Arguments
Is the CO2 effect saturated?
What the science says...
| Select a level... |
Basic
|
Intermediate
|
Advanced
| ||||
|
The notion that the CO2 effect is 'saturated' is based on a misunderstanding of how the greenhouse effect works. |
|||||||
Climate Myth...
CO2 effect is saturated
"Each unit of CO2 you put into the atmosphere has less and less of a warming impact. Once the atmosphere reaches a saturation point, additional input of CO2 will not really have any major impact. It's like putting insulation in your attic. They give a recommended amount and after that you can stack the insulation up to the roof and it's going to have no impact." (Marc Morano, as quoted by Steve Eliot)
The mistaken idea that the Greenhouse Effect is 'saturated', that adding more CO2 will have virtually no effect, is based on a simple misunderstanding of how the Greenhouse Effect works.
The myth goes something like this:
- CO2 absorbs nearly all the Infrared (heat) radiation leaving the Earth's surface that it can absorb. True!
- Therefore adding more CO2 won't absorb much more IR radiation at the surface. True!
- Therefore adding more CO2 can't cause more warming. FALSE!!!
Here's why; it ignores the very simplest arithmetic.
If the air is only absorbing heat from the surface then the air should just keep getting hotter and hotter. By now the Earth should be a cinder from all that absorbed heat. But not too surprisingly, it isn't! What are we missing?
The air doesn't just absorb heat, it also loses it as well! The atmosphere isn't just absorbing IR Radiation (heat) from the surface. It is also radiating IR Radiation (heat) to Space. If these two heat flows are in balance, the atmosphere doesn't warm or cool - it stays the same.
Lets think about a simple analogy:
What might we do to increase the water level in the tank?
We might increase the speed of the pump that is adding water to the tank. That would raise the water level. But if the pump is already running at nearly its top speed, I can't add water any faster. That would fit the 'It's Saturated' claim: the pump can't run much faster just as the atmosphere can't absorb the Sun's heat any faster
But what if we restricted the outlet, so that it was harder for water to get out of the tank? The same amount of water is flowing in but less is flowing out. So the water level in the tank will rise. We can change the water level in our tank without changing how much water is flowing in, by changing how much water is flowing out.
Similarly we can change how much heat there is in the atmosphere by restricting how much heat leaves the atmosphere rather than by increasing how much is being absorbed by the atmosphere.
This is how the Greenhouse Effect works. The Greenhouse gases such as carbon dioxide and water vapour absorb most of the heat radiation leaving the Earth's surface. Then their concentration determines how much heat escapes from the top of the atmosphere to space. It is the change in what happens at the top of the atmosphere that matters, not what happens down here near the surface.
So how does changing the concentration of a Greenhouse gas change how much heat escapes from the upper atmosphere? As we climb higher in the atmosphere the air gets thinner. There is less of all gases, including the greenhouse gases. Eventually the air becomes thin enough that any heat radiated by the air can escape all the way to Space. How much heat escapes to space from this altitude then depends on how cold the air is at that height. The colder the air, the less heat it radiates.
(OK, I'm Australian so this image appeals to me)
So if we add more greenhouse gases the air needs to be thinner before heat radiation is able to escape to space. So this can only happen higher in the atmosphere. Where it is colder. So the amount of heat escaping is reduced.
By adding greenhouse gases, we force the radiation to space to come from higher, colder air, reducing the flow of radiation to space. And there is still a lot of scope for more greenhouse gases to push 'the action' higher and higher, into colder and colder air, restricting the rate of radiation to space even further.
The Greenhouse Effect isn't even remotely Saturated. Myth Busted!
Basic rebuttal written by dana1981
Update July 2015:
Here is a related lecture-video from Denial101x - Making Sense of Climate Science Denial
Last updated on 7 July 2015 by pattimer. View Archives
stealth @224, given that it is the Top Of Atmosphere radiative forcing that we are discussing, the proper comparison is not with the back radiation (which is of secondary importance) but with the Outgoing Longwave Radiation (OLR). Granted that 1.3 W/m^2 is just 0.5% of the OLR, but then, just 0.5% of the Global Mean Surface Temperature (GMST) is 7 C. Percentages without perspective are not very informative here.
To calculate the temperature impact of a given radiative forcing prior to any feedbacks, you must recognize that a positive radiative forcing represents a reduction in the OLR. In order to restore the TOA energy balance, and assuming no feedbacks, the OLR must be restored to its original value. That requires an increase in the effective temperature of radiation to space. Assume that 240 W/m^2 OLR is required for the energy balance, then the effective temperature or radiation to space must be (240/(5.67x10^-8))^0.125, or 255 K. A radiative forcing of 1.3 W/m^2 then, reduces the effective temperature to 254.7. Consequently a 0.3 K increase, ignoring feedbacks, is required to restore radiative balance.
For the full 3.7 W/m^2 from a doubling of CO2, the reduction in effective temperature is 1 K, and hence a 1 K increase is required to restore radiative balance, ignoring feedbacks. Finally, because atmospheric temperatures within the troposphere are locked together by convection so as to follow the lapse rate, any change in temperature at the top or middle of the troposphere the results for the need to restore radiative feedback will result in a change in surface temperature of the same size. After that occurs, the increase in back radiation will be larger than the radiative forcing, but the energy balance at equilibrium will still be neutral because heat transfer from the surface by convection and latent heat will increase to make up the difference.
Finally, the most recent surface and TOA energy balance diagram is from Stevens et al 2012:
Also, just looking at the change in temperature from increased CO2 isnt that meaningful. You cannot change temperature without also invoking the water vapour feedback. Calculating the other feedbacks is complex (hence the range in estimates for climate sensitivity) but Planck feedback plus water vapour feedback should be the baseline for considering the effects on increased CO2.
scaddenp @227, it is incorrect to think of the water vapour feedback as a singular factor. To illustrate this, consider the procedure for estimating the planck response plus water vapour feedback using Modtran. I will do so just using the 1976 US Standard Atmosphere with no clouds for illustrative purposes. To do it properly, you should do it for a representative sample of environmental and cloud conditions, and take a weighted average, something it is not strictly possible to do with the University of Chicago Modtran model due to the limited number of environmental conditions specified. Bearing that caveatte in mind, however, we proceed as follows:
Now, at this stage we may want to calculate the snow albedo feedback to the Planck response plus WV feedback to the Planck response. That will again increase the offset temperature required, which will inturn result in another round of WV responses, and a further reduction in snow cover and so on.
It is because feedbacks iterate like this that it is not correct to talk about the WV feedback as a singular factor. Supose, for example, that the total cloud feedback were slightly negative rather than (as is more likely) positive. Then the total WV feedback will be less. On the other hand, if the snow albedo feedback is stronger than expected (as is known from observation), that will result in a stronger WV feedback.
Because feedbacks interact in this way, I think it is conceptually better to determine the Planck response, and then determine the feedback factors as a group to the extent that is possible.
Tom, I do realise that H2O isnt that straightforward a feedback - especially if you take into account clouds, (I've worked through the excellent series at SoD ) but my understanding from th IPCC reports is that uncertainties with GHE of water vapour are in the "well understood" category with good agreement between theory and experimental/observational data. (unlike say clouds, ice sheet loss, clathrates etc) Ignoring clouds, I understand the effect to be effectively double planck response. For that reason, I think claims of "only" 1K for double CO2 are particularly spurious. You can argue about the feedbacks from clouds and melting ice, and especially ocean saturation and methane release, but you cant argue too much about the water vapour.
scaddenp @229, it is not that the water vapour feedback is not well understood. Rather, because feedbacks are responses to warming of cooling, other feedbacks which also warm (or cool) will also result in an additional WV feedback response. Therefore you cannot quantify the WV feedback without quantifying all feedbacks. The IPCC recognize this. They quantify the WV feedback to the Planck response alone, but note that "... because of the inherently nonlinear nature of the response to feedbacks, the final impact on sensitivity is not simply the sum of these responses. The effect of multiple positive feedbacks is that they mutually amplify each other’s impact on climate sensitivity." Consequently, while the WV+Lapse Rate feedback increases the temperature response by 50% of the Planck response ignoring other feedbacks, their total contribution to climate sensitivity will be greater than 0.5 C.
Ignoring cloud feedbacks, the IPCC indicates the other feedbacks will result in an increase of temperature of 1.9 C for a doubling of CO2. If the cloud feedback would increase the temperature response by 50% by itself, then the final climate sensitivity will be > 2.85 C, with a combined WV, Lapse rate and Surface Albedo feedback greater than 0.9 C. If, however, it is -10%, the resulting climate sensitivity will be less than 1.71 C, and the contribution of the non-cloud feedbacks will be less than 0.9 C.
Hello,
I have a question that I was hoping might be answered here. I've read through the comments and admit that most of what is being discussed are not things I understand well. It seems that the results of difference spectra reported by Harries et al. are a smoking gun. IR measurements from space over time provide concrete, easy to interpret proof that the composition of the atmosphere has changed with time in such a way that more IR is captured.
In trying to understand the methodology better, I came across this more recent publication by the same author using the same approach. It included data from another satellite in 2003. Here is the result:
The paper states that "The CO2 band at 720 cm-1 ... shows some interesting behavior, with strong negative brightness temperature difference features for 1997-1970 ... whereas, the 2003-1997 ... shows a zero signature." (Edited for clarity relative to my question--the essence of it is captured)
The "expanation" offered in the paper is essentially that there most be some compensating effect since it is known that CO2 concentrations increased between 1997 and 2003.
I'm willing, in my ignorance, to grant that that's true. However, I wonder, if the presence of a difference between 1970 and 1970 is seen as proof that CO2 isn't saturated, why is zero difference between 1997 and 2003 not powerful evidence that it is? It just seems to me that if the former evidence is enough to make one feel sure CO2 is absorbing more that the later evidence should convince the same person that CO2 is not absorbing more (between those dates).
Any insight would be appreciated!
[RH] Fixed image width.
Sorry for my various typos. I should have read it over before submitting!
basnapple @232, I have difficulty reconciling your description of the "explanation" by Griggs and Harries with that which they offer in the preprint version of their paper.
Specifically, in that version they show (Fig 3 a&b) that two popular reanalysis products do not predict the observed changes in OLR. They also show, however, that there is a profile constrained by observations that does predict the observed changes in OLR (Fig 3 c). That means the observed changes in OLR are consistent with the expectations of radiative physics plus observed changes in gas and temperature profiles, even though the observations of those profiles is of insufficient resolution to permit accurate prediction of these small changes in OLR. As they put it,
This contrasts sharply with your claimed "explanation", which is of course no explanation at all. Converting a claim that changes in OLR lie within those expected given known limits of observation, and hence that there is no discrepancy, to a claim that a discrepancy exists for which there is no explanation is very substantial. I doubt that editorial review would have forced so large a change on the paper. Nevertheless I ask that you quote the original sections of the paper as published to show that you have indeed fairly represented Griggs and Harries.
Hello Tom,
Thanks for the reply. I read through the preprint as best I could. I am unfamiliar with the models used and some of the terminology. Is it true that, in their analysis, the temperature profile of the atmosphere is the fitting parameter?
As far as how well I represented their claims, I will quote with some (perhaps unnecessary) context:
First quote, commenting on the difference spectra figure I included in my previous post:
"An initial inspection indicates that the processing of the data has not caused any major artifacts. In all cases the difference spectra are seen to have consistent and reproducible
features. The only sign of asymmetry (which could indicate a mismatch of wavenumber scales between the spectra) is in the CO2 (0110 → 1000) band at 720 cm-1, which may be due to its position on the very steep high frequency wing of the CO2 fundamental centered
at 667 cm-1."
Second quote, in reference to the same:
"A negative brightness temperature difference is observed in the CO2 band at 720 cm-1 in the IMG–IRIS (1997–70) and the AIRS–IRIS (2003–1970) difference spectra, indicating increasing CO2 concentrations, consistent with the Mauna Loa record (Keeling et al. 1995). However, this channel in the difference is also sensitive to temperature, and we note that in the 2003–1997 difference, despite a growth in CO2 between these years, there is no signal at 720 cm-1."
Third quote is where the portion in my previous post comes from, now with context:
"The CO2 band at 720 cm-1, though asymmetric for the reasons stated earlier, nevertheless shows some interesting behavior, with strong negative brightness temperature difference features for 1997–1970 and 2003–1970: whereas, the 2003–1997 (a much shorter period, of course) shows a zero signature. Since we know independently that the CO2 concentration globally continued to rise between 1997 and 2003, we must conclude that the 2003–1997 result must be due to changes in temperature that compensate for the increase in CO2. This would mean a warming of the atmosphere at those heights that are the source of the emission in the center of this band. This is somewhat contrary to the general (small) cooling of the stratosphere at tropical latitudes." (emphasis mine, I'm only trying to show where my "explanation" came from)
Fourth quote, regarding differences between model results and observations:
"Finally, there exists a marked gradient in the simulated spectrum between 800 and 700 cm-1, which is absent in the observations. This coincides with the far wings of the strong CO2 band centered at 667 cm-1. In sensitivity tests, this gradient showed sensitivity to the amount of CO2, and is therefore related to the strong CO2 band, and may reflect reanalysis uncertainties in temperature."
The paper includes appendices detailing the temperature profiles used to coerce the models to the data.
I would appreciate some help digesting this.
basnappl @234, first a correction. I thought the copy of Griggs and Harries 2004 was a preprint of the paper you were looking at. In fact, you were looking at Griggs and Harries 2007, which is a seperate (although related) paper. Based on that, my comment @233 is correct so far as it goes. That is, observations of changes in OLR match those predicted by models, within observational limits of ghg, temperature and H2O profiles. That is, the slight discrepancy you have pointed out results from our limited knowledge conditions within the atmosphere rather than any failing of Line By Line (LBL) or Band models of radiative transfer.
One thing we do know is that those radiative transfer models are extraordinarilly accurate. This is shown by the match between one particular model and observations shown in the graph below:
(For more examples, see my discussion here and my article here.)
Because of these tests of accuracy when the conditions are well known, and because radiative transfer models are based on very well known physics, the discrepancy you point to is almost certainly the result of atmospheric conditions rather than deficiencies in the radiative transfer models. That being the case, there is no question of the greenhouse effect being saturated, for all radiative transfer models predict the logarithmic relationship between CO2 concentration and radiative forcing, ie, that you get approximately the same increase in forcing for each doubling of CO2.
It is interesting, however, to look at the reason for the "zero signature" between 1997 and 2003. Griggs and Harries are more explicit in 2007 than in 2004, attributing the lack of signature to temperature profile. As it happens, using the lapse rate it is possible to use the brightness temperature as a rough indicator of altitude. If we do so, we see that in order to have zero influence in the band in question, the increase in CO2 between 1997 and 2003 would need to be matched by an increase in temperature at 7-10 km altitude greater than that at the surface, and that expected by the models. In other words, it appears that Griggs and Harries have found a tropospheric hotspot.
KR @222. I went and read Myhre 1998 that produced the ΔF = 5.35*ln(C/C0) W/m2 estimate for CO2 forcing. The issue I have with this is that the alpha coefficient of 5.35 is derived from three different models, which assume that the models are accurate relative to the global atmosphere.
Being a software modeler myself (for 30+ years) dealing with RF energy through atmosphere, I understand that what is measured in the lab, what is model in software, and what the real world does are almost always very different. The real world is so noisy and chaotic that I have found models are almost useless in predicting what will really happen in the real world. I would be stunned if this is not also true for this forcing equation, and GCMs in general.
Trying to measure this value for the real world, on average, is probably impossible given that the atmosphere is so different moment to moment and place to place, and changes in long term trends may be hard to determine since we have so little empirical measurement data. So while this is the “best go-to reference” we have, I doubt it is realistic or correct relative to what is really happening in the real world. I admit it *might be* correct, but I cannot prove or disprove it, nor can anyone else. This isn’t meant as a criticism of experts in this field, only a realization of my experience that the atmosphere is impossible to model accurately.
Tom Curtis @226: I think you made a minor math error. If the average global temperature is 15C, or 288K, then 0.5% of this is 1.44K. I love the energy balance diagram by Stevens et al 2012. I went and read the paper and I found it quite interesting. Since Kevin Trenberth has generated a topic on the energy budget (http://www.skepticalscience.com/news.php?n=865), I am going to take my questions and comments about the energy budget over there.
Scaddenp @227. Just because warmer air can hold more water vapor, doesn’t mean it will. Charts I have seen (http://www.climate4you.com/GreenhouseGasses.htm) of humidity for the atmosphere over time has been about the same at low levels, but at higher altitudes relative humidity has been decreasing. If these charts are correct, this seems to suggest that water vapor has not increased as the atmosphere has warmed over this time period (65 years).
[TD] For your comment on water vapor, please read the counterargument to "Humidity is Falling," and if you disagree with the peer-reviewed evidence presented there, please comment there. Regarding water vapor in the stratosphere, see "What is the role of stratospheric water vapor in global warming?" Anyone who responds to Stealth's comment here about water vapor, please, please do so on those other threads, not here. Everybody must help to keep the conversations on the appropriate threads. Thank you.
[TD] Thank you for recognizing that another post is the right place to talk about energy budget!
Stealth - Regarding Myhre 1998:
"I doubt it is realistic or correct relative to what is really happening in the real world."
You would be wrong. Those model results have been proven out, empirically measured by the satellite observations, such as those discussed in the opening post (Harries et al 2001 in particular). Have you read the opening post of this thread?
Yes, the Myhre results are based on numeric models of radiative absorption/emission - using column estimates from three multiple latitudes, three different models to minimize bias and atmospheric variation. And they have been confirmed - the satellite spectra show the same outgoing radiation as predicted by those models. And therefore the modelling of slightly different atmospheric compositions is trustworthy. There is really no doubt about them, no significant uncertainties in direct forcing calculations.
If your model reproduces observations from basic physics, it's a good model. Your issues about uncertainties are unwarranted.
KR @237 To be honest, I’m fairly stunned by your response and I am not sure how to address it. Just out of curiosity, what is your background?
You seem to be asserting that the alpha coefficient of 5.35 of the CO2 forcing function is both accurate and precise because it has been empirically measured. That cannot be true, otherwise people would not be building models as a way to attempt to arrive at these values. Correct? Why build a model when you can just measure it. Your assertion that this is basic physics and models match the real world simply cannot be true. I understand physics (I have a physics degree) and I build software models for living (I also have a computer science degree) so I think I am qualified to speak to models and physics.
The climate is not that simple – far from it – so my assertion of uncertainty is, I believe, completely accurate and true. As further evidence to “prove” that there is enormous uncertainty in the climate, just read the TOA energy balance paper Stevens et al that was referenced by Tom Curtis @ 226. This is great paper! It is peer reviewed. At the end it states: “The net energy balance is the sum of individual fluxes. The current uncertainty in this net surface energy balance is large, and amounts to approximately 17 Wm–2. This uncertainty is an order of magnitude larger than the changes to the net surface fluxes associated with increasing greenhouse gases in the atmosphere.”
Think about that – the uncertainty in the energy budget is ten times larger than the fluxes associated with GHGs. This is clearly proof that my assertions of uncertainty are completely warranted.
@StealthAircraftSoftwareModeler:
Out of curiousity, which climate models have you analyzed in depth?
Stealth - Perhaps you should re-read just what you have quoted:
Since what we are discussing WRT Myhre 1998 are radiative transfer codes, and the change in forcings due to changes in atmospheric composition, we are indeed speaking of the 'net surface fluxes' which have far lower uncertainties. You seem to be conflating uncertainties in accounting for multiple energy flows into a total budget with uncertainties in computing atmospheric spectral response - applying an entire collection of uncertainties to a tiny portion of the puzzle.
If you wish to discuss the total energy budget, the sum of individual components (and their uncertainties) of the energy budget, there is an appropriate thread. However, the radiative transfer codes are well proven, giving results within under 1% of observations (Chen et al 2007) including dealing with compositional changes - arguing any significant uncertainties in that regard (as you have) is quite frankly unsupportable.
Stealth:
Appeals to personal qualifications and arguments from incredulity such as on display in #236 and #238 are not terribly convincing.
All this:
strikes me as practically equivalent to:
although I am sure it was not your intent to communicate such sentiments.
(I have highlighted in the quotes from your comments the three words that are often the butt of jokes on medical blogs: "in my experience" or variants there of are sometimes called "the most dangerous words in medicine". I see no reason why this maxim should not generally be applicable to other scientific domains, particularly when the person asserting it is arguing against the weight of evidence, as you are.)
Stealth - you say. "That cannot be true, otherwise people would not be building models as a way to attempt to arrive at these values."
I think you are confusing different models here. I dont think anyone is doing much on work on refining the 5.35 value from Myhre. As pointed out, the uncertainities are low and matches observation.
By comparison global GCMs are trying to model what will be climate response to this deltaF. These models do have significant uncertainities resulting in varying estimates for climate sensitivity from 2-4.5. Dont confuse the difficulties with modelling climate response with the modelling required for calculating the forcings. Different models.
ps. After getting the full picture on humidity from the link indicated, I'd be interested in your assessment of tactics used by Climate4you to mislead.
Stealth
A very important distinction needs to be made and clarified here. The models that are being referred that calculate the 5.35 ln(C/C0) result are not climate models! They are Radiative Transfer Codes; solutions to the equation of Radiative Transfer. As such what they do is, given a known state for a column of gas - temperature, pressure, humidity and composition profiles - they calculate the instantaneous radiative state at any point in that column. As such, the underlying maths is actually relatively simple. And they work from databases of very well established spectroscopic data. There are no assumptions or time based modelling or projections, they calculate a single snapshot.
A bit like engineering stress analysis programs, where they do the same simple calculations many times over for small cells to build up the composite picture. And their results are used in a wide range of applications, Climatology is only one of them. They are used in astronomy, military,, satellite communications modelling, a whole host of different domains. And their results have been extensively tested in the field and in the lab.
John Hartz @239: I would love to examine the source code of some GCMs to see if Dr. Freeman Dyson’s claim that GCMs are full of fudge factors is true. I suspect that it is true because software modelers always have to make assumptions and design trades in order to get software to run in a reasonable amount of time. Can I get the source to any of the GCM models? I doubt that I can, but it is worth a try. I know there are a dozen so models, and if source code is available, which would you recommend I look at? I don’t have time to look at them all, and they give widely different projections based on the spaghetti graphs I’ve seen, so I’d only like to see the one that is considered the best.
Source code is available for many of them. See here for GISS ModelE. Weather is chaotic so the same climate model will different wiggles for different initialisations. They dont pretend to be able to predict weather. 20-30 years are what climate is about. There is a very useful article on interpretation here.
Understanding the real differences between different modelling approache is what CMIP5 (and its predecessors) is about.
Stealth, the fancy computer models merely fine tune the basic projections that have turned out to be pretty accurate, starting in the 1800s, and the early ones certainly did not involve computer code because computers had not been invented yet. You can try some of those simple models yourself by getting an introductory textbook such as David Archer's Global Warming: Understanding the Forecast, or by taking notes while watching his free online lectures from his class at the University of Chicago.
Tamino has illustrated a simple climate model you can run without a computer if you have a lot of time, or with a spreadsheet if you don't mind using a computer. He also has a followup that's only a bit more complicated.
There are a bunch of other climate models that are simple enough for learning and teaching. One list has been compiled by Steve Easterbrook.
Code for slightly more complex or narrow models also is freely available. RealClimate's "Data Sources" page has a handy but short list. Even the full-blown General Circulation Models (GCMs) have freely available code. RealClimate's Data Sources page also has a handy list of links to those codes. Steve Easterbrook has a three-year old list of GCMs with links to whatever info he could find about getting their codes. See also Tamino's Climate Data Links.
Stealth - RealClimate has links to both climate data and a number of model codes here, including GCMs and others.
"...GCMs are full of fudge factors..."
Um, no. They are full of physics. There are parametric approximations of small-scale phenomena, and for limitations of sampling and scale - but those are anchored in physical measurements, they are not "fudge factors" or tuning knobs for giving a specific result. Temperature projections and climate sensitivity are outputs of the models, not inputs, and a great deal of the variation between individual model runs comes from differing initial conditions. That variation is in fact part of the results, indicating to some extent the range of potential weather we might see around climate trends. See scaddenp's link above for more discussion.
You do, I hope, realize that "fudge factor" claims are essentionally accusations of fraud aimed at the model authors? And unsupportable ones, to boot?
Stealth,
I don’t have time to look at them all, and they give widely different projections based on the spaghetti graphs I’ve seen, so I’d only like to see the one that is considered the best.
This demonstrates a fundamental misunderstanding of what climate models are doing.
Weather is chaotic. The timing of even signficant events like El Nino/La Nina cannot be predicted years into the future. In order to distingish between long term climate change, and the effects of internal variability, it is essential that repeated runs of the same climate model exhibit different realisations of that internal variability. This allows them to be averaged together so that the random variations cancel out leaving behind the systematic changes that will dominate in the longer term.
Even then, there is a risk that the forecasts of individual climate models inadvertently contain systematic biases that won't be cancelled out by this procedure due to the choices that were made in designing them and potentially even software bugs. So different models from completely different groups are also combined, to see what is common in their forecasts and what varies between them. Given their varied nature, if they all predict the same thing to within some tolerance, then we can have a certain amount of confidence that the true answer lies within that range; if they disagree about something, then our confidence is reduced.
Of course, since they are all embodying known physics to varying degrees, they will all contain systematic biases towards "reality, as we understand it" in that regard.
Anyway, far from being an indication of failure, those spaghetti graphs are an essential element of determining the reliability of the forecasts and trying to predict the range of possible outcomes.
Can I get the source to any of the GCM models? I doubt that I can, but it is worth a try.
Perhaps you should check a little bit harder before beginning to "doubt"?
Stealth - "Can I get the source to any of the GCM models? I doubt that I can..."
From the first two pages of a Google search on "code for climate model":
Not too difficult to find...