent.

    Like

  10. I have presented peer reviewed studies that address the problems with the models.

    Yes, and I showed that one is meant to narrow uncertainty, meaning the “best estimate” of 3.0C forcing to a doubling of CO2 is strengthened.

    Do you always shoot yourself in your (private member)?

    Like

  11. Sorry, but I have seen nothing that shows the were the models are good enough to use as a basis for policy.

    No one gives a shit about your opinion, in science, or outside science.

    I’m sure this is frustrating to you, as you’re so certain that you know more about the science, math, etc than thousands of professionals.

    But – that’s life!

    Like

  12. Vernon –

    Sorry, but I have seen nothing that shows the were the models are good enough to use as a basis for policy.

    Again, you’re avoiding the question. What would Dr. Vernon do to test his own climate model, what data would he input into it, and what would he look for. You make a lot of claims without much substantiation, and shit on actual scientists quite a bit, so why don’t you stick your own neck out and say straight up what results would convince you.

    I have listed many studies that have found problems.

    Again, the ability of the models to hindcast the 20th century accurately demonstrates that, despite whatever problems they have, they are largely accurate. The real question is, if we fix all these problems, will it fundamentally change the modeling results? Ask dhogaza points out to you, this is pretty unlikely.

    Why not try to use current peer reviewed science.

    The first link I provided is a graph taken directly from the IPCC 4th assessment, which is the synthesis of the peer reviewed research as of 2007. And again, of the very few references you actually make, none of them actually, erhm, say what you wish they said.

    Why don’t you try to answer peer reviewed studies with something other than opinion pieces.

    The only opinions I expressed was a discussion on the relationship between complexity and accuracy when constructing mathematical models and how to test model validity. Nothing I said is particularly controversial. Anyone who has ever constructed a model of anything will tell you this. You can never be perfect, but you can be good. Also, Coby’s post has links to a subset of the successful modeling predictions, if you had bothered to read it you might know that.

    Like

  13. dhogaza,

    to bad you have nothing to contribute to the discussion other than ad hom attacks. Coby indicates your a smart person that must have much to offer, yet I rarely see more from you than personal attacks.

    Adam,

    I have list listed around seventeen studies, most that are after the IPCC 4th AR which all address problems with the models. Even over on RC there is a discussion of how there may be no warming till 2020.

    When the models can explain how we get into and out of a period of glaciation, can explain periods like the MWP, RWP, LIA, etc. then I would consider them good enough to make policy decisions on.

    Like

  14. to bad you have nothing to contribute to the discussion other than ad hom attacks.

    Once upon a time, upon first meeting you on the internets, I was nice to him and tried to point out his errors regarding climate science.

    So did many others, including professionals like those at RC and Tamino’s Open Mind.

    Your response was to continue making the same claims, repeatedly, in many venues, over and over.

    So I stopped being nice to you. So did many others …

    Like

  15. I already answered your post on that, Vernon.

    Once again you’re interpreting a scientist saying that current models don’t capture some of the ways in which things might be MUCH WORSE (ocean model, in this case) as being supportive of your notion that there’s nothing to this AGW stuff.

    Like

  16. Vernon said:

    Well, you do not like it when I point out the flaws so how about some one else

    Vernon, you are obviously not a biologist since you keep on mixing apples and oranges. Do you really think that Terrence Joyce is discussing GCM’s in that article?

    Just to save you the embarrassment of repeating it on other blogs, he is discussing ocean circulation models not GCM’s. They are much further behind than GCM’s since there is a paucity of data on the oceans unlike the atmosphere. The introduction of the ARGO system and the exploration of the abyssal depths for temperature and salinity data will greatly improve these models in the future.

    Like

  17. Just to save you the embarrassment of repeating it on other blogs, he is discussing ocean circulation models not GCM’s.

    Both HadCM3/4 and GISS GCM Model E include dynamic, gridded ocean models coupled to the atmospheric models. They can be turned off (for instance when the HadCM model’s run in UK weather forecasting mode – they use the latest observed ocean “weather” data and run the atmospheric model with that representation of the ocean). Well, actually apparently at the moment Model E has five different ocean models to choose from, depending on what questions are being asked of the model.

    So the criticism of ocean circulation models is fair since they get coupled to the atmospheric models.

    But Vernon’s still missing the point …

    Like

  18. Dhogaza, I think we are talking about two different kinds of ocean circulation models. The GCM’s are linked to surface circulation whereas the models Joyce is working with involve overturning circulation involving thermo-saline conditions.

    Here are two quotes from Joyce:

    A 3-4 meter, high latitude buildup of fresh water over this time period has decreased water column salinities throughout the subpolar N. Atlantic as deep as 2000m. At the same time, subtropical and northern tropical salinities have increased.

    The degree to which the two effects balance out in terms of fresh water is important for climate change. If the net effect is a lowering of salinity, then fresh water must have been added from other sources: river runoff, melting of multi-year arctic ice, or glaciers. A flooding of the northern Atlantic with fresh water from these various sources has the potential to reduce or even disrupt the overturning circulation.

    Whether or not the latter will happen is the nexus of the problem, and one that is hard to predict with confidence. At present we do not even have a system in place for monitoring the overturning circulation.

    Models of the overturning circulation are very sensitive to how internal mixing is parameterized. Recall that internal mixing of heat and salt is an integral part of overturning circulation. One recent study shows that for a model with constant vertical mixing, which is commonly used in coupled ocean-atmosphere climate runs, there is only one stable climate state: our present one with substantial sinking and dense water formation in the northern N. Atlantic.

    ———————————————

    For the ocean, our data coverage is wholly inadequate. We can’t say now what the overturning circulation looks like with any confidence and are faced with the task of predicting what it may be like in 10 years!

    http://www.whoi.edu/page.do?pid=12455&tid=282&cid=10046

    Like

  19. Dhogaza, I think we are talking about two different kinds of ocean circulation models. The GCM’s are linked to surface circulation whereas the models Joyce is working with involve overturning circulation involving thermo-saline conditions.

    That could be, I was just looking at the surface-level (pun intended) claim that GCMs don’t include dynamic circulation models. They do, but as to how close they are to what Joyce is working on, I don’t know. Maybe I’ll spend some time looking, this week.

    However, we’re both agreed, I think, that …

    1. Vernon misrepresents Joyce’s comments in regard to the utility of GCMs

    2. He’s either too uneducated to realize it, or dishonest.

    3. And well maybe it’s just me, not you, but I think he’s a member of the Unteachable Caste.

    Like

  20. Oh, and this quote you provide backs up what I was trying to tell Vernon, at least:

    One recent study shows that for a model with constant vertical mixing, which is commonly used in coupled ocean-atmosphere climate runs, there is only one stable climate state: our present one with substantial sinking and dense water formation in the northern N. Atlantic.

    He’s saying that accurate modeling might show that possible response to AGW will be a second stable state, different than ours. Which, regardless of the change in average global temps, means hugely different regional climates based on shifting warm/cold water currents, etc.

    In other words, as I’ve been trying to tell Vernon, THINGS MIGHT BE MUCH WORSE THAN WE IMAGINE.

    Since the whole thrust is in the context of “response to AGW”, there is no comfort there for Vernon etc denialist camp. Far from stating that AGW isn’t real, or the response might not be important, they’re saying “HOLY SHIT! We have no *effing* idea because the models aren’t capable of modeling a state flip in the ocean!”

    Vernon – the measured statement cited by you is how scientists respond when they’re worried.

    Like

  21. dhogaza,

    You and Forrester are a trip. The WHOI clearly say that GCMs must include the oceans, since they are the major driver after the sun, which they do not do now. Just taking sea surface temperature does not work. To have you and Forrester say that that page has nothing to do with GCMs is an complete lie. I guess that is ok but I suggest that anyone that doubts what I quoted go read the whole thing at:

    http://www.whoi.edu/page.do?pid=12455&tid=282&cid=10046

    Like

  22. The WHOI clearly say that GCMs must include the oceans, since they are the major driver after the sun, which they do not do now.

    You’ve been told you’re wrong, and you’re repeating it again.

    That makes you a liar.

    HadCR3:

    “The ocean model has a resolution of 1.25×1.25 degrees and a timestep of 1 hour. Thus there are 6 ocean grid points for every atmospheric one.”

    GISS GCM Model E has several available models, including:

    “GISS Dynamic ocean model (Russell)

    This model is a fully dynamic, non-Boussinesq, mass-conserving free surface ocean model (Russell et al, 1995, 2000, Liu et al 2002). The dynamics is based on a modified Arakawa scheme on the C-grid, with a linear upstream scheme for advecting tracers. Vertical mixing uses the KPP scheme of Large et al (1996). Momentum mixing is modelled as a spatically varying laplacian as in Wajsowicz (1993). The effects of mesoscale eddies and isopycnal diffusion are parameterised as in Gent and McWilliams (1996), but using variable coefficients (Visbeck et al, 1997), and coded as in Griffies (1998).

    The model contains up to 12 variable depth subgrid scale straits which contect ocean grid boxes, which would not be connected at the resolution used. In particular, the Straits of Gibraltar, Hormuz, and Nares straits are so modelled. All ocean components are fluxed through these straits as a function of the end to end pressure gradients, balanced against a drag proportional to the straits ‘width’ which serves as a tuning parameter to get reasonable fluxes.”

    Like

  23. they are the major driver after the sun

    And this deserves a emphatic WTF?

    So how is the ocean generating energy? Cold fusion of water molecules?

    Like

  24. dhogaza,

    Once again… The sun is the only source of energy of any significants. The oceans store almost all the energy (~99%) that runs the climate. Do you disagree with this? The atmosphere does not warm the oceans, rather the oceans hold and release heat for reasons that I have not seen answered yet.

    That was Joyce and Keigwin are saying. That a GCM that does not incorporate a model reflecting the total oceans circulation will not work. They further say we do not know enough at this point to create one.

    Like

  25. Yet another study that finds problems with the current GCMs.

    Hunt (2005)
    http://www.springerlink.com/content/677kx61w6812gx81/

    It was therefore concluded that these fluctuations were generated by stochastic processes intrinsic to the nonlinear climatic system. While a number of characteristics of the MWP and the LIA could have been partially caused by natural processes within the climatic system, the inability of the model to reproduce the observed hemispheric mean temperature anomalies associated with these events indicates that external forcing must have been involved. Essentially the unforced climatic system is unable to sustain the generation of long-term climatic anomalies.

    Like

  26. Well, I have presented a lot of studies that have found problem with the models. S&T’s new study shows that natural climate drivers are much stronger than CO2 based warming.

    Coby,

    ready to concede that the models are not ready for prime time? Since I have not seen anyone actually showing that the models do not have the problems which I have presented.

    Like

  27. blueshift

    Vernon probably won’t reply, because his previous post was two years ago. But it does beautifully show how deniers have no idea about science.

    Note how his link and his quote make this statement:

    “….. Essentially the unforced climatic system is unable to sustain the generation of long-term climatic anomalies….”

    There has never been a clearer statement about the idiocy of the ‘natural variation’ claim. That paper very clearly states that the climate MUST have some sort of external forcing mechanism if it is to sustain long term anomalies. It doesn’t just change on its own.

    It proves – once again – that deniers don’t read science. They rely on others to give them their opinions. And those others are equally bad at understanding science and rely on ideology rather than evidence for their views.

    Like

  28. Mandas,

    There was some sort of spam comment that made me think it was an active thread. After I submitted I realized my mistake.

    You are of course right, Vernon’s post illustrates the ignorance of the deniers.

    Like

  29. So have they worked out what the “error” this introduces is worth?

    I.e. it could be that the model UNDERESTIMATES warming because of that lack. It could be that the error makes the mean 2.95C per doubling rather than 3.0C per doubling.

    What we observe now indicates that any sensitivity below the lower end of the IPCC projections is proven false.

    Also, if all those other scientists have it wrong, what about Joyce and Keigwin?

    Like

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s