Let’s assume for the sake of argument that there’s so much slop in that graph that any feature we observe might be off by +/- 13 years. In that case, how did the models end up fitting so beautifully with the post-1950 instrumental record? What an amaaaaaazing coincidence. Maybe the IPCC reviewers should buy some lottery tickets!!
Talking generally here, not about this particular research: When you have two competing models, it’s common to adjust the parameters in each model, and see how close each model can get to the observed data. If model A is a good match (within the error bar of the data) and model B is obviously off, that’s a very good indication that model A is correct.
In other words, there may be free parameters (“fudge factors” if you insist) in both models. But the point is, model A could be made to fit the data, while model B couldn’t fit the data even by adjusting the parameters.
My understanding is that the result in question was produced the way I described, but I haven’t read the actual paper.
I am reminded of a quote from the late great Johnny Von Neumann:
I’m not sure what you are referring to by “model A” and “model B.”
But my understanding is that the blue area in the graph is created by taking the models that have been tuned to the instrumental record and eliminating the inputs that represent anthropogenic forcings. Obviously this will result in a non-match.
Not if you have 5 or more data points.
Really? I thought the blue area was the best case result of trying to reproduce the observed temperatures without using anthropogenic CO2.
Could you remind me which paper that plot came from?
I don’t see why. Metaphorically speaking, an elephant with a wiggling trunk would seem to have a lot of data points.
jshore linked to it above.
Are you intentionally misunderstanding stuff? It is not a matter of being off by 13 years. It is a matter of there being small ups and downs that are due to noise.
And, your understanding is obviously wrong. But, you will continue to believe it anyways so it is useless to continue to correct you.
What are you talking about? Did you even bother to read this thread?
Here’s what scr said:
Are you intentionally misunderstanding stuff?
Do you have any cite at all to support your position? Any evidence? Any coherent argument?
I think you’re misunderstanding what I said. I merely pointed out that you have no idea if the plot says any of the things you think it says, because you don’t know which features are statistically significant and which are not.
Well, you’re the one making the claim about how the simulation was done. How about quoting the relevant section of the paper, or even the abstract, that describes how these simulations are done?
We’ve spent pages discussing this stuff. However, you are the one who is making wild statements about a whole scientific field being utterly incompetent. What is more amazing is that you do so without even demonstrating that you have the slightest understanding of how climate models actually work. It is up to you to actually back such wild statements up.
This stuff is really becoming tedious. Other posters have in the past complimented me on my patience in dealing with climate change “skeptics” who post here and continually treading the same ground again and again…but I have to say that you are really trying my patience. I see little use in discussing this stuff with you further.
It must seem thankless. But I appreciate seeing someone who knows what they are talking about continuing to knock this stuff down. If these arguments were left to stand unrebutted, I think a lot of lurkers would find them more plausible for that reason alone.
Your claim was a little more specific than that. Among other things, you said the following:
Those were your words, not mine.
Here was my response:
And here is jshore’s repsonse to me:
He completely missed the point. It WAS a matter of being off by 13 years. Because of what you said.
Nice try at shifting the goalposts, though.
I’ve laid out my case based on the output of the simulations. There is strong evidence of fudging. But for the heck of it, I dug up a couple abstracts. Here is one of them:
I take it by “ensemble simulations,” what they mean is that the same model was run multiple times with different inputs representing the presence or absence of anthropogenic forcings.
In other words, this would not appear to be a situation where somebody tried to construct a model including anthropogenic and succeeded (in matching the temperature record); then tried to construct a model without anthropogenic forcings and failed (in matching the temperature record).
Sure, and you have failed to adequately answer my simple questions. Most recently, how a group of models with 1 set of forcings produce outputs so noisy that an obvious peak in 1950 is just random noise; and yet a different set of forcings give outputs that beautifully match the instrumental record.
Lol. Would you care to use the quote feature to back that statement up?
Please show me a wild statement I made about a whole scientific field being utterly incompetent.
You were the one who brought up the 13-year “discrepancy” in the first place. You then admitted that you don’ t know the accuracy or resolution of the data, implying you don’t know whether the data supports your point or not.
I don’t understand what that evidence is. Just because it fits too well?
Your quote says: “The late-twentieth-century warming can only be reproduced in the model with anthropogenic forcing” (bolding mine). That means they tried to reproduce it with and without anthropogenic forcing, and only succeded with the model with anthropogenic forcing.
My point is supported either way, for reasons stated earlier.
No. I’ll lay it out for you one more time: If the resolution of the graph is sufficient that the obvious peak in 1950 is indeed a significant peak, then it shows that the models are fudged. Because there was no natural forcing that can account for the peak.
On the other hand, if the graph is so noisy that the obvious peak in 1950 might very well represent a peak in 1963, then it’s extremely unlikely that the same models – even if they were realistic – would yield an output (with all forcings) that beautifuly matches the instrumental record after 1950.
So either way, it’s pretty likely that the models are fudged.
This is partially corroborated by the IPCC admission that some or all of the models contain adjustments that don’t correspond to any physical principles.
It depends how you define “try.” It doesn’t look to me like they “adjusted” anything in order to make the output with non-anthropogenic forcing fit the temperature record.
I’m not even sure if we’re looking at the same graph. We’re talking about figure 9.5 in the IPCC WG1 AR4 report, right? Page 684 of this PDF document? I don’t see an “obvious peak in 1950”. I see a steady temperature from 1950 to 1960, a slight (insignificant) drop to 1962, then an obvious drop between 1962 and 1965 which corresponds to the Argung volcanic eruption.
In any case, if there is so much “fudging” in climate modeling, I’m sure there would be papers claiming their models were able to reproduce the observed temperatures without using anthropogenic forcings. Do you have such a cite?
Where in the IPCC report does it say that?
Of course it’s a very poor fit, if that’s what you mean. That’s the point of that research result: they couldn’t get the non-anthropogenic model to match the observed temperature. Probably because there are no observed positive non-anthropogenic forcings that could even be tweaked to explain the observed temperature change.
Actually, I’ve mainly been referring to the figure in the FAQ section of the same document.
I’m not sure of that. But I agree it would be an interesting challenge. If I had the time; the computer skills; and the source code to the models in question, I would give it a try.
In any event, one individual has made the following observation:
See my quote from a few posts back.
With the anthropogenic forcings removed, it doesn’t appear that anyone tried to “tweak” the model or to “get” a particular result.
Why do you think that they did try?
Okay…
So, in other words, you are accusing the entire field of being so stupid that they tune their model so it fits the historical temperature record when they include anthropogenic forcings and are then surprised when it doesn’t fit well once they are removed.
You are accusing them of failing to understand that with many free parameters properly chosen in a model you can fit anything.
You are, in short, accusing the whole field of being utterly incompetent as scientists. (Strangely enough, given that you believe they have the parametric freedom in the models such that they can tune their models to reproduce the historical temperature record, you are unable to provide one cite where a group tuned the model to reproduce the historical temperature record with only natural forcings. So, you must think there is some vast conspiracy to prevent anyone with access to a climate model from performing such an elementary experiment.)
Which figure? FAQ 8.1 is the same as 9.5.
Which post?
This conflates two different issues: One is whether the reason that the models can simulate the historical temperature record (and, in particular, the warming in the last ~30 years) with anthropogenic forcings but not without is because they have been tuned so that they fit with the anthropogenic forcings.
The second is whether, given the known uncertainties in the equilibrium climate sensitivity and in the component of anthropogenic forcing due to aerosols, it is possible to get good agreement with the historical temperature record with different combinations of equilibrium climate sensitivity and aerosol forcing.
The answer to the first question is NO. However, the answer to the second is YES, at least to a degree,…and this is (much of the reason) why we still have the range of possible values that we do for the equilibrium climate sensitivity and resulting uncertainty in projections for future warming. (Of course, there is further uncertainty introduced into the projections of future warming due to uncertainty in future greenhouse gas and aerosol emissions.)
So, yes, it is true that, given our current lack of understanding of the aerosol forcing, the historical instrumental climate record does not give all that strong limits on the possible equilibrium climate sensitivity. However, this is not a new fact that has been discovered by “skeptics” but is the known fact regarding the current uncertainties in climate science and is incorporated into the IPCC projections of future warming. If you asked any climate scientist what the two biggest uncertainties are in the science that set limits on the accuracy with which we can predict the response of the climate system to greenhouse gases, I think nearly all of them would say the radiative forcing due to aerosols and the behavior of clouds (and the two are not completely independent since part of the uncertainty in the aerosol forcing is due to how they affect clouds…the so-called “indirect aerosol effect”).
If you actually read the IPCC report, you can find long involved discussions about climate sensitivity and the extent to which it is constrained by (1) results of the computer models, (2) the historical instrumental temperature record, (3) other historical records such as the ice age - interglacial cycles.
See, in particular, Section 9.6 of the report and also Section 10.5, especially Box 10.2.