From this NASA press release I’ll have more on this later. The timing of this release is interesting.
› Play animation (Quicktime) | › Play animation (Windows Media Player)
› Related images and animations
WASHINGTON – Researchers studying carbon dioxide, a leading greenhouse gas and a key driver of global climate change, now have a new tool at their disposal: daily global measurements of carbon dioxide in a key part of our atmosphere. The data are courtesy of the Atmospheric Infrared Sounder (AIRS) instrument on NASA’s Aqua spacecraft.
Moustafa Chahine, the instrument’s science team leader at NASA’s Jet Propulsion Laboratory, Pasadena, Calif., unveiled the new product at a briefing on recent breakthroughs in greenhouse gas, weather and climate research from AIRS at this week’s American Geophysical Union meeting in San Francisco. The new data, which span the seven-plus years of the AIRS mission, measure the concentration and distribution of carbon dioxide in the mid-troposphere–the region of Earth’s atmosphere that is located between 5 to 12 kilometers, or 3 to 7 miles, above Earth’s surface. They also track its global transport. The product represents the first-ever release of global carbon dioxide data that are based solely on observations. The data have been extensively validated against both aircraft and ground-based observations.
“AIRS provides the highest accuracy and yield of any global carbon dioxide data set available to the research community, now and for the immediate future,” said Chahine. “It will help researchers understand how this elusive, long-lived greenhouse gas is distributed and transported, and can be used to develop better models to identify ‘sinks,’ regions of the Earth system that store carbon dioxide. It’s important to study carbon dioxide in all levels of the troposphere.”
Chahine said previous AIRS research data have led to some key findings about mid-tropospheric carbon dioxide. For example, the data have shown that, contrary to prior assumptions, carbon dioxide is not well mixed in the troposphere, but is rather “lumpy.” Until now, models of carbon dioxide transport have assumed its distribution was uniform.
Carbon dioxide is transported in the mid-troposphere from its sources to its eventual sinks. More carbon dioxide is emitted in the heavily populated northern hemisphere than in its less populated southern counterpart. As a result, the southern hemisphere is a net recipient, or sink, for carbon dioxide from the north. AIRS data have previously shown the complexity of the southern hemisphere’s carbon dioxide cycle, revealing a never-before-seen belt of carbon dioxide that circles the globe and is not reflected in transport models.
In another major finding, scientists using AIRS data have removed most of the uncertainty about the role of water vapor in atmospheric models. The data are the strongest observational evidence to date for how water vapor responds to a warming climate.
“AIRS temperature and water vapor observations have corroborated climate model predictions that the warming of our climate produced as carbon dioxide levels rise will be greatly exacerbated — in fact, more than doubled — by water vapor,” said Andrew Dessler, a climate scientist at Texas A&M University, College Station, Texas.
Dessler explained that most of the warming caused by carbon dioxide does not come directly from carbon dioxide, but from effects known as feedbacks. Water vapor is a particularly important feedback. As the climate warms, the atmosphere becomes more humid. Since water is a greenhouse gas, it serves as a powerful positive feedback to the climate system, amplifying the initial warming. AIRS measurements of water vapor reveal that water greatly amplifies warming caused by increased levels of carbon dioxide. Comparisons of AIRS data with models and re-analyses are in excellent agreement.
“The implication of these studies is that, should greenhouse gas emissions continue on their current course of increase, we are virtually certain to see Earth’s climate warm by several degrees Celsius in the next century, unless some strong negative feedback mechanism emerges elsewhere in Earth’s climate system,” Dessler said.
Originally designed to observe atmospheric temperature and water vapor, AIRS data are already responsible for the greatest improvement to five to six-day weather forecasts than any other single instrument, said Chahine. JPL scientists have shown a major consequence of global warming will be an increase in the frequency and strength of severe storms. Earlier this year, a team of NASA researchers showed how AIRS can significantly improve tropical cyclone forecasting. The researchers studied deadly Typhoon Nargis in Burma (Myanmar) in May 2008. They found the uncertainty in the cyclone’s landfall position could have been reduced by a factor of six had more sophisticated AIRS temperature data been used in the forecasts.
AIRS observes and records the global daily distribution of temperature, water vapor, clouds and several atmospheric gases including ozone, methane and carbon monoxide. With the addition of the mid-tropospheric carbon dioxide data set this week, a seven-year digital record is now complete for use by the scientific community and the public.
Animation of the 3-D transport and distribution of water vapor as measured by AIRS from June through November 2005. Image credit: NASA › Play animation (Quicktime) | › Play animation (Windows Media Player)
enlarge imageFor more on AIRS, see http://airs.jpl.nasa.gov/ .
Ryan Stephenson says:
It doesn’t actually matter in a positive feedback system how big the positive feedback is – because the effect will simply rotate around the feedback system getting bigger and bigger until it was first noticeable and ultimately disastrous.
No…When climate scientists talk about a net positive feedback, they are including the case where the feedback is such that the response is amplified by a finite amount. The idea is one of a converging geometric series, such as 1 + 1/2 + 1/4 + 1/8 + … The sum of such an infinite series is not infinite…It is two, meaning that a water vapor feedback behaving in this manner would amplify the warming by a factor of 2.
Note that climate scientists talking of a positive feedback in this way are defining feedback in a somewhat different way as it is apparently defined in system control theory (or whatever one calls it). If you want to adopt the “control theory” point of view, you have to consider the increased radiation as the earth warms that is described by the Stefan-Boltzmann Equation as a feedback and, indeed, if you do then when climate scientists talk about a positive feedback, you have to convert that to “positive feedback not counting the radiative feedback implied by the S-B Equation but still a negative feedback overall when you do”. Some climate scientists have adopted terminology more compatible with the control theory point-of-view, for example, Dennis Hartmann in his book “Global Physical Climatology” (see p. 231 here: http://books.google.com/books?id=aKPxctcJNNUC&printsec=frontcover&dq=dennis+hartmann+global+physica+climatology&cd=1#v=onepage&q=&f=false ).
However, the important point is that, whether you think the choice that many climate scientists have made in using the term “feedback” is bad or good, it doesn’t affect the actual physical result. It is just a matter of definition: Climate scientists like to think of the “base case” as the temperature rise that would be implied by considering the Stefan-Boltzmann Equation and to ask what the temperature rise is relative to that, whereas a viewpoint in line with system control theory would consider the radiative feedback implied by the Stefan-Boltzmann as a negative feedback on the original change in radiative balance.
Unfortunately, your logic is based on your misunderstanding of how climate scientists use the term “positive feedback”. And, in fact, the paleoclimate record, including the glacial-interglacial cycles, are hard to explain without invoking a positive feedback (in the sense that many climate scientists use it, meaning one that amplifies the climate sensitivity relative to the value implied by the radiative forcing and the S-B Equation).
The logic is fine, although one might quibble with the terminology that climate scientists have adopted. Issues of terminology are often confusing…I just had lunch with someone today who was complaining about the awful choices that he thought physicists had made in terminology, such as using “specific heat” to refer to something that is not heat (He thought it should be called “specific heat capacity”) and calling lots of things constants, like “optical constant” or “dielectric constant” that can in fact vary considerably as a function of wavelength, temperature, or what have you.
I made a formatting error in my previous post: The paragraph starting “It doesn’t actually matter in a positive feedback system how big the positive feedback is…” should have been included in the quote from Ryan. They are his words, not mine.
I’m puzzled. Heterogeneous CO2 distribution would imply that, after factoring out solar irradiance, a heterogeneous temperature field. But I can’t find that data, only a high-level summary of consequences. I’m still trying to understand the data, but I’m suspicious
Joel Shore (18:58:52) :
So, the thing I think everyone is trying to understand here is why this process needs anthropogenic CO2 to begin. Wouldn’t this process be at play already in a warming world? Which leads to the question why would water vapor not have already instigated this feedback cycle before anthropogenic carbon dioxide became an issue in the atmosphere? Water vapor is a GHG, after all, so it should exhibit the same influences as CO2, albeit on a grander scale.
Joel Shore
The models are still models of the theory which have not been validated against our real planet. I’ve always found it interesting that your answer to any arguments, and there is no question you know your stuff, are always repetition of aspects of the theory. I think that even you will admit that even if some features have a valid mathematical basis, some of the numbers used may be off the mark thus generating outcomes not consistent with the real earth. Perhaps even more than an initial values problem.
And where have you been the last few weeks. The CRU mess puts a lot of the models problems in some perspective. How do the models when backcasting address the MWP and the LIA? If they don’t ‘understand’ natural variation in the past, how could they possibly understand the 20th century let alone the future. And further, while I felt there has been some fudging in the global temp records, I still believed the arc of the warming was mostly valid. Now I’m not so sure of even that.
It’s no longer business as usual and not safe to pretend it is. Thus intelligent, articulate, expression of the theory is simply not good enough anymore without the recognition that something is rotten somewhere and we don’t know where.
cool, more neat toys to play with! XD
Joel Shore,
You state:
“In fact, since much of the same physics of the transport of water vapor controls both the water vapor and lapse rate feedbacks, climate models tend to show considerably less spread in their prediction of the sum of these two feedbacks together than they do for each of the feedbacks separately.”
The models have considerable divergence with each other when the water vapor and lapse rate feedbacks are considered separated. However, the important point is that the convergence when the two are combined is also just with each other, not with the climate. As already discussed, the Wentz paper points out that the models reproduce less than one half of the precipitation increase observed with the recent warming. In other words, they under represent the negative wet lapse rate feedback.
The convergence of the models is thus just wrong agreement. Lambert of the Hadley Center and Stine, Krakauer and Chiang of UC Berkely write: “Thus if GCMs do underestimate global precipitation changes, the simulation of other climate variables will be effected.” Eos Vol 28 No. 21
In the same issue of Eos, Previdi and Liepert explain: “This non-radiative energy transfer takes primarily the form of latent and sensible heat fluxes with the latent heat flux being about 5 times larger than the sensible heat flux in the global mean. The latent heat flux from the surface to the troposphere is associated mainly with the evaporation of surface water. When this water condenses in the troposphere to form clouds and eventually precipitation, the troposphere heats up and then radiates this energy gain out to space. The radiative energy loss from the troposhere is equal to the energy heat gain at the surface. The global water cycle is therefore fundamentally a part of the global energy cycle and any changes in global mean precipitation and evaporation are consequently constrained by the energy budgets of the troposhere and surface.”
As Wentz stated in the journal Science: “The difference between a subdued increase in rainfall and a C-C increase has enormous impact, with respect to the consequences of global warming. Can the total water in the atmosphere increase by 15% with CO2 doubling but precipitation only increase by 4% (1)? Will warming really bring a decrease in global winds? The observations reported here suggest otherwise, but clearly these questions are far from being settled.”
It is unfair to criticize the IPCC based upon diagnostic work published since the AR4 report, however, there was plenty of diagnostic literature available to the authors at the time that should have given them pause, since the documented errors were larger than the phenomenon of interest. With the developments since the TAR, they were justified in assuming the models were much improved, but that development was accompanied by improved diagnostic studies, that should have made them less confident than the authors of the TAR were in their ignorance. Instead, the AR4 authors in denial of concerns raised in the draft reviews had the hubris to declare themselves more “confident”.
Keep in mind that there are two separate aspects of airborne water — vapor and clouds. Water vapor is a greenhouse gas, but clouds aren’t — they’re actually aerosols & act like grey-body particles. So there’s a distinct difference in radiational characteristics.
Positive water vapor feedback would occur during warming/cooling from any cause — solar, CO2, ocean-current changes, etc. And it occurs in both directions, magnifying temp increases or declines. But it could be overridden by the resulting clouds, depending on when they occur (day or night). From my experience, clouds in fact do have a daytime bias — observe any warm-season day here in the US or almost any tropical region. Max cumulus cloud cover occurs in late-afternoon or early evening and a min during the night generally. So the total water (clouds and vapor) temp effect may be near zero or even negative, tho warmer temps should cause more convective precipitation.
“Time to concede the “observations don’t back the models” meme.” ”
Models predict a 30% increase in CO2 yields a 1C warming, and the observation of the last 70 years has been more like 0.3C warming (if the temp records are honest).
Time to concede it back.
Dave F says:
Yes…the water vapor feedback operates in the absence of the CO2 rise and, in fact, if it is turned off in climate models, these models have trouble reproducing the natural variability seen in the real climate system ( http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.143.6379&rep=rep1&type=pdf ):
I see Joel Shore is at it again.
The ghg-h2o vapor connection is one of temperature. There are some interesting limits imposed on this that are not supportive of gcm claims. One of the claimed measured results is that on average relative humidity remains constaint and this is really not the good news for the CAGW crowd that many think it is.
Like co2, h2o vapor has a log effect which means roughly the same increase or decrease in forcing per doubling or halving of the actual amount of vapor in the atmosphere (or section of atmosphere). Like co2, h2o is well saturated and has many times the low concentration where it would no longer have a similar effect for halving again. While co2 has an attenuation or forcing increase per doubling of about 3.5 w/m^2 as would be measured at 70km altitude, h2o vapor would be about 2 to 3 times – 8-10 W/m^2 that (all this assumes clear sky and uses a simple 1-d strictly radiative model that uses actual physics and no kludges or fudge factors).
Making the assumption that a 5 deg C rise in temperature occurs, ostensibly due to a doubling of co2 and positive feedback from h2o vapor, we find that the increase in h2o vapor content (with the assumed constant RH) results in a 1.3 x increase. This is well under a doubling of h2o. The actual calculated h2o forcing is about 3.1W/m^2 in addition to the 3.5 W/m^2 of co2 which means the h2o feedback alone even for a 5 deg C rise will be less than the contribution of co2 alone. The co2 contribution without feedback is considered to be around 1 deg C. A simple average value for sensitivity suggests that the Earth has a 33 deg C rise for about 150 W/m^2 total ghg forcing or 0.22 deg C per W/m^2 forcing change that causes 0.22 x 3.5 = 0.77 deg C and the h2o would add about 0.68 giving a total of just under 1.5 deg C. This is a far cry from the necessary 5 deg C predicted by some gcm models or from our original assumption. If one assumes a 2 deg C rise, the same thing happens except that the h2o vapor contribution is only a small fraction of the co2 contribution and we are still missing a substantial portion of the forcing necessary to achieve the 2 deg C rise. So far as the h2o vapor contribution goes, it’s going to be insignificant compared to whatever else is causing the T rise. That means if a 1 deg C rise happens, it’s because something else drove up the T by almost a deg C. It’s the rather linear nature of absolute humidity, relative humidity and the fact that absorption is not linear.
This leaves the CAGW crowd only with some sort of positive feedback caused by clouds to have any significant effect. What’s going on with clouds though is poorly understood but perhaps not poor enough to see serious problems with that presumption.
Several things are known concerning clouds. For one, they form after the relative humidity hits 100% to no more than 200% (supersaturated) and formation depends upon particulates and those can somewhat determine characteristics of the cloud. On average, clouds increase albedo, especially thick lower clouds. Higher ones may do little in that area and clouds will block outgoing IR far more effectively than some slight increase in co2. Clouds at night will only block IR but during the day will tend to block more incoming than outgoing. Note that cloud tops will radiate broadband IR but at a lower blackbody T than the mean surface T. Clouds, especially thicker lower ones tend to form and dissipate during the day providing a net negative feedback not positive. Also, clouds are associated with increased convection, reducing the reliance on radiative effects as the water vapor cycle transports more energy above the lower troposphere. The fact there might be a bit more IR coming down and hitting primarily the ocean results not in lower ocean heating but in increased evaporation because IR cannot penetrate the water. This increases h2o vapor (necessary to the assumption of constant RH) which increases the h2o vapor cycle since moist air weighs less than dry air and also since it has more concentrated h2o, it is warmer air having blocked more incoming and outgoing IR so there must be more transport of heat out of the low troposphere. Finally, as it reaches higher altitudes, it’s got to become supersaturated due to lower temperature values higher up. There it either has to form clouds or it’s got to condense and rain out. All of these factors point to a very strong negative feedback, not to some ultrastrong positive feedback. The fact that so many gcm outputs claim otherwise is something that should call into question just what is happening in these to cause such an apparently nonphysical answer.
[REPLY – Okay, but bear in mind that Joel Shore is welcome here, as are all points of view. I am glad to see him back. ~ Evan]
Joel Shore (17:39:26) :
Coolio, but still, where does the water vapor feedback stop? A factor of 2? 2 times what? Where are the beginnings and the ends of this cycle?
Also, if the water vapor feedback is where most of the warming is expected to come from, well, it has been overpowered in the past, why won’t it be in the next 100 years?
Dave F
One can mathematically create infinite series that converge to any particular number. Others blow up to infinity. Given any particular temperature, water vapor will have contributed a certain amount to that temperature based upon that temperature and the availability of water to become vapor. What I did above to avoid messing with an infinite series is to assume a new temperature and to determine how much the increase in water vapor would contribute to causing that increase. I also did it for a 2 degree rise which indicated a much less water vapor increase was available to contribute but I don’t recall what the calculation values turned out to be other than it was much less than the co2 contribution, with a total of around 1 to 1.25 deg C rise. In both cases, the co2 contribution alone to increasing temperature was taken to be 0.75 deg C for a doubling and in both cases, the co2 + h2o vapor contributions were less than what was required to raise the temperature by that amount and the h2o contribution in the latter was far less than that of the co2 alone.
Reducing this to crude and overly simple, in today’s atmosphere, the water vapor contribution (vapor only / no cloud effects) will result in contributing no more than 1/4 as much to a temperature increase as will the original co2 forcing change.
cba (05:55:16) :
Thanks, that was pretty helpful! I do wonder what Mr. Shore has to say about that.
There is another side of the coin also, yes? If the water vapor falls as precipitation, then there is no feedback in the other direction?
Joel Shore (18:31:44) :
“Even Roy Spencer has expressed a lot of skepticism regarding Lindzen’s results…mainly because he makes a comparison of the ERBE data to climate models run in a mode that isn’t really relevant for comparison to real data.”
“A lot of skepticism” is your interpretation. But, in any case, Spencer was speaking to the fact that Lindzen looked at AMIP, while climate forecasts are made with CMIP. However, CMIP is validated based on agreement with AMIP. The positive feedback Lindzen spoke of in AMIP is a necessary kluge, without which CMIP would not agree with AMIP over the validation period. Hence, Lindzen’s analysis is relevant, and AMIP and CMIP stand or fall together.
For persons with a control systems perspective, the proper way to look at it is an inner positive feedback loop, which normally would lead to instability, with an outer negative feedback loop which dominates (SB T^4 feedback is pretty powerful, for example). Such an inner positive feedback leads to amplification in overall response.
That is why the question is so important. One of the critical legs of AGW is the assumed positive feedback from water vapor. And, it must be significantly positive, and certainly not overall negative when cloud cover is taken into account.
Martin Lewitt: I found the EOS papers that you mentioned and it seems that your reading of them was rather selective. For example, you quote the description of the global water cycle in Previdi and Liepert but don’t bother to note the main conclusion of their paper, which is:
The discussion in F. H. LAMBERT, A. R. STINE, N. Y. KRAKAUER,
AND J. C. H. CHIANG is more complex but they also are not convinced that the difference between GCMs and observations regarding precipitation is a real discrepancy.
Oh…Here are links to those Eos articles: http://www.atmos.berkeley.edu/~zan/Site/Alexander_Stine_files/Hugo_Eos_2008.pdf
http://www.nwra.com/resumes/liepert/pdf/Previdi_Liepert_EOS_May20_08
Dave F says:
I don’t know exactly what to say about cba’s musings except that he seems to come up with different results than are found in the published literature. He should probably tell us where he is getting these number from or, if they are his own, then he should be publishing them.
”
Dave F (09:03:17) :
cba (05:55:16) :
Thanks, that was pretty helpful! I do wonder what Mr. Shore has to say about that.
—–
There is another side of the coin also, yes? If the water vapor falls as precipitation, then there is no feedback in the other direction?
”
LOL, I hope he tries. I need some more humor in these trying times.
—–
the h2o is falling as liquid or ice. At this point the energy absorbed at the surface by the vapor as it formed from liquid or ice (heat of evaporation, heat of fusion) was given up high up in the atmosphere and also the temperature will be lower too having given up that energy. Often some of the falling liquid will evaporate, again absorbing energy into a phase change rather than in an increase in temperature. Net result is a continuation of the ongoing cycle which conveys heat higher into the atmosphere where the molecules are no longer blanketed and can radiate more of their thermal energy into space. Once there is solid or liquid h2o high up, it is not constrained to radiate in the narrow discrete line spectra of a gas molecule that can then be recaptured by a nearby molecule of the same type.
But as Roy has been saying all along, the details is in the clouds, not just h2o vapor.
Joel Shore,
You were rather selective in your reading of the EOS articles. I quoted the part that pertained to the climate, however careful and complete reading of the article shows that the variability conclusion you want to quote is model variability not climate variability based:
“Figure 1 shows distributions of hydrological sensitivity for 20-year periods in the twentieth and 21st centuries based on output from eight coupled atmosphere-ocean GCMs.”
It is the skill of the models which is at issue. At best the papers excuse for the models is that they didn’t match the climate of the recent warming because they had the aerosols wrong. Whatever 20 year random variability in precipitation the models are able to simulate, the models didn’t come anywhere near close to bracketing the increase in precipitation associated with the recent warming. Given the number of models, and the fact that none were within half of the increase in precipitation, either the climate itself must be considered an unrepresentative random outlier, on precipitation but apparently not on temperature, or there must be some other errors in forcing or internal state. To the authors credit, they didn’t criticize the climate for being an unrepresentative outlier but hypothesized an issue with the aerosols and mentioned top of atmosphere radiative diagnostic issues with the models as well.
What is apparent is that the authors of both EOS articles take the models’ discrepancy with the observed increase in precipitation seriously.
The models are not yet ready to be used as “evidence”. Regional and state studies which use them to project increased risk of droughts without qualifying the results with a discussion diagnostic literature for the models should be rejected by peer reviewers, and if the authors or peer reviewers were aware of the diagnostic literature and went forward with publication, that is arguably scientific fraud, not merely negligence and incompetence.
All the talk of non-linearities and tipping points strikes me as mathematically highly unlikely . Such events must generally be of measure zero in any differentiable function . The fact that we are only about 8c warmer than the StefanBoltzmann&Kirchhoff temperature for a gray body ( the proper 0th term rather than the non-physical “cold earth” hypothesis of absorption as a gray body but emission as a black body which is where the commonly seen 30c+ purported warming comes from ) shows that there have been no substantial tipping points or non-linearities in the first 278k warming of our temperature .
Bob Armstrong: Well, maybe it’s just me, but I consider the difference between the current Rochester conditions and the condition when we were buried under a mile or two of glacial ice sheet to be kind of dramatic, not to mention the difference between the hot-house and snowball earth climates that may have existed further in the past.
Martin Lewitt: If you really think that you are reading those papers correctly, I challenge you to contact the authors directly and see if they agree with your assessment. The point of the figure caption that you quote is to illustrate that the models predict there to be large decadal fluctuations in the change in precipitation and hence that it doesn’t make sense to compare century-long integrations of the models to a twenty-year period in the real climate system. So, no, the issue is not just a question of the skill of the models. It is also a question of what sort of variability one expects to see in the climate system.
And, I don’t know what you mean about the models being used as “evidence”. The models are used to understand what the future climate will likely be under various scenarios of emissions. The fact that the models are not be perfect in all respects (no models are) does not mean that they are not useful. The predictions of greater drought in some areas are based on a combination of shifts in storm tracks and greater drying due to the hotter climate.
Joel Shore (17:51:54) :
“Greater drying due to the hotter climate” and water vapor feedback leading to hotter climate do not seem to add up, isn’t that a negative feedback that removes most of the theoretical warming?
Joel Shore,
The authors of the EOS paper are using the model internal variability to defend or excuse the model failure to match the climates increase in precipitation. When the model skill is at issue, it is circular reasoning. Of course, the defenders of the models would like the null hypothesis to be that the models are skillful, so their internal variability informs our understanding of the climate’s internal variability. Sorry, but models are not entitled to that null hypothesis. They first have to be able to replicate things like multi-decadal oscillations, increases in precipitation, surface albedo feedback, the signature of the solar cycle on the climate, ENSO, the observed radiative imbalance at the top of the atmosphere, etc in order to prove some skill.
The authors are not the arbiters of what their papers mean, and whether the evidence and arguments they put forward support their proposed interpretation. They presumably have passed some standards of peer review, but we can render our own judgement as to the quality of their evidence, the number and validity of their assumptions (such as model skill in certain areas), and the extent to which their conclusions are actually supported.
It is hard to dismiss the failure of ALL the AR4 models to reproduce an increase in precipitation within in even half the range of the observations as a chance occurrence, when there are probably hundreds of runs in about a couple dozen GCMs. The discrepency in the models simulation of the increase in evaporation and water vapor without a better corresponding increase in precipitation suggests that some fundamental parameterization of the physics is wrong. Whether the models get this wrong by the same means is unknown, what is known is they get it wrong in the same correlated direction. It doesn’t look like random error.