Curry Reviews Lindzen and Choi

As per a request initiated by Bender, here is a critique of the recent Lindzen and Choi paper.

Citation: Lindzen, R. S., and Y.-S. Choi (2009), On the determination of climate feedbacks from ERBE data, Geophys. Res. Lett., 36, L16705, doi:10.1029/ 2009GL039628.
http://www.seas.harvard.edu/climate/seminars/pdfs/lindzen.choi.grl.2009.pdf

Abstract: Climate feedbacks are estimated from fluctuations in the outgoing radiation budget from the latest version of Earth Radiation Budget Experiment (ERBE) nonscanner data. It appears, for the entire tropics, the observed outgoing radiation fluxes increase with the increase in sea surface temperatures (SSTs). The observed behavior of radiation fluxes implies negative feedback processes associated with relatively low climate sensitivity. This is the opposite of the behavior of 11 atmospheric models forced by the same SSTs. Therefore, the models display much higher climate sensitivity than is inferred from ERBE, though it is difficult to pin down such high sensitivities with any precision. Results also show, the feedback in ERBE is mostly from shortwave radiation while the feedback in the models is mostly from longwave radiation. Although such a test does not distinguish the mechanisms, this is important since the inconsistency of climate feedbacks constitutes a very fundamental problem in climate prediction.

I didn’t pay much attention to this paper (hereafter referred to as LC) when it came out, but it received quite a big play in the blogosphere. It was touted by some as the “end of the AGW scam.” On the other hand, numerous aspects of methodology received serious criticism. Why is this paper important (if correct)? The two main findings of LC are that: the sensitivity of the Earth’s climate is much smaller than the conventional estimates (e.g. IPCC); and climate models substantially disagree with observations and produce a sensitivity that is far too high (and hence are producing falsely alarming projections).

Some background: The equilibrium climate sensitivity (in K) is defined as a change in the equilibrium annual global surface temperature in response to a doubling of CO2. Alternatively, the equilibrium climate sensitivity can be express (in K/Wm2) as an equilibrium temperature change per unit radiative forcing. The climate sensitivity is not prescribed in global climate models but is determined as a result of the climate model integation including parameterization of various physical processes in these models. The equilibrium climate sensitivity is a useful parameter for comparing climate models. Climate system without feedbacks would have an equilibrium climate sensitivity of 0.3 K/Wm2 corresponding to the global average warming of about 1.1 K for doubling of CO2. Mostly positive feedbacks in the climate system have been estimated to increase the climate sensitivity to 1.5 to 4.5 K. LC is one of many studies that have tried to estimate climate sensitivity from observations.

For reference, some of the more substantive critiques of LC in the blogosphere include:
http://chriscolose.wordpress.com/2009/03/31/lindzen-on-climate-feedback/
http://motls.blogspot.com/2009/11/spencer-on-lindzen-choi.html
Would appreciate any other good links that you know of. The most significant critiques include: using an old (uncorrected) version of the ERBE data, ignoring a known temporal aliasing effect in the ERBE data, comparing to the AMIP (atmosphere only, with specified sea surface temperatures) rather than the CMIP (coupled atmosphere ocean) climate model simulations, incorrect handling of the direct response to sea surface temperature change. Each of these issues in implementation of the methodology could easily be fixed, but I would expect that their individual and cumulative impact on the analysis results would be substantial.

My assessment of this paper is quite critical (details to follow). In addition to the critiques of the methodology and its implementation linked to above, I have further concerns with the overall methodology that I hope provides a broader framework for criticism of this paper.

1. The first thing about this paper that struck me is that LC infer global climate sensitivity from an analysis of only the tropical oceans (20N to 20S). This is clearly stated in their methodology section and explicitly in the caption for Fig 1a, although in their title, abstract, and discussion the word “tropics” isn’t mentioned and the authors clearly interpret this analysis to be germane to global climate sensitivity. Why only the tropics are analyzed is not explained, other than by a vague statement in the last paragraph that refers to the “neutral higher latitudes.” Huh? Consider a global map of observed 20th century temperature changes (Fig 1B) http://ruby.fgcu.edu/courses/twimberley/EnviroPhilo/GlobalWarmingThreatProfJamesHansenNASA.pdf and also the IPCC climate model projections: http://www.ipcc.ch/publications_and_data/ar4/wg1/en/figure-10-8.html. The warming over the tropical oceans is lower than that for the tropical land masses, and far lower than that for the higher latitudes particularly the Arctic. But somehow (Fig 3) LC manage to determine a value of climate sensitivity from AMIP models for the tropical oceans that is between 2-4.5 K, which is essentially the same you would expect for the entire globe (e.g. the IPCC range of values) and in spite of the fact that the sensitivity of the tropical oceans is arguably at least a factor of 2 lower than the global average. So something is either wrong with the AMIP models (they definitely provide a different answer than the CMIP models, see Roy Spencer’s discussion on this) or with the LC methodology for calculating climate sensitivity.

2. The time scale of the feedbacks considered here are short term processes (over the tropical oceans) associated with clouds, water vapor and lapse rate, which are assumed to have equilibrium responses on time scales from a few months to less than 2 years. Even if this assumption re the timescale of equilibrium response is correct (see #3 below), Lindzen and Choi admit that their feedback analysis is relevant only for negative feedbacks since positive feedbacks have much longer equilibrium response times. It seems that this study is motivated by trying to document a negative feedback in the tropics, to support Lindzen’s iris hypothesis that addresses upper tropospheric water vapor and cirrus cloud feedbacks associated with tropical deep convection (which has received far more substantive criticism than support). The iris hypothesis and the tropical upper tropospheric water vapor and cirrus cloud feedbacks, while arguably still open to debate, are not by any stretch of the imagination a major driver in global climate feedback. Even in the limited context of local short term feedback processes over the tropical ocean, with the combination of issues raised in #1 and #2 I would expect the local feedback factor to be essentially zero.

3. Given that LC focus their analysis on the tropical oceans, the results from their analysis of ERBE data seems very implausible: a strong negative feedback in the shortwave (SWR), with a small positive feedback in the infrared (IR). The negative SWR feedback is basically an increase in the planetary albedo with increasing temperature, without a correspondingly large decrease in outgoing IR. How could this possibly be? The possibilities are:

• an increase in surface reflectivity (impossible since tropical ocean surface reflectivity doesn’t change with surface temperature)
• an increase in water vapor amount (doesn’t work, since water vapor changes have a much larger signal in the IR)
• increase in aerosols that are making clouds more reflective with relatively small impact on IR emission (no major volcanoes during this period, but Saharan dust and biomass burning could have the desired effect. However, no known relationship between surface temperature and dust/biomass aerosols)
• increase in low cloud amount (that reflects sunlight while emitting IR at nearly the same temperature of the surface). This works only if the increase in low cloud amount is not obscured by high clouds (which dominate the radiation fluxes at the top of the atmosphere when they are present). For this to be a significant effect, we would need to see a decrease in deep convective clouds in the tropics, which to my knowledge hasn’t been observed.

So, this large negative SW feedback with small positive IR feedback is not presently associated with a likely physical mechanism; e.g. Lindzen’s iris feedback wouldn’t produce this type of SWR/IR signature. I suspect that the large negative SW feedback identified from the ERBE data is an artifact of previously cited problems with the ERBE data analysis.

4. Even with a redo of the LC calculations fixing the implementation errors (e.g. using the correct version of the ERBE data, etc.), I am not convinced that the overall methodology used by LC can give a credible result for the climate feedback factor and climate sensitivity. The issue I want to focus on here is the nature of the energy balance model used to calculate the feedback factor and climate sensitivity. LC defines the feedback parameter as the change of the top of atmospheric radiation flux with change in surface temperature. If this seems counterintuitive to you and you don’t see how this relates to climate feedback, well it is based upon a lot of simplifying assumptions. This feedback parameter is derived from a simple linear feedback analysis of a simple energy balance model. Chapter 13 of my text Thermodynamics of Atmospheres and Oceans explains this
http://curry.eas.gatech.edu/climate/pdf/Ch13_GalleyC.pdf
http://curry.eas.gatech.edu/climate/pdf/chapter13_figs.pdf
For a lucid explanation of the specific equations used by LC (based upon a simple equilibrium planetary energy balance model) and discussions of the key assumptions and uncertainties, see this paper by Steve Schwartz (which by the way is highly controversial owing the short equilibrium response time that he determined)

Frame et al. ( http://www-atm.physics.ox.ac.uk/user/das/pubs/constraining_forecasts.pdf ) show that any estimate of climate sensitivity is critically dependent on subjective prior assumptions of the investigators, not simply on constraints provided by actual climate observations. Further, equally plausible approaches using the same model and observations can yield very different estimates of the risk of a high climate sensitivity. Assumptions inherent in the model that LC use include:

• radiative equilibrium at the top of the atmosphere. If the climate system rapidly equilibrates, then climate sensitivity can be inferred from the top of atmosphere forcing and the increase in temperature over a given time period. In contrast, if the climate response time is long, inferring climate sensitivity in this way would lead to an estimate of sensitivity that would be too low. Steve Schwartz argues that the response time is fairly rapid, whereas Hansen and others argue for a much longer response time http://ossfoundation.us/projects/environment/global-warming/summary-docs/oss-reports/slr-research-summary-2008/2005_Hansen_etal_1.pdf
Arguably, the equilibrium climate sensitivity cant be obtained directly from observations, since the Earth’s climate system is always changing. LC assume a short response time, which they say is justified for negative feedbacks.
• the model does not consider spatial variations in climate sensitivity (we have already seen evidence of much higher sensitivity in the Arctic)
• the model does not consider the frequency dependence of feedbacks (in sign and magnitude)
• in response to a change in external forcing (e.g. solar, CO2), the top of atmosphere fluxes (SWR, IR) can be determined by many combinations of surface temperatures and albedos, vertical distribution of temperature and humidity, cloud vertical and horizontal distributions, and aerosol particles. It is only in the context of a simplified model of radiative convective equilibrium that surface temperature provides a unique value of SWR, IR. The unusual features in Delta Flux / Delta SST identified by Roy Spencer are probably an artifact of this complexity http://wattsupwiththat.com/2009/12/17/spencer-on-his-agu-presentation-yesterday/.

Steve Schwartz concludes: “Finally, as the present analysis rests on a simple single-compartment energy balance model, the question must inevitably arise whether the rather obdurate climate system might be amenable to determination of its key properties through empirical analysis based on such a simple model. In response to that question it might have to be said that it remains to be seen. In this context it is hoped that the present study might stimulate further work along these lines with more complex models. It might also prove valuable to apply the present analysis approach to the output of global climate models to ascertain the fidelity with which these models reproduce “whole Earth” properties of the climate system such as are empirically determined here. Ultimately of course the climate models are essential to provide much more refined projections of climate change than would be available from the global mean quantities that result from an analysis of the present sort. Still it would seem that empirical examination of these global mean quantities – effective heat capacity, time constant, and sensitivity – can usefully constrain climate models and thereby help to identify means for improving the confidence in these models.”

Summary: No confidence in the analysis of LC.

Prognosis: So where do we go from here in evaluating climate sensitivity and understading feedbacks? The basic assumptions behind this type of sensitivity analysis based on top of atmosphere fluxes used by LC need to be tested by climate models. Personally, I don’t have confidence in this method. I spent the 1990’s working of the feedback problem (mainly in the Arctic). I hosted a workshop on feedbacks and sensitivity in 2003, the workshop summary contains much food for thought http://www.gewex.org/reports/workshop02.pdf

—-
Update: this essay was finished on Dec 26. Steve decided to wait a few weeks before posting, given all the interest in climategate. Since then, there is now a formal reply to the LC paper that is in press in GRL by Trenberth, Fasullo, O’Dell and Wong, this is dicussed at RealClimate http://www.realclimate.org/index.php/archives/2010/01/first-published-response-to-lindzen-and-choi/ and http://www.realclimate.org/index.php/archives/2010/01/lindzen-and-choi-unraveled/. See also DotEarth http://dotearth.blogs.nytimes.com/2010/01/08/a-rebuttal-to-a-cool-climate-paper/

1. stephen richards
Posted Jan 18, 2010 at 9:47 AM | Permalink

Judith

What would be the point of testing empirical measurements against models, please?

2. NokTang
Posted Jan 18, 2010 at 9:57 AM | Permalink

“I didn’t pay much attention to this paper (hereafter referred to as LC) when it came out”
* Why not? Lindzen and Choi are prominent climate scientists. Was it because of the “end of the AGW scam” that urged you to rebuke this paper in order to reinstall the status quo?

Anywayz the LC09 paper has been rebuked already wideley in the sceptic camp where, unless in the Church of Climate, there’s no place for group thinking.

3. Cameron
Posted Jan 18, 2010 at 10:10 AM | Permalink

Certainly not a scientist and have no math or science background upon which to asses the nature of the science being discussed. But am very interested in the topic for reasons that are obvious since this concerns us all. After surfing the net and reading what I can it was my understanding that the fingerprint of AGW according to the climate models would appear in first in the tropics. If so then why wouldn’t that be a valid location to search for the fingerprint of AGW? Or am I mistaken?

• Judith Curry
Posted Jan 18, 2010 at 3:52 PM | Permalink

Cameron, the fingerprint of of AGW is most evident in the high latitudes of the northern hemisphere

• Michael Smith
Posted Jan 18, 2010 at 4:17 PM | Permalink

I am confused by this response. I thought the GCMs were universal in predicting that the greatest warming should occur in the tropical troposphere. So why would there be reason to suspect an AGW “fingerprint” in the high latitudes of the northern hemisphere?

• Tolz
Posted Jan 18, 2010 at 8:37 PM | Permalink

Dr. Curry,

First of all, thank you! As to the “fingerprint” of AGW in the high latitudes of the northern hemisphere (or anywhere else for that matter), can you describe in general terms what that is? Steve has had many postings regarding the search (thus far in vain, to my reckoning) for “Waldo”. That sounds like the answer to the $64,000 question. • John A Posted Jan 18, 2010 at 8:59 PM | Permalink But not in the high latitudes of the Southern Hemisphere. Why is that? • Chris BC Posted Jan 19, 2010 at 12:52 AM | Permalink I too am confused as to why only “the high latitudes of the Northern hemisphere”. Why not high latitudes elsewhere? Why not elsewhere period? And what of high latitude areas where melting snow and ice reveals previous forests and/or man made structures in the NH? Aren’t those fingerprints of non-AGW warmer periods? As far as limiting scope to the tropics (not mentioned in this comment but several others close by), is that not the area of the earth with the greatest heat content and thus an excellent place to look at the heat flux to determine sensitivity? Finally, maybe slightly OT but what of past geologic events that caused plant death on massive scales? Would those events not have released far more CO2 and thus triggered the positive feedback mechanisms forecast by AGW theory? What of the ice core records showing CO2 levels following that of temperature? With high CO2 levels and positive feedback how could global temperatures ever dropped until after CO2 had dropped considerably? • Gord Richens Posted Jan 19, 2010 at 7:58 AM | Permalink “the fingerprint of of AGW is most evident in the high latitudes of the northern hemisphere” Thank you Dr. Curry for your input. It concerns some of us that climatology refrain from using the observed temperature increase in the northern latitudes, such as it is, to fortify the claim of “A”GW without explaining why human activity would affect temperatures in the north before elsewhere. • Gord Richens Posted Jan 19, 2010 at 8:54 AM | Permalink …mindful of the profound difference to policy implications between anthropogenic warming attributable to CO2 versus other human-induced causes (soot, etc.). • Jim Olsson, Dr. Posted Jan 19, 2010 at 9:02 AM | Permalink Seemingly, the temperature record “for the high latitudes of the Northern hemisphere” does not support a significant global warming compared to 1930-1940. Check for example Greenland and Faroe Islands at Danish state institute dmi.dk. The same applies to other areas for example Scandinavia and Norther Russia. Note that this argument and question was raised by Wibjörn Karlen in the Climategate emails. He is an expert in Arctic temperatures and does not believe in AGW. CRU and IPCC has produced graphs with increasing temperatures not traceable to the real data. Wibjörn Karlen did not received any satisfying answer. • JamesG Posted Jan 19, 2010 at 6:07 PM | Permalink You mean the place that’s officially now just as warm as it was earlier this century and which certain Russians now say is in the start of a new cooling cycle? I think you’re right! • JT Posted Jan 19, 2010 at 8:34 PM | Permalink Given that heat, which develops in the tropics as a result of absorption of shortwave radiation there, is transported to the Arctic by both the atmosphere and the oceans how can the Arctic’s “sensitivity” to CO2 enhanced longwave radiation from the atmosphere be disentangled from its sensitivity to convected heat? My conjecture is that because the Arctic’s temperatures are naturally low there is a larger temperature gradient between imported warm air and water and the local matter than exists in the tropics or mid latitudes, and that the gradient enhances the rate at which energy is transferred from warm atmospheric currents and warm oceanic currents to Arctic ice. In other words, the fact that the Arctic was melting recently does not necessarily imply that the Arctic is more “sensitive” to CO2 “forcing”. I suspect its likely it’s more sensitive to heat transport than to CO2 back radiation from the atmosphere. 4. James Chamberlain Posted Jan 18, 2010 at 10:15 AM | Permalink Judith, I appreaciate your answer and tone. I hope that more climate scientists take your route in defending themselves. James • Judith Curry Posted Jan 18, 2010 at 3:54 PM | Permalink Thank you James! 5. Paul Posted Jan 18, 2010 at 10:23 AM | Permalink It would appear that no one in the climate community can evaluate and compare modelled climate sensitivity to observations. At best modelled climate sensitivities are a guesstimate. Summary: No confidence in the models. It is a pretty fix that both empirical and modelled approaches are beyond our ken. 6. george h Posted Jan 18, 2010 at 10:48 AM | Permalink How well does what we know about geologic history support the extreme kinds of amplifying feedbacks assumed by the IPCC? Is there ANY evidence of net positive feedbacks from the geologic record? I think not. On the contrary, it seems to me that we have a climate system and planet which over eons is remarkably stable thermodynamically. Testing vs models? How about testing vs 4.5 billion years of reality? • Judith Curry Posted Jan 18, 2010 at 3:57 PM | Permalink We all wish we had good enough data over the geologic record to do such studies. But the planet has gone through some pretty dramatic climate changes, not sure ice age cycles should be classified as thermodynamically stable? • Bruce Posted Jan 19, 2010 at 8:01 AM | Permalink Depends on the energy inputs. What if the sun has wilder “teenage” behavior than it likes to “admit” now? That would imply a substantial negative feedback. What is disturbing is the apocalyptic pronouncements of the Team when these issues, among many, were not ever physically closed out, much less completed 10-15 years ago. So far I see the debate is just starting, and the Team is physically behind, winning only in their Gameboy simulations. • Henry Posted Jan 18, 2010 at 4:17 PM | Permalink It has been a while (50 million years?) since the Earth was completely ice covered or ice-free, but I don’t understand why that relative stability prevents some positive (or negative) feedback. Ignoring the strong possibility that feedbacks may be non-linear, and not knowing either the words or symbols, I don’t see why first order feedback by a factor $f$ cannot lead to eventual amplification by a factor $1/(1-f)$, which will not lead to meltdown or chronic instability so long as $-1. • RB Posted Jan 18, 2010 at 4:36 PM | Permalink You mean -1 < f < 0, and your statement is correct. I really wish we had sceptics who took up Richard Alley and Hansen on the 420 million years of evidence to which a CO2 sensitivity of 2.8C per doubling yields remarkable agreement from model-based simulations (see the Bjerknes' Lectures of both). Also, I don't know if I"m misunderstanding this, but wasn't snowball earth a positive feedback effect? • RB Posted Jan 18, 2010 at 4:36 PM | Permalink Oops, the way you had defined it, it should have been 0<f<1 • Henry Posted Jan 18, 2010 at 7:04 PM | Permalink Can you explain why it cannot be positive and negative? E.g. if $f=1/3$ then you get $1 + \frac{1}{3} + \frac{1}{9} + \frac{1}{27} + \cdots = \frac{1}{1-1/3} = \frac{3}{2}$ i.e. one-and-a-half times the original effect, while if $f=-1/3$ then you get $1 - \frac{1}{3} + \frac{1}{9} - \frac{1}{27} + \cdots = \frac{1}{1+1/3} = \frac{3}{4}$ i.e. three-quarters of the original effect. • RB Posted Jan 18, 2010 at 7:15 PM | Permalink Actually, if you have a negative feedback with forward gain G and feedback gain H, you would write the equation classically from input,X to output Y as Y=G*(X-H*Y) For a forward gain, G of 1, that would be Y=X-H*Y since you would subtract H*Y from the input before amplifying it with a gain which yields Y=X/(1+H) in which case amplification without instability will be for the range -1 < H <0 However, if you had defined your input-output relationship as a positive feedback, you would have defined it as Y=X+H*Y which would yield Y=X/(1-H) which is the relationship you had defined. In this case the relevant regime for amplification without unbounded instability would be 0<H<1 Its just a matter of definition of how you define feedback. • RB Posted Jan 18, 2010 at 7:17 PM | Permalink Oh, sorry, specifically with regards to your question, if amplification is 1/(1-f), for -inf<f<0, you have a negative feedback system (gain always less than 1). For 0<f<1, you have bounded positive feedback. At f=1, you have unbounded positive feedback. • RB Posted Jan 18, 2010 at 4:51 PM | Permalink It is helpful to discuss the timescales involved. If there are natural feedback effects in Earth’s thermostat, what are the relevant timescales? “In the long run” is not useful for the discussion if the timescales involved are million-year silicate weathering. • Marc in VA Posted Jan 16, 2011 at 7:33 PM | Permalink George H. I don’t know if this is responsive to your original question (“How well does what we know about geologic history support the extreme kinds of amplifying feedbacks assumed by the IPCC?”), but I just look at this and it seems germane. Title: “Climate sensitivity constrained by CO2 concentrations over the past 420 million years” Their bottom line: “[Climate sensitivity] of at least 1.5 uC has been a robust feature of the Earth’s climate system over the past 420 Myr, regardless of temporal scaling.” I am waaaay unqualified to critique the paper itself. 7. Posted Jan 18, 2010 at 11:12 AM | Permalink Dear Judith, thanks for your detailed analysis. Is it OK to admit that a majority of it is a (somewhat cherry-picked) review of critical points that have usually appeared elsewhere? http://motls.blogspot.com/2010/01/ipcc-types-read-lindzen-choi-2009.html I have seen a newer version of the Lindzen-Choi paper and pretty much all the things I was dissatisfied with – or convinced by others to be dissatisfied with – have been fixed, as far as I can say. However, you seem to believe that there exists a deeper path of criticism of the very basic framework by LC. The most far-reaching point in this approach of yours is the following sentence: “The basic assumptions behind this type of sensitivity analysis based on top of atmosphere fluxes used by LC need to be tested by climate models.” I don’t understand the logic here. One can’t say that the LC theoretical model – which is just about the framework of quantities to be tested and their conjectured relationships – is falsified just by showing that the existing climate models don’t reproduce its assumptions, can he? Quite on the contrary, the logic here is that the disagreement between LC and the climate models shows that that climate models are wrong – as long as one compares the models results with the empirical data. The objects that LC are calculating are really phenomenological ones – they can be defined operationally from the empirical data, in a simple way. Whatever the theories and models are, they must be able to calculate the corresponding quantities because they’re measurable. One can’t “debunk” a criticism (or falsification) of models by assuming that the models are correct, can he? I have clearly had different zeroth-order expectations about the dependence of the radiation on the temperature etc. than Richard had. I always imagine that the black body radiation is the bulk of it and all the other effects including the albedo are perturbations. But Richard knows quite something about the climate and his and Choi’s test is a very clever challenge – a method to isolate the sensitivity that reveals many other patterns and data that the models should agree with. I am surprised that you disagree with this modest general proposition. Another point. You say that the radiation is not determined just by the sea surface temperature but also by the vertical profile of the temperature and humidity, among other things. And I fully agree. But that doesn’t mean that one can’t get anything interesting by neglecting these extra things. The idea is that all these other factors behave “randomly”, according to some statistical distribution. The vertical gradient of temperature at a certain altitude and the humidity over there are random variables that are – in the tropics – distributed in some way. The point is that if you neglect their influence on the energy flux, you will just add some noise to the LC curves. Indeed, that’s a part of the noise that is shown in the Lindzen-Choi graphs. But as long as the slopes in the graphs can be determined at some appropriate level of statistical significance – both in the models and in the empirical observations – it just means that the noise, including the non-temperature influences on the energy flux, has simply averaged out, right? Cleverly, LC obviously look at the time intervals where the changes have been fastest because these are the best points to suppress the noise. But the vertical humidity profiles etc. change the time-derivative of the fluxes in some way, and the mean value of this change is how the atmosphere works “normally”. The average influence of the humidity profile is a part of the question, and it’s completely correct that it is a part of the question. LC study the actual climate system which does include humidity and its vertical profiles and their fluctuations in time. They want to determine the sensitivity in this real climate. So they look at the corresponding empirical, observed data – and they try to adjust parameters in a phenomenological model that most accurately reproduces these observed data. This is a clearly valid methodology. When it comes to CO2, this logic of averaging the “noise” out is applied all the time. All temporary and local changes are just “irrelevant weather” effects while what we get by average over the globe and 3 decades is the “important CO2 climate signal”. Sorry to say but the fact that the sea temperature decides about the energy flux above it is much more solid – and much more “dominant an influence” – than the influence of CO2 on the temperatures. So if the methodology of “treating all other effects as noise that is supposed to average out, to isolate the signal” can be applied to CO2, it can surely be applied to the sea surface temperature. I think that the original paper by Lindzen and Choi had mistakes but in general, it is a clever new way to isolate some “parts of the climate device” that can be separately measured and separately compared to the theories and models. This is exactly how science should be done. One must look at the reality from new angles – and find the new best observables that allow us to decide about the agreement or disagreement between the reality and the theories most efficiently and accurately. This is much better than to compare – and test – just some incomprehensible overall mixture of millions of influences that haven’t been verified in isolation. It’s much better than to use still the same test – that has been “programmed” to pass because it’s the only thing that the climate models have been trained to pass. Such a single test doesn’t prove any validity. Many more tests are needed and they have to be done all the time. One has to study dependences of many pairs of quantities on one another, the critical exponents in temporal and spatial autocorrelation, the dependence of fluctuations and various correlations between fluctuations on the time scales, seasons, latitude, altitude – everything. Too bad you disagree. I feel that the criticism – an eclectic synthesis of critical points found at various places – is primarily an excuse for not learning new ways how to look at the climate system (perhaps because of fear that it could reveal that something you have believed is incorrect). But the new way by Lindzen and Choi is certainly worth considering. Best wishes Lubos • Judith Curry Posted Jan 18, 2010 at 4:19 PM | Permalink Lubos, thanks for your thoughtful message, and for pointing out your recent post on LC. With regards to the model they use, it is based on a simple energy balance climate model and radiative convective model (that have been around for decades). A simple feedback analysis of the simple energy balance climate model was first done I think by Jim Hansen and MIchael Schlesinger in the 1970’s. This kind of model tied to the TOA radiation fluxes doesn’t make any sense for a region that is sub global (such as what LC did); to apply this would require a transport term to account for advection into and out of the region. So there is nothing novel about the model that LC used, but the model (at least in the form they use) should only be applied globally and not to a region. And if people believed the simple energy balance and radiative convective climate models, we wouldn’t need global general circulation climate models. These simple models are useful conceptual tools, but are not convincing to me in terms of providing a useful estimate of climate sensitivity when applied in the manner of LC • Artifex Posted Jan 18, 2010 at 8:45 PM | Permalink I buy Dr. Curry’s point about the transport term needing to be considered and the correctness of LC may very well depend on the magnitude of this flux. However if this flux is very small compared to the incoming and outgoing energy LC may still hold. Does Dr Curry have a guess at the magnitude of this term ? There is another interesting point here. It would seem to me that if there are big arguments over the size of the transport term,the existing GCM’s could be used to provide estimates of the transport term. Then if the LC model still points to a dramatically lower sensitivity, what does that say about the GCM’s ? • Wondering Aloud Posted Jan 19, 2010 at 11:18 AM | Permalink As an often hard line trouble maker I want to change my usual approach and thank you Judith for your efforts and thoughtfulness. this is the kind of discussion I like. I hate the tendancy, mine as much as anyones to be knee jerk and pick sides all the time. • EdeF Posted Jan 18, 2010 at 5:53 PM | Permalink After reading LC2009 and the critique and the response from Lubos, I have to concur with Lubos. BTW, here is the link to the ERBE experiment: http://asd-www.larc.nasa.gov/erbe/ASDerbe.html • Dave Posted Jan 19, 2010 at 9:29 AM | Permalink Maybe I’m showing my ignorance here, since Dr Curry hasn’t picked up on these points, but: “The idea is that all these other factors behave “randomly”, according to some statistical distribution. ” Surely, if you assume that, and your results diverge from reality, you have proven, reductio ad absurdem, that they are not random? Or, you can only have as much confidence in them as in your initial assumption. “One can’t “debunk” a criticism (or falsification) of models by assuming that the models are correct, can he?” Of course you can. If you assume something to be true, and it leads to completely nonsensical conclusions, you can say with some confidence that the assumption is wrong. “as long as the slopes in the graphs can be determined at some appropriate level of statistical significance – both in the models and in the empirical observations – it just means that the noise, including the non-temperature influences on the energy flux, has simply averaged out, right? ” This seems totally bizarre to me. You seem to be contending (reasonably) that a trend may be made up of signal+noise, but *assuming* (unreasonably?) that the part you are looking at is the signal and not the noise. • Gerald Browning Posted Jan 20, 2010 at 12:46 AM | Permalink Lubos, I enjoyed your objective review of the LC manuscript. I look forward to reading their new manuscript. Jerry 8. Posted Jan 18, 2010 at 11:26 AM | Permalink ‘ in their title, abstract, and discussion the word “tropics” isn’t mentioned ‘ It is mentioned clearly in the second sentence of the abstract. Summary: No confidence in the analysis of JC. • David P Posted Jan 18, 2010 at 11:44 AM | Permalink Ouch. And that was her #1 gripe. • Tolz Posted Jan 18, 2010 at 11:52 AM | Permalink That’s a little harsh, PaulM. It’s good that Dr. Curry posts here. Even in finding fault with her analysis, I think she deserves respect and appreciation. Things are a whole lot more interestng around here when a qualified person has a different perspective. • Dave Dardinger Posted Jan 18, 2010 at 12:48 PM | Permalink From past experience I’ve concluded that Dr. Curry has a sufficiently thick skin to handle the sort of slap-back PaulM made. After all, a conclusion sentence like “Summary: No confidence in the analysis of LC.” is just begging for a response in kind. This is simply a heat-kitchen situation. And, yes, there have been any number of important scientists who have posted here and then failed that test, but Judith isn’t one of them. She does what should be done and ignores those who should be ignored. What is more to be looked forward to is how she responds to Lubos’ post. Since she links to his site, I think she’s required to make some sort of response. Then we in the peanut gallery can start learning the real issues. • TerryMN Posted Jan 18, 2010 at 2:13 PM | Permalink Snippy, perhaps, but it really does blow a large hole in the credibility of the critique. I saw the comment before David replied, and his reaction (Ouch.) was mine as well. I echo Mike B’s question below also. Even given the commonly stated claim that the arctic is a proverbial mine canary that will see warming first and fastest, it’s 4% of the globe and the antarctic is not showing any statistical signs of warming (another 4%). It sounds a bit hypocritical when we hear so often that “The US is only 2% of the globe, so trends here are unimportant.” • Steven Mosher Posted Jan 18, 2010 at 5:59 PM | Permalink Huh? How does that blow a hole in the analysis. Remove it and her other issues still stand. It’s not like a bristlecone. • TerryMN Posted Jan 18, 2010 at 8:33 PM | Permalink Read my comment again, Mosh – I didn’t say it blew a hole in the analysis. The credibility of the critique, for me, was damaged with this: Abstract: Climate feedbacks are estimated from fluctuations in the outgoing radiation budget from the latest version of Earth Radiation Budget Experiment (ERBE) nonscanner data. It appears, for the entire tropics, the observed outgoing radiation fluxes increase with the increase in sea surface temperatures (SSTs). …………………. 1. The first thing about this paper that struck me is that LC infer global climate sensitivity from an analysis of only the tropical oceans (20N to 20S). This is clearly stated in their methodology section and explicitly in the caption for Fig 1a, although in their title, abstract, and discussion the word “tropics” isn’t mentioned and the authors clearly interpret this analysis to be germane to global climate sensitivity… I wouldn’t have taken PaulM’s tone, but Dr. Curry has a false assertion in point number one. Not a good way to start a critique, IMO, but what the hey. Whether 20S to 20N is a better sample than the arctic or antarctic in studying trends (or lack thereof) is probably another discussion for another day. Cheers, Terry • Steven Mosher Posted Jan 18, 2010 at 11:42 PM | Permalink fair enough, but I still dont think it goes to the credibility of the critique. What it does, for me, is it puts me on the look out for similar mis readings. If the mis readings form a pattern then I might not proceed to the next level which is to push down into the analysis. • Judith Curry Posted Jan 18, 2010 at 3:52 PM | Permalink Assessments without analysis aren’t useful • Judith Curry Posted Jan 18, 2010 at 6:43 PM | Permalink You are missing the point here. LC conduct an analysis of the tropics, using a global energy balance model that is designed for the tropics. The global model cannot be used for a regional analysis without including energy transfers in and out of the global box. Apart from that, LC infers global sensitivity based upon the tropical analysis. Clearly LC is using an analysis of the tropics to infer global climate sensitivity. • harold Posted Jan 19, 2010 at 1:14 AM | Permalink I don’t see leaving out large parts of the globe and related energy transfers as being a big problem. Every model out there leaves out things and makes simplifying assumptions. Empirical quasi-physics has it’s limitations, and the only thing we can be 100% sure of is none of the models are right. Some may have a preference for one or another model, or disfavor some other but it really isn’t as if it’s physics. The notion of using quasi-empirical models to evaluate postulated constants is interesting – I wouldn’t pin anything in the results. 9. Carl Gullans Posted Jan 18, 2010 at 11:51 AM | Permalink PaulM: Could have pointed it out w/o being snippy; don’t we all want Judith to have a conversation here? No need to be rude. • Posted Jan 19, 2010 at 12:37 PM | Permalink I don’t think I was rude. I just pointed out an error rather briefly, expanded on by Terry above. Indeed it is to be welcomed that JC posts here and responds to comments. I am sure she is aware that anyone posting on a climate blog needs a thick skin. Yes she deserves some respect, but so do L&C. She should acknowledge her mistake. 10. Mike B. Posted Jan 18, 2010 at 12:07 PM | Permalink Judith, I don’t understand why you are so critical of the fact that LC only looks at the sensitivity in tropics. LC makes it quite clear, throughout the paper, that they have focused only on the tropics. They also explain why they looked only at oceans (satellite data was unreliable, and in any case land area represents only22% of the tropics). This criticism is particularly strange when you later note that you yourself have in the past conducted studies focused on the Arctic. Less than 4% of the earth’s surface is above the arctic circle, while about 35% is between 20% North and South. Finally, I don’t think LC’s “neutral” comment regarding the extratropics means what you imply. • Judith Curry Posted Jan 18, 2010 at 3:53 PM | Permalink Mike, thanks for your question. I am not critical of them focusing on the tropics, but rather for inferring global sensitivity from analysis of the tropics • Ron Cram Posted Jan 18, 2010 at 5:31 PM | Permalink Judith, I thought LC was inferring radiative imbalance in the tropics. Is that not correct? Wouldn’t radiative imbalance in the tropics be a fairly decent proxy for global climate sensitivity? • Ron Cram Posted Jan 18, 2010 at 5:37 PM | Permalink Judith, I see you now that you had already answered my question when you answered Motl’s. Thank you. • Judith Curry Posted Jan 18, 2010 at 6:40 PM | Permalink Ron, part of the issue is that LC’s model should be applied globally, not regionally • WillR Posted Jan 18, 2010 at 6:56 PM | Permalink I need to read everything again — I just skimmed it — however… A fellow (Gene” gave this reply one day in a short discussion about modeling… (not here) I recall one instance when I wanted the modelers to run a specific type of prediction, which they refused because they said we can’t get the right answer. I tried to tell them that I knew that, but I only want to know: “What are the important variables.” So I have to ask the same question. What was the the intent of LC — to model the earth? Or to get an indication of the “the important variables”. I am used to doing large models — NP problems / time series. But I though he had a good point. It’s nice to know what’s important in the model — and where it’s important. So I have no trouble accepting their idea of a limited model. 11. commieBob Posted Jan 18, 2010 at 12:23 PM | Permalink Hi Judith, In Electronics (my field) we prefer to deal with systems, as much as possible, as black boxes. The response of a black box is most easily characterized by applying an impulse or a step input and looking at the output as a function of time. The advantage of the above approach is that it doesn’t matter how the system is actually realized. For instance, one could have an analog box and a digital box and the system function would be the same. You couldn’t tell the difference between the boxes and you wouldn’t care anyway. All you care about is the function. The reason I like LC is that it looks like a reasonable test of the (very very complex) black box that is the planet Earth. LC considers energy in at the top of the atmosphere and energy out as radiated from the surface but received above the atmosphere; in other words, energy in vs. energy out. If you don’t like LC, can you perhaps suggest a better test? Do remember, though, that if you start worrying about the physical processes involved within the system, you probably don’t have a valid black box. • Judith Curry Posted Jan 18, 2010 at 4:23 PM | Permalink commiebob, your point is interesting. The energy balance climate models assume a black box that is in equilibrium. Unfortunately the earth’s climate is never in equilibrium. Averaging over an ideally long time scale might get around this problem, but for the length of the data record that we have to work with, this isn’t very feasible. Given this, personally i think we need to dig into the processes and make sure that they are all treated correctly in the climate models and that they interact correctly. • P Gosselin Posted Jan 19, 2010 at 4:48 AM | Permalink “The energy balance climate models assume a black box that is in equilibrium. Unfortunately the earth’s climate is never in equilibrium.” No, the black box is not in equilibrium. You apply an input, and then you get a corrsponding output. In engineering mechanics, a bridge designer starts with free body diagrams, where all the relatively minor imperfections, anomalies and internal factors are neglected, not unlike a black box. It works very well and is a great place to start in the complex analysis of how the system will behave under different loads. I think this is analogous to what LC have done. It’s a good way to start. They’re on the right track. I could never imagine a bridge engineer starting a sturctural analysis by examining the weights of the rivets and the paint on the structure. A crude analogy, I admit. • Bernie Posted Jan 18, 2010 at 5:09 PM | Permalink commieBob: I assume that you are not looking at any old set of black boxes. There is presumably a set of criteria for choosing a “reasonable” black box beyond simply the input/output parameters to begin with – hence Dr Curry’s approach makes sense because we already strongly suspect aht many of the processes inside the black box are poorly specified. The problem is that the level of investment in the models is very large. Where I think your analogy has some real power is the notion that it is meaningful to assess the value of the black GCM boxes by averaging across all the black GCM boxes. I know of no hardware designer who would ever dream of such a strategy to suggest that the “right” design must be among the black boxes because we get the answer we are looking for!! • commieBob Posted Jan 18, 2010 at 8:44 PM | Permalink Hi Bernie, Given my druthers, I would deal with only one black box at a time. What I was referring to as reasonable was the choice of input and output. Energy in, energy out, sea temperature. The black box, as I refer to it here, is physical. This black box is not an analytical technique, it is that which is to be analyzed. • Chris BC Posted Jan 19, 2010 at 1:13 AM | Permalink I also have experience in Electronics and testing and I’m with commieBoB (although I hope he’s not really a commie). It seems to me that it will be very difficult to convince everyone until the black box response can be measured adequately. Inputs vs. outputs. Whether LC have it exactly right, they’ve got the right idea. I have zero confidence in GCMs that aren’t even agreed upon by the warmists, and certainly not the skeptics. When NASA/NOAA and the Met Office can start making reasonably accurate 3, 6, 12, and 24 month forecasts then maybe I’ll believe they have enough understanding of what’s going on inside the box to model it accurately. • Judith Curry Posted Jan 18, 2010 at 6:16 PM | Permalink commiebob, a question for you. Would you expect a complex nonlinear system (like the climate) to be treatable by a simple linear feedback analysis? What is the scope of system complexity that you try to analyze using control theory? • mpaul Posted Jan 18, 2010 at 7:04 PM | Permalink Much work has been done to apply stochastic process theory to black-box simulation. The method has been applied to moderately complex systems with great success. By contrast, deterministic computer modeling generally fails (and quite spectacularly) when the number of variables becomes moderately large. So perhaps the question is best directed back to you: what is the scope of system complexity that you think deterministic computer models can be applied to? Computational fluid dynamic models can not reliably predict air flow through a city (for example), so why do we think a deterministic model can be created for the climate (a much more complex system). • Calvin Ball Posted Jan 18, 2010 at 11:07 PM | Permalink Lucia said it very well once. We can model the air flow around an airplane quite well, as long as its going forward. Turn the airplane sideways, the modeling doesn’t work any more. • bender Posted Jan 18, 2010 at 11:36 PM | Permalink The standard counter argument is that at a certain time-scale the “stochastics” (i.e. determinstic nonperidic flow) are irrelevant compared to the deterministic forcings. This is a logical argument. That doesn’t mean it’s true, however. Read everything Tom Vonk has written at CA. • Steven Mosher Posted Jan 18, 2010 at 11:46 PM | Permalink I tried I honestly tried. needs pictures Vonk does. ( sorry star wars marathon is on ) • bender Posted Jan 18, 2010 at 11:49 PM | Permalink When you read Vonk, do the raspy Vader voice. Eerie. Like jeez doing Cousteau. • mpaul Posted Jan 18, 2010 at 11:55 PM | Permalink Agreed, but the periodicity for the primary mechanisms is 1,000s and often 10s of thousands of years. The stochastics are still largely in play when the time scales are 50 or 100 years. I’m afraid that climate scientist are confusing weather with climate. • bender Posted Jan 19, 2010 at 6:35 PM | Permalink “When are the stochastics in play vs not?” That is the question. If you’d have asked Gavin 10 years ago, during a warm trend, he would have said over 10-year scales. You ask him now that it’s not warming and he says, umm, 30 years. You say 50-100. I’d like to see that question discussed by the authorities of the settled science. I’d like to see them prove that weather stops “here” and climate starts “there”. Instead, hand-waving about about “what you expect” vs. “what you get”. I want to see it in an IPCC report. • Steven Mosher Posted Jan 19, 2010 at 11:24 PM | Permalink ha, did you ever read the mail where jones discusses the 30 year period? Its in there.. very funny. I posted it over at Lucia’s the period is actually 35.. Perhaps we should go hunt up the study he refers to. If interested I can go look the mail up.. or just search on the word “35” only a couple mails pop up. • Posted Jan 20, 2010 at 8:48 AM | Permalink “When are the stochastics in play vs not?” … I’d like to see that question discussed by the authorities of the settled science … I want to see it in an IPCC report. It’s outrageous that this has not been a major emphasis on this at the IPCC level. The superciliousness of the typical alarmist – I’m thinking George Monbiot in the December debate I attended with James Delingpole in London, or John Hirst, head of the Met Office, confronted by Andrew Neil on BBC TV more recently – that those of us less than impressed by the Met’s regional forecasting 3-6 months out or by the lack of success of the GCMs 10-15 years out haven’t grasped the basic fact of the settled science – the difference between weather and climate. And the magic number of thirty years where one wonderfully morphs into the other is deduced from real world data where exactly? Feynman would make mincemeat of these people. • Judith Curry Posted Jan 24, 2010 at 8:24 AM | Permalink The weather/climate issue is an interesting one. NOAA formally defines climate as anything longer than two weeks (which is the period for which they do weather forecasts). Does this make sense? of course not. There is a weather/climate continuum. In the old days, before we understood the longer period ocean oscillations like PDO and AMO, 30 years was a good “averaging period” for climatology. If you are doing trends, 30 or 35 years doesn’t work if you are say starting at the coldest part of the AMO and ending at the warmest part. I have argued (in the context of my analysis of hurricane activity) that averaging should be done over discrete periods of AMO/PDO phase combinations, then trends can be determined by differencing between two successive like phase combinations (e.g. the warm amo-warm pdo periods of the 1930’s vs 1995~2002). The statistical significance of doing something like this is rather daunting, but the physics of what is going on suggests that this is the way to do trends for the 20th century. Back to the weather/climate continuum. The main impacts of climate change are mainly associated with changing frequencies and intensities of extreme weather events! Tim Palmer of ECMWF is advocating that the best climate models (e.g. ECHAM5) have the same atmospheric and ocean components as the weather models and seasonal climate models, which are evaluated on a daily basis. He further argues that to get the fast feedbacks correct (e.g. cloud and water vapor), that the weather systems need to be resolved and modeled correctly, which requires requires resolution on a scale of order 10 km (which the ECMWF model now has). ECHAM model has just completed 12 km resolution simulations of the climate model on the ORNL computer, one would hope that these simulations would be a definitive source of info on modeled cloud/water vapor feedbacks • Charlie A Posted Jan 24, 2010 at 1:40 PM | Permalink Slightly off topic, …..snip • EdeF Posted Jan 20, 2010 at 12:59 PM | Permalink Any computer simulation can be evaluated. I propose the following tests of the GCMs (pick any one) over a very limited future timescale. Here are my tests: a) The number of named storms in the Carribean July-Nov 2013. b) The monthly rainfall at Seoul, Korea for the monsoon month of July 2012. c) The number of days above 110 deg F at Death Valley, Calif. August 2015. d) The average temperature for the month of January 2011 at Managua, Nicauragua. e) Total snowfall for Elmau, Austria Dec 2013 to March 2014. I will guess the values by averaging the previous 5 years data. If the GCMs can beat me I will buy their team two cases of Dos Equis XX. Test to be concluded September 2015. Any takers? • WillR Posted Jan 20, 2010 at 1:35 PM | Permalink Considering that the confidence is very high on the other side of the debate — maybe they can go 10-to-1 against you. Just be careful of alcohol poisoning… • Posted Jan 19, 2010 at 4:47 AM | Permalink Dr Curry, You asked, “commiebob, a question for you. Would you expect a complex nonlinear system (like the climate) to be treatable by a simple linear feedback analysis? What is the scope of system complexity that you try to analyze using control theory?” A complex, nonlinear system is still a black box from the standpoint of an external observer. It receives an input from an external source, and produces an output which is some function of the input. While mechanisms inside the box determine that function, we have no hope of elucidating it without good data on I/O. While measurements of the TOA energy balance in the tropics are surely insufficient, they are necessary and relevant. Moreover, if there are differences in the radiative balance with latitude, the models should embody an explanatory mechanism for those differences and predict them. Thanks for joining the discussion. This is “peer-to-peer” review at its best! • Craig Loehle Posted Jan 19, 2010 at 9:43 AM | Permalink Of course the problem here is that with the climate system we are only allowed one perturbation of the real system (20th Century climate), which is under-determined wrt dynamics of the complex model. We can of course perturb the simulations any way we want, but only if “we” are part of one of the modeling teams. • Bernie Posted Jan 20, 2010 at 9:48 AM | Permalink Doesn’t this point severely limit the value of feedback analysis for climate systems? It seems to me that this type of analysis is necessary but not sufficient for determinibng future states. • commieBob Posted Jan 19, 2010 at 12:13 PM | Permalink Hi Judith, Sorry for the slow reply. To answer your question: “Would you expect a complex nonlinear system (like the climate) to be treatable by a simple linear feedback analysis? What is the scope of system complexity that you try to analyze using control theory?” 1 – I bend over backwards to eliminate complexity especially in cases where the system is not well understood. A poorly understood complex nonlinear system is not reliably treatable with simple linear feedback analysis (or often, to be fair, any analysis). 2 – On the other hand, a well understood (albeit complex and nonlinear) system is often treatable using a rather simple analysis. The history of my own field (Electronics) is one of simplification. As we come to understand systems, the analysis becomes simpler because we know what we can safely ignore. What I didn’t include in my original post were the constraints on my trust of LC: a – The analysis deals with an obvious first order effect b – which occurs over a short period of time. As long as the above constraints are met, the nature of the system within the black box doesn’t usually matter. (rant)The trouble is that, once we have found a valid correlation, we are tempted to take that as proof of a bunch of other stuff that we haven’t actually proved.(/rant) 12. Dennis Wingo Posted Jan 18, 2010 at 1:31 PM | Permalink Just as a reference….. I am working on Nimbus data from the 1960’s and have found probably the first paper related to Earth Radiation Budget in the Arctic and Antarctic regions, based on IR data from the Nimbus II HRIR sensor. The Radiation Balance of the Earth-Atmosphere System Over Both Polar Regions Obtained from Radiation Measurements of the Nimbus II Meteorological Satellite” Some interesting stuff in there, especially as it was in the middle of the 1960’s cooling interlude. document is accessible from the NASA Technical Reports Server (http://ntrs.nasa.gov) as document number 19670024960_1967029460.pdf Just a heads up to interested parties. 13. Jimchip Posted Jan 18, 2010 at 1:58 PM | Permalink Professor Curry and Mr. McIntyre, I want to sit behind Bender so I can look over his shoulder. He’s been studying up. I know I’m going to learn a lot but I won’t have much to say, at least until I finish my reading, and then, probably, just to ask some questions. • Judith Curry Posted Jan 18, 2010 at 4:23 PM | Permalink Jimchip, i am DEFINITELY waiting to see what Bender has to say 🙂 • Steven Mosher Posted Jan 18, 2010 at 6:03 PM | Permalink I’ve relieved bender from his pool cleaning duties. If we say his name three times he should appear. bender. bender. bender. • bender Posted Jan 18, 2010 at 6:56 PM | Permalink Does it make sense to reply to a Dec 26 essay? Maybe makes more sense to let the “experts” quibble a bit, until they’ve settled on an equilibrium state of mutual disagreement? 🙂 . Ah, breathe in that clear air of the consensus. . Let’s wait a bit yet. Let the auditing task clarify itself somewhat. . [P.S. Although I am a curious persion, “curious” is not me – despite the ambiguous iconery.] • Steven Mosher Posted Jan 18, 2010 at 11:58 PM | Permalink Ok.. I’m bothered by things on both sides of this. The models used by LC and the generalized unwillingness to put models into contact with observational data. • Patrick M. Posted Jan 18, 2010 at 8:02 PM | Permalink But can you make him disappear again? • bender Posted Jan 18, 2010 at 11:22 PM | Permalink Genies don’t go back in that easily once out. • Patrick M. Posted Jan 19, 2010 at 7:09 AM | Permalink Sounds like the kind of thing a genie would say… • Jimchip Posted Jan 19, 2010 at 2:11 AM | Permalink what to do about bender? this is just me but i thought maybe nonlinear thermodynamics? third order semi-implicit Runga-Kutta? Anything, nonlinear. (Ahem, no hockeysticks allowed) Or, anything else that Professor Curry wants. I’ve got a beach. maybe…pool duty for bender? Mosher… • bender Posted Jan 18, 2010 at 11:33 PM | Permalink There’s amazingly little to comment on. First, her review of L&C. She thinks extrapolation from tropics to globe is perilous. She could be right. But her concerns could also be overstated. For her to dismiss the whole of the paper on the basis of one (albeit important) criticism, this is going a bit far IMO. FWIW I had the same concern. So did Motl, apparently. Notable that Lindzen appears to have a reply. Not surprising. That extrapolation may be invalid does not mean it IS invalid. Especially if tropics and extratopics are “teleconnected”, the same mechanism operating tropically could be operating extratropically. No one will know for sure until the study is done. . Second, I started by asking her to assess Lindzen’s POV, not just review the L&C paper. Her focus in excessively narrow. That there are shortcomings in this paper does not mean his skepticism is unwarranted. It seems almost nitpicky to focus on one paper when his POV is founded on a corpus of insight and experience. • Posted Jan 19, 2010 at 12:05 AM | Permalink It seems almost nitpicky to focus on one paper when his POV is founded on a corpus of insight and experience. Very well said. Yet how can one hope to pick up all of that (and of those that disagree)? Still, it’s cool listening in to the pool cleaners and others trying to make head or tail of it. • Cameron Posted Jan 19, 2010 at 1:45 PM | Permalink Still, it’s cool listening in to the pool cleaners and others trying to make head or tail of it. Ooo that is a rather revealing statement. I have a good friend, well not anymore we’ve had a falling out, but he is a theoretical mathematician whose expertise is in optics. A very eccentric fellow who earns a living cleaning pools. • Posted Jan 19, 2010 at 2:11 PM | Permalink I’m a slow learner but one thing I definitely remember from Climate Audit is that bender is in the same line of business as your friend. • Cameron Posted Jan 19, 2010 at 4:07 PM | Permalink Oh really? I did not know. It’s not generally accepted yet, even by the skeptics, but politically speaking AGW as a world wide political movement is dead. There is just way too much ammunition out there now that is easily understood for the political opponents of the movement. The claim the science is “settled” is laughable. The Medieval Warm Period, and Little Ice Age are back on the map. Not to mention the drip, drip, drip, of data manipulation, or the ethical and financial malfeasance that been exposed. How long before it’s finally buried who knows, but it’s now only a matter when not if. Anyway I think I’ll just go back to lurking and not posting. Have a good day. • Posted Jan 19, 2010 at 7:23 PM | Permalink As you browse look out for references to bender’s occupation. You may decide it’s a spoof – but I have no final proof of that. • Jimchip Posted Jan 19, 2010 at 2:19 AM | Permalink spell it all out. too many acronyms. • MrPete Posted Jan 19, 2010 at 7:32 AM | Permalink Re: Jimchip (Jan 19 02:19), On the web and elsewhere, many of those acronyms are now meaningful words. We live in a textified world where spoken and written language need not match. InMyOpinion ForWhatItsWorth PointOfView IMHO, of course 😀 • Posted Jan 19, 2010 at 2:34 PM | Permalink Re: Jimchip (Jan 19 02:19), IIRC, the excellent CA101 includes non-tech abbreviations as well as techie ones. • Jimchip Posted Jan 24, 2010 at 11:02 AM | Permalink I joke with Bender, maybe too much. I said before I’d only have a few comments, maybe a few questions. My, admittedly, obscure comment was maybe even for Prof. Curry. Communication partly involves ‘knowing one’s audience’. It was mostly for Bender 🙂 FWIW, I know acronyms. I’m learning a lot re: Bender’s request. No questions on my part, yet. I’ll ask Bender. BTW, Lucy, good CA101 link, IMHO. Thanks, Professor Curry. • Harold Posted Jan 19, 2010 at 8:28 AM | Permalink If perilous extrapolation is to be avoided, it seems much of climate science is imperiled. On a slightly different note: Judith said: The negative SWR feedback is basically an increase in the planetary albedo with increasing temperature, without a correspondingly large decrease in outgoing IR. How could this possibly be? The possibilities are: One glaring possibility not enumerated in what followed was that there is something(s) going on we haven’t figured out yet. This is always the most interesting situation to me, since it’s the first step in expanding understanding. • Judith Curry Posted Jan 24, 2010 at 8:10 AM | Permalink Bender, what else would you like me to comment on? Lindzen’s POV is focused solely on the tropics (and he mistakenly extrapolates what is going on in the tropics to global sensitivity/feedbacks). Do you mean the iris hypothesis and the tropical upper tropospheric water vapor feedback? 14. Judith Curry Posted Jan 18, 2010 at 2:12 PM | Permalink I have a short time to comment, more time this eve. I am not critical that they only examine the tropics, but that they claim that an examination of the tropics is sufficient to diagnose global sensitivity. It is not sufficient, as I outline in my post. • Dave Dardinger Posted Jan 18, 2010 at 2:33 PM | Permalink Dr. Curry, I don’t know if you’ve actually posted on this new version of CA before, but you might want to press the Reply button on the post you’re responding to so people will be able to tell who you’re answering. In this case I assume it’s the Mike B message of 12:07. Also, as this is likely to be a very long thread, you probably also need to have the CA Assistant (see top right corner of the page), as we’ve got nested comments which are a bear to navigate without it. • Judith Curry Posted Jan 18, 2010 at 4:24 PM | Permalink Thanks Dave, i think i have it figured out • JamesG Posted Jan 19, 2010 at 5:40 PM | Permalink I thought this was because most feedback papers do the same thing, on the assumption that the feedback will be strongest at the tropics, Dessler for example explains it here: http://pielkeclimatesci.wordpress.com/2010/01/06/guest-post-by-andrew-dessler-on-the-water-vapor-feedback/ “Q. Why is the tropical upper troposphere so important for the water vapor feedback? A. It is the changes in water vapor in the tropical upper troposphere that plays the major role in the water vapor feedback. While photons from these water vapor molecules do not directly heat the surface, they do primarily regulate emission of energy to space. Because the troposphere is rapidly mixed by convection at a rate much faster than radiation, the effect of changes due to radiation fluxes that are entirely internal to the troposphere (e.g., due to changes in lower tropospheric water) will be rapidly wiped out by convection and have a small net impact on surface temperature. The tropics dominate the effect because of the smaller temperature difference between the surface and the upper troposphere in the mid-latitudes combined with smaller column abundances of water vapor there. “ 15. mccall Posted Jan 18, 2010 at 2:21 PM | Permalink FWIW (and echoing Lubos point), Dr. Lindzen had acknowledge much of criticisms before Dr. Schmidt in a recent RC post went (arguably) well over the line in claims against the paper. As a result, a follow-up LC paper is circulating (not sure if it’s a re-issue or a corrigendum?) that addresses the criticisms with little/no impact on the “models still substantially overstate CO2 sensitivity” summary. • Judith Curry Posted Jan 18, 2010 at 4:28 PM | Permalink mccall, my original essay was written on Dec 26 (see a post i just made further down on the thread), before the recent spate of postings on RC and elsewhere. I haven’t seen the revised LC paper (would appreciate a link if anyone has it). Apart from the critique by Trenberth et al. (which is what i assume LC respond to), my main concerns are inferring global sensitivity from this analysis of the tropics, misapplication of the sensitivity analysis based on a global energy balance model to a specific region, and deficiencies in the overall approach to inferring climate sensitivity using this kind of model. • Judith Curry Posted Jan 18, 2010 at 4:45 PM | Permalink i am having difficulty posting this as a new comment, “awaiting moderation”, so i’m posting here for now I just noticed that Steve didn’t post my final version of this little essay, here is the updated text: ———— Update: this essay was finished on Dec 26. Steve decided to wait a few weeks before posting, given all the interest in climategate. Since then, there is now a formal reply to the LC paper that is in press in GRL by Trenberth, Fasullo, O’Dell and Wong, this is dicussed at RealClimate http://www.realclimate.org/index.php/archives/2010/01/first-published-response-to-lindzen-and-choi/ and http://www.realclimate.org/index.php/archives/2010/01/lindzen-and-choi-unraveled/. See also DotEarth http://dotearth.blogs.nytimes.com/2010/01/08/a-rebuttal-to-a-cool-climate-paper/ ———– • mccall Posted Jan 18, 2010 at 9:42 PM | Permalink Thank you for clarifying the time line, Dr. Curry. Dr. Lindzen’s comment and response (I’m trying to locate it) were to Gavin’s Schmidt’s RC posting. As I recall, RL mentioned that most of the scientific critiques came to him via e-mail or other direct correspondence (I believe Wong was mentioned specifically), but perhaps yours are some of them. RL’s more pointed retort was to GS/RC was that their criticisms were mostly/if not all addressed in “revision” , and the high CO2 sensitivity claims for modeling (including feedbacks), still could not be sustained. Absent the revision, it’s not clear whether L-C have addressed your questions/objections, nor any others (fundamental or stretch) for that matter… • Judith Curry Posted Jan 19, 2010 at 7:41 AM | Permalink I haven’t written to RL on this. When I wrote this around Xmas, all was quiet on the LC front in the blogosphere (as far as I can tell), then a recent flurry of activity and I haven’t caught up on all of it. would appreciate pointers (i’ve seen RC, dotearth, and now Lubos) 16. geo Posted Jan 18, 2010 at 2:30 PM | Permalink It might be nice if someone changed the author on this piece from “Steve McIntyre” to “Judith Curry”. 17. Jack Okie Posted Jan 18, 2010 at 2:48 PM | Permalink Thank you, Dr. Curry, for being willing to engage on this site. It is enormously helpful to understand the issues behind the findings. Please continue. • Judith Curry Posted Jan 18, 2010 at 4:28 PM | Permalink Jack, thanks for your comment! 18. LennartS Sollentuna Posted Jan 18, 2010 at 2:49 PM | Permalink Has anyone a pointer to the up-dated Lindzen & Choi paper? 19. Craig Loehle Posted Jan 18, 2010 at 3:39 PM | Permalink A fundamental problem with the models wrt feedbacks is that no one can point to a simple experiment/data set that would be/is a clear unambiguous test of the way the models handle feedbacks. Spencer has shown one type of interpretation error. Lindzen/Choi attempt another. The 2002 GEWEX workshop linked above shows that at that time there was no ability to assess these critical assumptions. Models are all about assumptions and the best way to evaluate a complex model is to evaluate the assumptions individually–much more tractable when we can’t do experiments on the system. Evaluation, for example, of the input forcing of clouds and aerosols shows that the different models use different input data which means 1) we are uncertain about these forcings and 2) models using different forcing but getting the same answer either have a kludge to force the right answer or can not all be right. Either way, a problem. If Lindzen or Spencer are even roughly right, the models have a big big problem. • Judith Curry Posted Jan 18, 2010 at 4:31 PM | Permalink Craig, thanks for your comment, i agree with most of what you say. If LC have identified a real problem in the models, it is almost certainly in the way that tropical deep convection is parameterized. This is a known deficiency in the models that is very vexing owing to the complex scales of motions involved and the difficulties in treating them with the relatively coarse resolution of global climate models • Judith Curry Posted Jan 18, 2010 at 6:38 PM | Permalink Actually, LC’s results are most consistent with a problem in treating low level clouds in the tropics • BLouis79 Posted Jan 21, 2010 at 3:44 PM | Permalink But don’t the models have problems dealing with low level cloud everywhere? And don’t low level clouds provide significant negative feedbacks? • Harold Posted Jan 19, 2010 at 8:38 AM | Permalink Craig – I’ve seen the comment from time to time that controlled experiments can’t be done to evaluate the models (IPCC says this as well). This is not true – there are a number of geoengineering proposals which could be used in exactly this way. People who don’t want geoengineering argue that it’s too risky, because something really bad might happen. On the other hand, they are usually arguing we have to do something because something really bad IS going to happen. Where’s the logic in that? In any event, I tend toward the previous modelers desire for a step function to analyze. 20. J. King Posted Jan 18, 2010 at 3:42 PM | Permalink Dr. Curry’s willingness to engage this community is incredibly refreshing and the comments give the lie to the statement that CA is populated by flat earthers and anti-science lunatics – most of us just want to understand. (Unfortunately my background will not permit me to say anything intelligent about Lindzen and Choi in this lifetime). • Judith Curry Posted Jan 18, 2010 at 4:33 PM | Permalink Thanks for your comment! • Leo G Posted Jan 18, 2010 at 5:39 PM | Permalink J. King, Your comment applies to me as well. Thanx! Dr. Curry, many thanx for your time here helping those of us to understand both sides of this debate in your forthright ans appealing way! Leo G • Judith Curry Posted Jan 18, 2010 at 6:17 PM | Permalink Thanks Leo 21. Judith Curry Posted Jan 18, 2010 at 3:50 PM | Permalink I just noticed that Steve didn’t post my final version of this little essay, here is the updated text: ———— Update: this essay was finished on Dec 26. Steve decided to wait a few weeks before posting, given all the interest in climategate. Since then, there is now a formal reply to the LC paper that is in press in GRL by Trenberth, Fasullo, O’Dell and Wong, this is dicussed at RealClimate http://www.realclimate.org/index.php/archives/2010/01/first-published-response-to-lindzen-and-choi/ and http://www.realclimate.org/index.php/archives/2010/01/lindzen-and-choi-unraveled/. See also DotEarth http://dotearth.blogs.nytimes.com/2010/01/08/a-rebuttal-to-a-cool-climate-paper/ ———– 22. D Bonson Posted Jan 18, 2010 at 4:28 PM | Permalink Thank you Judith Curry for engaging in these discussions. I would like to read from you a response to Luboš Motl’s post (Posted Jan 18, 2010 at 11:12 AM). • Judith Curry Posted Jan 18, 2010 at 6:07 PM | Permalink I’ve posted one now 23. Judith Curry Posted Jan 18, 2010 at 4:32 PM | Permalink I just noticed that Steve didn’t post my final version of this little essay, here is the updated text: ———— Update: this essay was finished on Dec 26. Steve decided to wait a few weeks before posting, given all the interest in climategate. Since then, there is now a formal reply to the LC paper that is in press in GRL by Trenberth, Fasullo, O’Dell and Wong, this is dicussed at RealClimate http://www.realclimate.org/index.php/archives/2010/01/first-published-response-to-lindzen-and-choi/ and http://www.realclimate.org/index.php/archives/2010/01/lindzen-and-choi-unraveled/. See also DotEarth http://dotearth.blogs.nytimes.com/2010/01/08/a-rebuttal-to-a-cool-climate-paper/ ———– 24. curious Posted Jan 18, 2010 at 4:56 PM | Permalink How come everyone here seems to think this post is by Judith Curry and not Steve McIntyre as the by-line would seem to indicate? • Craig Loehle Posted Jan 18, 2010 at 5:08 PM | Permalink steve warned us it was coming–and might I ask why you have a Bender picture as your icon? • curious Posted Jan 18, 2010 at 6:37 PM | Permalink I think it is to differentiate himself from me (or other curious types) as I don’t use a pic. 🙂 Good to see some civil debate on this thread! C • curious too Posted Jan 18, 2010 at 8:21 PM | Permalink Sorry didn’t realize there was already someone who was curious here! • bender Posted Jan 18, 2010 at 11:40 PM | Permalink yep, and there’s already a “bender” aka Mr Bending Rodriguez • Mike B. Posted Jan 18, 2010 at 5:13 PM | Permalink Hmmm. Good question. Perhaps someone could clear that up? • curious too Posted Jan 18, 2010 at 8:27 PM | Permalink Looks like the by-line is fixed now. 25. Slabadang Posted Jan 18, 2010 at 5:09 PM | Permalink LC´s method is genius! Because now the evidence of the “sensitive” climat has to be proofed by the warmists. I´m reading and cant find the logic of the critisism.LC underlines that they focused on the tropics.Where is the mesured evidence that the rest of the globe has less outgoing radiation?.And to compensate for the tropics it has to be far less than the “models” predicts.It just doesn´t add up! LC delivered a vital part of the understanding of the globes radiation.Curry seems to have no clue of radiation not in the tropics and not at the Arctic. RomanM: If you have a valid point to make, do it. There is no reason to include a personal attack on Dr. Curry. Please refrain from doing so in the future. • Derek H Posted Jan 18, 2010 at 9:44 PM | Permalink 1. Slabadang, please be polite. I would like to see more of Dr. Curry’s explanations here because she at least is striking a reasonable tone and appears to be intellectually honest. I have found Schmidt, Mann, Hansen et al to be rather strident, arrogant and either dishonest or misleading; I would like a conversation with AGW proponents who are none of that. 2. Can someone please translate this diatribe into English? • Steven Mosher Posted Jan 19, 2010 at 3:41 PM | Permalink No need to compare her behavior with others. The difference speaks for itself. • Bill Hammock Posted Jan 20, 2010 at 8:03 AM | Permalink Yeah, actions speak louder. So, you went on and published the personal e-mails of the CRU lab. And started screaming that ‘trick’ is paramount evidence of dishonesty, while it was just an innocent expression. You are among the group of people that polarized the discussion and polarising it even further. What Dr Curry is doing is exceptional. It is a single voice in the denier cacophony. • Bill Hammock Posted Jan 20, 2010 at 5:02 PM | Permalink It is exactly this type of replies that drive away people and diminish constructive criticism on these blogs. You give a vague reply and you quickly move on to something else. It is always moving to something else and never finishing a subject. The word ‘trick’ was used to describe a smart technique in the research work. Simply by looking into the context of the e-mails it screams that there was nothing suspicious about it. If the ‘trick’ was the best of the examples of suspicious work by the CRU, then, frankly, you are out of arguments. What I know is that those who polarize the debate are those without good arguments. • Bill Hammock Posted Jan 20, 2010 at 9:12 PM | Permalink OT • Bill Hammock Posted Jan 21, 2010 at 12:36 PM | Permalink So, what happened here is the moderator erased my text about the ‘trick/decline’ issue and replaced it with “OT”. A respectable moderator would at least mention that the ‘OT’ is his text, as they do in other blogs. For someone who wants to search, they can find that ‘trick/decline’ is a non-issue, and not dishonest at all. The denier club is losing it. Just like the smokers were misled by the tobacco industry in the 20th century and kept smoking, you are taken for a ride by the fossil fuel industry. Good luck to Dr Curry. Steve: This has nothing to do with Judy Curry’s post. • Posted Jan 21, 2010 at 12:51 PM | Permalink The moderator here is terse but effective. There’s no point comparing with other places – they probably haven’t kept focus on the science as well as this one has. ‘Hide the decline’ has been judged off topic, no doubt partly because it’s been dealt with so often before, for example on 10 Dec and before that on 26 Nov. They are worth careful study. 26. Jeremy Posted Jan 18, 2010 at 5:20 PM | Permalink JC, wrt your second point… why does a focus on shorter time-scale (Nth order) affects on climate negate LC results? If the climate system were an Nth order ordinary differential equation (which isn’t too far from the truth) the shorter time-scale affects would be the higher orders and should therefore have the capability to completely overwhelm the lower-order (N-x) influences, am I right? You seem to be making the point that since these lower-order (longer time-scale) issues are not considered in LC, then the study does not apply. This to me sounds like saying that in an equation such as this: X(t)= a*y(t) + b*z(t)^2 + c*w(t)^3 + d*u(t)^4 (where a,b,c,d,t,u,w,X,y,z >=0) …that the u(t)^4 term can be ignored in favor of the z(t)^2 term. This is mathematically difficult to justify unless your coefficients are known, and/or known to be non-linear. If we allow for that to be the case, which I will admit is a reasonable allowance given our limited analytical understanding of climate, then I should think that some very simple laboratory experiments would demonstrate these lower-order positive feed-backs to rule the day. We know that changes in cloud-cover affects the temperature of an area in a very short amount of time, and alters it more drastically than the longer-time-scale influences in perhaps hundreds of years. Put simply, why does a study of the large influence of shorter-time-scale negative feed-backs to global temperature not apply in a discussion of where climate will go simply because longer-time-scale positive feed-backs are not considered? Certainly that which has the greatest influence per unit time on temperature has the greatest potential to alter the total energy content of our atmosphere. • Dave Dardinger Posted Jan 18, 2010 at 6:00 PM | Permalink Put simply, why does a study of the large influence of shorter-time-scale negative feed-backs to global temperature not apply in a discussion of where climate will go simply because longer-time-scale positive feed-backs are not considered? This is the way of looking at things that I was needing. Let me try repeating what you’re saying, and you can tell me if I’ve got it right. If we have two feed backs, one a short time-scale negative feedback and the other a long time-scale positive feedback, the short time-scale negative feedback will keep the temperature rise (in the case we’re looking at)lower; essentially as much as it can be held low. So if the non-feedback rise was 2 deg C, then the negative feedback might keep it to .5 deg C. This, then, is what the positive feedback would have to work with, not the 2 deg C rise. It might eventually produce a large rise, but it would take an even longer time than in the 2 deg C case (assuming some sort of quasi-linear effect). And depending or what sort of negative feedback we’re dealing with, it’s feedback might become even more pronounced at higher temperatures thus dampening the positive feedback even more. • Judith Curry Posted Jan 18, 2010 at 6:14 PM | Permalink Jeremy, a focus on shorter (rather intermediate) time scales doesn’t negate their results, but it limits what kind of conclusions can be drawn from the analysis. Dave, you are hitting on an important point that the feedbacks are frequency dependent and the overall feedback for what might happen on the time scale of a century is a complex interaction of a number of nonlinear feedbacks operating over a range of time scales. Doubtful that all this can be captured by a simple linear feedback analysis such as LC • Dave Dardinger Posted Jan 18, 2010 at 6:29 PM | Permalink Doubtful that all this can be captured by a simple linear feedback analysis such as LC But can it be captured by GCMs either? The problem isn’t so much that a GCM can’t crunch the numbers properly, but have the physical problems been properly worked out and described in the literature so that they can be put into a GCM? The IPCC has admitted that things like clouds are the least known theoretically. That being the case, why should we reject what is essentially a cloud feedback by LC without careful examination of the properties it has? Not that I mind a discussion of the yeas and nays here. I welcome it and I hope that’s what happens. There’s also the point a couple of people have made that even if the LC formulation can’t be extended to a global formulation it still may be useful for estimating the feedback situation. • Judith Curry Posted Jan 18, 2010 at 6:36 PM | Permalink Dave, the LC method should ONLY be applied globally, it doesn’t make sense for a region, that is part of the problem. GCMs reproduce all scales of feedback processes, the question is whether cumulative feedbacks produced by the GCMs are correct. If LC attribute what they find to a cloud feedback, as i said in my post, the only cloud feedback that might have the fingerprint they find is one associated with low clouds over the ocean (that are unmasked by higher clouds above). There may be issues with how low clouds in the tropics are treated by the climate models, but how the climate models treat deep tropical convection seems to be the bigger issue (and was the issue that lindzen’s iris hypothesis raised). • Geoff Sherrington Posted Jan 18, 2010 at 10:06 PM | Permalink Is there a systematic change in the modelled feedbacks for each traditional latitude band that indicates that findings from non-tropical regions will weaken or strengthen the LC proposition? That is, if the proposition is found acceptable for the tropics, does it follow that the other global regions will be consistent with the conclusions, or inconsistent? So far, all I have read is that they will be different (as intuition might expect), but I have not read that they will be damaging or terminal. • Dave Dardinger Posted Jan 19, 2010 at 1:03 AM | Permalink If LC attribute what they find to a cloud feedback, as i said in my post, the only cloud feedback that might have the fingerprint they find is one associated with low clouds over the ocean (that are unmasked by higher clouds above). I’m somewhat confused here. I thought LC started out talking about increased CO2 resulting in the IR being emitted at a higher level, and thus a lower temperature, without considering feedbacks. LC then says the OLR increases (giving a negative feedback) in the Tropical ERBE readings while the models show decreases in OLR requiring a positive feedback in surface temperatures to produce equilibrium. Now LC don’t worry about the mechanism producing the increased OLR. Perhaps you’re correct as to what this would imply, but I’d like more than just hand waving as to low level; unmasked clouds being the “only” cloud feedback having the required fingerprint. Can you point me to a paper available online which discusses why this must be the case? I’d think there could be a variety of cloud configurations which would work. • Clif C Posted Jan 18, 2010 at 7:27 PM | Permalink I’m still waiting for the attractor. 27. 40 Shades of Green Posted Jan 18, 2010 at 5:41 PM | Permalink I would echo the general welcome already extended to to Professor Curry by others. 28. Judith Curry Posted Jan 18, 2010 at 6:09 PM | Permalink My new comments are “awaiting moderation” for some reason. Here is a comment without the web links (that may be causing the problem I just noticed that Steve didn’t post my final version of this little essay, here is the updated text: ———— Update: this essay was finished on Dec 26. Steve decided to wait a few weeks before posting, given all the interest in climategate. Since then, there is now a formal reply to the LC paper that is in press in GRL by Trenberth, Fasullo, O’Dell and Wong, this is dicussed at RealClimate. See also DotEarth ———– RomanM: Sorry, but the spam filter can be unpredictable in its choice of what to keep from posting. I have relaeased the three comments from limbo. 29. Posted Jan 18, 2010 at 6:14 PM | Permalink I glad to see that the L & C paper is now being audited on CA. Lord Monckton rather foolishly IMO has jumped on its conclusions rather as the experimental proof that the models are fundamental wrong in regard to climate sensitivity. OT here I know but if anyone here is interested in looking at ‘global warming’ in full colour then take a look at this new thread that I’ve just put up on ‘digginintheclay’. Mapping global warming The main conclusion reached in the thread is that global warming is hardly global and that based on the evidence shown in the colour coded trend maps presented in the thread, ‘global warming’ is not global but is in fact largely NH winter warming. I’ve stated that given what the maps show, it’s hard to see how CO2 could be the cause of this warming unless ‘Demon CO2’ is happy to allow notable exceptions by being choosey and selectively allowing warming in parts of the planet while allowing other parts to cool at the same time. I’ve suggested that Western Australians apply for a rebate on their carbon taxes and have also recommended where ‘pommies’ like myself should all go if we want a good tan this summer. Regards KevinUK 30. Tom Posted Jan 18, 2010 at 6:21 PM | Permalink This may go against SOP but the FOIA, CRU files reference these verses… Job 37:14-24 I think it was a PJ email, tongue in cheek I suspect; that deals with these very questions and answers. Quite apt. 31. Slabadang Posted Jan 18, 2010 at 6:23 PM | Permalink Kevin! How come your modells dont show what LC measured? 32. Steven Mosher Posted Jan 18, 2010 at 6:27 PM | Permalink Dr. Curry. Thanks for this. Click to access workshop02.pdf • Judith Curry Posted Jan 18, 2010 at 6:37 PM | Permalink Steve, you’re welcome. Google some of Rossow’s papers, he did a nonlinear feedback analysis of Lorenz’s simple chaotic model, shows how complex it is to analyze feedbacks in nonlinear systems • JamesG Posted Jan 19, 2010 at 6:01 PM | Permalink Which normally means we need to rely much more on real data than on models that neither capture all the physics nor the meteorology and have insufficient resolution anyway. 33. Kenneth Fritsch Posted Jan 18, 2010 at 6:54 PM | Permalink I am glad to see that CA is back to analyzing climate papers again as the climategate discussions were getting a bit stale. I have nothing to add to this discussion, by way of technical comments, but I note that when papers are criticized/analyzed, like the Santer and Douglass exchanges and this one, we see more readily just how uncertain the science and models really are. I did particularly like Craig Loehle’s post on having no experimental means of evaluating climate models (no absolute standard to compare to) and therefore we need to compare climate models and look for significant differences. If all models agreed, we might have more confidence that the chance of agreement is better than all being exactly wrong to the same extent. Significant differences between model results, of course, means that one might be correct and one wrong or they might both be wrong and different. I have been taking this approach in comparing temperature data sets. • Judith Curry Posted Jan 18, 2010 at 10:28 PM | Permalink Kenneth, I agree that it is good see something on the blogs other than climategate! Craig Loehle has hit the nail on the head in terms of lacking good experimental means of evaluating climate models. The gewex link at the end of my essay was targeted at trying to figure out how to more effectively use satellite data to evaluate climate models. this is very important, but the climate modelling community hasn’t really come up with good evaluation metrics LC (and Douglass et al) have been taking the approach of using satellite datasets for the tropics and evaluating climate models. The metrics they propose are reasonable for evaluating climate models (delta Flux/delta SST; trend of upper tropospheric temperatures). Most of the models don’t compare very will with observations. this is valuable information, especially if the observations can identify what aspects of the models need improving. My objection is when the results are over interpreted, i.e. in my opinion the LC analysis is not a convincing observational determination of climate sensitivity. • bender Posted Jan 18, 2010 at 11:46 PM | Permalink My objection is when the results are over interpreted Agreed. But, JC, is this a “fatal flaw” which would cause you to reject the paper outright, with no hope of publishing? Or would you reject with invitation to re-submit after dialing back the rhetoric? Or would you accept after “major revision”? I’m trying to understand to what degree you disagree with this paper. Sounds like you’re already back-pedaling somewhat from your original review (which is what I would have predicted). • Judith Curry Posted Jan 24, 2010 at 8:04 AM | Permalink If i had been a reviewer of this paper, i would say “major revisions needed.” I would have asked that they they focus their interpretation only on evaluating the climate model simulations of the tropics, and not claim that this is of any particular relevance to global climate sensitivity. In any event, the paper certainly should have been published, even if some of the analysis and much of the interpretation is not correct. This paper is stimulating further research and is challenging models with observations which is always valuable. The overhyping of this paper as “the end to the AGW hoax” is annoying, but at the end of the day not all that important (although this is the sort of thing that seem to send those involved in the hacked CRU emails into quite a tizzy). • Derek H Posted Jan 19, 2010 at 10:33 AM | Permalink Dr. Curry, I agree, I think some ideologues went way overboard in labeling LC as an AGW-killer but you raise a fundamental point that I never see the mainstream press cover and IPCC doesn’t seem to want to acknowledge: many of the models don’t compare well to actual observations. I am still astounded that apparently legitimate scientists accept paleoclimatological analysis using proxies when those proxies apparently don’t conform to the last 30-40 years of observations (and probably the BEST period of observation to boot!). Perhaps you could explain that paradox to me and I can get rid of the hobgoblin in the back of my mind? • Posted Jan 19, 2010 at 12:59 PM | Permalink I think people should keep the discussion to the paper and critique thereof… 34. eddy Posted Jan 18, 2010 at 7:07 PM | Permalink Many thanks for this, Judith. Overall: More of this kind of stuff, please! • Judith Curry Posted Jan 18, 2010 at 7:53 PM | Permalink Thanks eddy! 35. Alexander Harvey Posted Jan 18, 2010 at 7:17 PM | Permalink Judith, This is a bit of topic, but I wish to use this opportunity to highlight something that troubles me. Whenever the subject of feedbacks comes up I always get the unnerving feeling that I am in Denmark with Hamlet yet the ghost never appears. The ghost in question being phase. Feedbacks, both positive and negetive, have important implications for both diurnal and seasonal phase lags. Focussing on seasonal lags, it seems to me that there may be very little room for large feedback effects either positive or negetive. The seasonal lags can, with care, be computed from a gridded climatology like the Hadley Centre´s. In the midlatitudes, they show interesting global patterns reflecting both the differences between land and sea but more crucially the effects of the prevailing winds and ocean gyres as they move phase about. These efects I feel can be emulated by a model that moves phasors associated with the atmosphere and oceans across the globe. In such an approach the positions of the phase minimum located around Mongolia and the phase maximum in the Pacific can be accounted for. Now large positive feedbacks would have the tendency to increase phase lags noticiably, and negetive ones to minimise them. Also any latency in the feedback response could cause instability to set in where the simpler approach of ignoring phase would not have indicated it. Such latencies particularly if cascaded could give rise to instability with negetive feedbacks once the total phase lag of the signal fed back exceeds pi/2. In my opinion, and it is just that an opinion, the phase lags that can be computed from the Hadley climatology are consistent with low levels of feedback. The seasonal phase over the oceans falls largely in the range (pi/3 +/- pi/12) which is quite moderate. It seems to me that large amounts of positive feedback, particular with a latency that is a significant proportion of one year, would indicate infeasible small underlying phase lags. Also I doubt that signifiacnt negetive feedbacks with similar significant latencies could be applied without triggering instabilities. I did quite a lot of work on this many moons ago and I should like for someone with the resources and necessary wit to consider such an approach. I admit that it would probably only provide a sense of the limits to feedbacks as the phasors associated with the land, ocean and atmosphere are hard to model with great accuracy, as is the modelling of the interaction of the phasors in relative motion, but the good news is that the data is I think quite robust in that great time and care has gone into the production of the climatologies, which I feel is evidenced by the coherent and detailed phase picture one gets from an analysis of the midlatitudual Northern Hemisphere. As an aside: the climatologies also contain important information regarding linearity due to what I consider to be very low production of harmonics over the land masses. Over the oceans there is detectable non-linearity possible due to the seasonal augmentation and suppression of the well mixed layer. I thank all for reading this far, it is appreciated. I do have procedures and results (but not readily to hand) for those that are interested. Alex • Judith Curry Posted Jan 18, 2010 at 7:52 PM | Permalink Alex, very good point. the feedbacks are frequency depended with complex phase lags and they are not easy to diagnose. The black box approach of LC “hopes” that there is some averaging period where there is equilibrium, and then the higher frequency feedbacks within this averaging period can be diagnosed. The LC approach is also based upon a linear feedback analysis. The fact that isolated feedback processes (e.g. water vapor, ice albedo) are not additive reflects the nonlinearity of the system and gives hints why the LC approach is inadequate. I would be interested in hearing more about your thoughts on this 36. vg Posted Jan 18, 2010 at 7:31 PM | Permalink I find this statement extraordinary “Cameron, the fingerprint of of AGW is most evident in the high latitudes of the northern hemisphere” Please explain why Why not the low latitudes of the southern hemisphere? The whole concept of “global” has been suddenly thrown out Thanks for your contribution Dr Curry this is more like a real debate 37. Robert of Ottawa Posted Jan 18, 2010 at 7:36 PM | Permalink Dr. Curry, I am not the first to pick this up: The basic assumptions behind this type of sensitivity analysis based on top of atmosphere fluxes used by LC need to be tested by climate models. Surely, it must be the other way around, the models being tested by the observations? Ans another thing: parameterization does not a physical process make. • Judith Curry Posted Jan 18, 2010 at 7:47 PM | Permalink Robert, the point i was trying to make is this. the LC model is a very simple one. If such very simple models could simulate the climate accurately, there would be no need for more complex models. What LC have done is take a parameter (sensitivity) derived from a very simple climate model and has adopted it as a metric for comparing observations and more complex climate models. No matter what the metric, one would expect the observations and models to agree if the models are to have any credibility. The question that i am asking is whether this metric used by LC is an appropriate one to assess the feedbacks/sensitivity of the global climate (and of the climate models). Whether this metric is a good one, and to what extent the assumptions that go into the simple climate model make sense could be evaluated using one of the more complex climate models. Such as whether the climate model ever shows “equilibrium” of the radiation fluxes at the top of the modeled atmosphere. And to what extent changes in the TOA radiative fluxes are coupled to changes in surface temperature. Etc. • RB Posted Jan 18, 2010 at 7:58 PM | Permalink Dr. Curry, I suppose there are multiple feedback effects on timescales of similar magnitudes? I’m guessing first order linear analysis may be OK if one feedback is over 100 years while another is over a million years and has an effect that is small over 100 years and can therefore be neglected – I guess you are saying that is not the case. • RB Posted Jan 18, 2010 at 8:10 PM | Permalink I guess you partially answered this in reply above • WillR Posted Jan 18, 2010 at 8:39 PM | Permalink I guess I look at this the other way. If complex models that could simulate the climate accurately existed — there would not be any need for simple models that model a region or system. I guess the other question that could arise from this is: Would we be better off with a monolithic model that “did everything” or a series of simpler models that interconnect and communicate — say like self organizing systems (see Holonics). Systems like this can often be given a “goal” and told to find a way to achieve a goal given historical figures and what the historical results were. Point being simply — that there is more than one way to model. I prefer the “self organizing” models that are composed of a series of (simple?) models that solve part of the problem and communicate goals and feed-backs to other systems. It is essentially a set of NP problems (imo) and that is one way that has been shown to work in other branches of math and science. Any further discussion would be a side track — but I did want to make the point that they made a modeling choice — then made their case (fwiw) — and that there are ways to incorporate that research into other models through coupling mechanisms. I only made this post since you opened the door to modeling techniques — and I will now try to be quiet — since it does seem to be a digression. Aside from that … appreciate your passing on your thoughts. • Carrick Posted Jan 19, 2010 at 1:26 AM | Permalink WillR: I guess I look at this the other way. If complex models that could simulate the climate accurately existed — there would not be any need for simple models that model a region or system. . If you can replicate the important phenomenology with a simplified model, it is informative of what assumptions need to be made to replicate that phenomenology. So what you claim is not true. Simple models often provide insight that complex models cannot. • WillR Posted Jan 19, 2010 at 6:18 AM | Permalink I think that was my point… • Clif C Posted Jan 18, 2010 at 9:43 PM | Permalink Dr. Curry, thank you for your post and for stimulating an interesting discussion. If I understand you correctly, you are saying (here and there) that the models LC employ are elegant but flawed. That LC indirectly “measure” a real physical property of the tropical atmosphere, but that the results may be off. That LC’s “measurements” cannot be compared to corresponding results of the GCMs because the GCM’s are global by design and cannot be falsified at a local scale. 38. Geoff Sherrington Posted Jan 18, 2010 at 8:18 PM | Permalink For what it is worth, the UAH lower troposphere tropics temperature over oceans was lower in the hot year 1998 than in the more normal prior year 1997. The difference was about -1 deg C. I wonder how this observation can be woven into model explanations? • Ron Cram Posted Jan 19, 2010 at 12:41 AM | Permalink Geoff, that is fascinating and counter-intuitive. Do you have a link for that? • Geoff Sherrington Posted Jan 19, 2010 at 1:45 AM | Permalink Ron, It was wrong. I read the wrong graph. Corrected below. Apologies. 39. Tom Posted Jan 18, 2010 at 9:04 PM | Permalink I would like to ask those who are able to answer; this question. In real terms, you all agree that mankind, has the ability to determine with certinty the 1/2 a degree(F),(or less) global warming over the last ten years? The world we live upon is very dynamic, to say the least. The over-under has got to be at least a degree or more:), on a planet like ours. You all think you really can get it down to fractions? And over these time frames and all this “new science” right out of the box?… It bogles my mind to think you can be so certin about such small measurements over such a vast area. • jaffa Posted Jan 18, 2010 at 10:39 PM | Permalink This point was the source of my initial scepticism. I simply can’t believe that ‘global temperature’ can be determined to fractions of a degree even using measurement data. Once proxies are introduced the whole claim seems to be ridiculous. Can someone clarify what is meant by ‘global temperature’? Is it an average of daily maximums, daily means or something else. How do those analysing the data compensate for the huge (in relation to the suggested warming trend) daily or even hourly changes, day/night changes, seasonal variation and longer term natural variations. Finally I had assumed that high quality, reliable data was being used, with satellite measurements covering the entire globe. Isn’t it? Perhaps someone will be kind enough to point those of us who think there are lies, damned lies, and statistics to a web page that makes sense of the apparent nonsense. • See - owe to Rich Posted Jan 19, 2010 at 5:13 PM | Permalink Jaffa, the law of large numbers means that a lot of data with medium uncertainty can be averaged to give a result with fairly low uncertainty. By looking at, for example, the HadCRUT3 data and the UAH satellite data you can get a feel for how much uncertainty. The standard deviation of the last 10 years’ of HadCRUT3 annual data is less than 0.1degC I think. Nevertheless, the law of large numbers requires that independent errors be allowed to cancel out, approximately. Hence auditting is vital to ensure that biasses, accidental or otherwise, are not allowed to creep in. HTH, Rich. • Carl Gullans Posted Jan 20, 2010 at 8:59 AM | Permalink Yes… although you are aware of this, I’ll mention this for jaffa and others who are unfamiliar… HadCRUT3 is *known* to have biases that have creeped in. So, you are right that using adjusted data, and claiming a low standard deviation, is completely wrong. Somebody wrote a pretty good explanation of this recently (ChiefIO, Bishop Hill? anybody who knows what I’m talking about, please help), showing that all adjustment decisions need to carry an error term going forward, but that they currently don’t. • Steven Mosher Posted Jan 19, 2010 at 11:37 PM | Permalink Jaffa. The “average” is estimated as follows Tave= (Tmax+Tmin)/2 That is the max of the day and the min of the day are added and divided by 2. It’s been shown ( and you can see for yourself by looking at CRN data which is collected every 5 minutes) that this is roughly equal to the integrated value of temperature for a 24 cycle. As others have noted the law of large numbers is your friend and you get a good estimate of the monthly average ( arguments about the physical ‘reality’ of this number abound) The most important thing however is the understanding of the tren over long periods of time. Good places to start are Lucias blog which addresses some of these issues and threads all over CA. 40. Michael Larkin Posted Jan 18, 2010 at 10:32 PM | Permalink So refreshing to see some good-natured interaction going on here – kudos to Dr. Curry. As things got nearer the end of the thread (not being an expert), I began to get some idea what the issues were. Even for a bozo like me, it’s very plain: this is very far from settled science and we could do with much, much more of this constructive dialogue. Keep building bridges, Dr. Curry! 🙂 • Judith Curry Posted Jan 19, 2010 at 6:47 PM | Permalink Michael, thanks for your comment 41. Charlie A Posted Jan 18, 2010 at 10:36 PM | Permalink Dr Curry, I would greatly appreciate any comments you have on 1) the general validity of using the seasonal temperature variations to estimate climate sensitivity, and 2) comments on the specific web article by George White at http://www.palisad.com/co2/eb/eb.html . It seems that he arrives at roughly the same sensitivity number as Linzen and Choi, but using a relatively independent approach. • Posted Jan 19, 2010 at 2:02 PM | Permalink Re: Charlie A (Jan 18 22:36), George White’s science looks to be of high quality. Using the same data, but separating hemispheric seasonal fluctuations, he has far bigger quantities, he can easily show phase lag where we have always felt it: the lag of season behind solstice, and he can quantify the energy balance to show a strong negative sensitivity. He’s done original quality work in other areas too, but he seems to be beyond the horizon of well-known work. I think he deserves a hearing. • Judith Curry Posted Jan 19, 2010 at 6:47 PM | Permalink Charlie, this is an interesting analysis, i’ll have to take some time going through this, thanks for pointing it out. 42. Posted Jan 18, 2010 at 10:42 PM | Permalink Judith, thanks for posting here and making yourself accessible for responses, very much appreciated. I have only an elementary grasp of the physics of climate but when I read LC I didn’t really understand what they were demonstrating. The problem seems to be in the timescales. Here’s one dilemma for the LC method – if the radiative imbalance with 380ppm CO2 is the primary force causing the climate to warm up, how come some years it cools down? 365 days of +2W/m^2 radiative forcing – and yet cooler at the end. Unless we just aren’t very good at measuring temperature it seem as though the feedbacks are many and varied in time response. So how can analyzing one part of the globe (as you point out, energy flows in and out of the region are also a problem for a non-global analysis) over short time scales capture the response? If their method worked, wouldn’t we in fact have monotonic increases in temperature year on year? Seems like a problem for climate models too, but that’s another story. And do you have a link to more chapters of your book? • Judith Curry Posted Jan 19, 2010 at 6:34 PM | Permalink There are feedbacks and also forcings. a forcing is “A causes B”. A feedback occurs if the change in B causes a further change in A. In the LC data set, coolings are caused by La Nina events and teleconnections with the higher latitudes. This further emphasizes the problems with trying to do this analysis for a limited region Re chapters in my book, if you send me an email, i can email you some more chapters (my publishers won’t let me post major portions of it on the web). 43. Dirk Posted Jan 18, 2010 at 11:42 PM | Permalink Perhaps I misunderstood, or something was mis-spoken, but the change in which stations are included in the temperature data over the past 30 years- removing high-altitude stations- couldn’t this account for the tropical land “warming”? One example given on the recent John Coleman report was that Bolivia’s temperature is now determined by low-altitude stations as high altitude stations are no longer included- creating a big red warming blip over Bolivia- true? Also- why would altitude change the properties of radiation relative to wavelength and CO2- is it concentration, or temperature, or albedo of sea ice vs. oil vs. plastic vs. plankton? So what are the possible variables that create concern over applying the results of the tropics to global behavior? • Judith Curry Posted Jan 19, 2010 at 8:09 AM | Permalink Dirk, different regions with differing surface characteristics, clouds, water vapor, and locations with different wind conditions respond differently to global forcing and the natural modes of climate variability. See the links i posted for the global maps of observed and projected temperature change in my original post. The tropics are characterized by low surface reflectivity, a very large amount of water vapor, and clouds that include deep convection and high cirrus. The tropical atmosphere is fairly opaque to IR radiation, and the so called “water vapor window” has a shutter over it in the lower tropical atmosphere (owing to large concentrations of water vapor) so virtually no IR emitted by the surface reaches space as it is absorbed in the lower atmosphere. This is why most of the radiative action in the tropics is in the upper troposphere. • Patrik Posted Jan 19, 2010 at 10:07 AM | Permalink Dr Curry>> “The tropical atmosphere is fairly opaque to IR radiation, and the so called “water vapor window” has a shutter over it in the lower tropical atmosphere (owing to large concentrations of water vapor) so virtually no IR emitted by the surface reaches space as it is absorbed in the lower atmosphere.” Isn’t this a very short and precise description for why the LC paper is meaningless? Because, if what You write is true, then added CO2 should make no difference what so ever in the tropics – right? But, wouldn’t it also then be true that the man made warming can’t happen in the tropics, because there is no additional greenhouse effect possible in the tropics? The question that follows is; in the polar regions we have virually no water vapor, right? So, even though added CO2 can warm the polar regions to some extent (~1-2 K?) – no feedbacks will happen there, because of the non-existing water vapor? So, the extra greenhouse effect that our CO2 can create, can only really appear and interact with H2O in other climate zones? Temperate, sub-tropical (?)..? Is my reasoning correct? • Judith Curry Posted Jan 19, 2010 at 6:29 PM | Permalink Patrik, you make an important point. Because of the opacity of the tropical atmosphere, doubling of CO2 had very little direct effect on the tropical surface temperature, most of the radiation action in the tropics is in the upper troposphere where there is less water vapor. This reasoning further makes the point that it doesn’t make any sense to try to determine the climate sensitivity from an analysis of the tropics • Patrik Posted Jan 20, 2010 at 10:18 AM | Permalink Ok, thanks for the answer. Could You please explain what kind of “action” You’re referring to in the upper troposphere above the tropics – since no LWR will reach the higher layers? • David A Posted Jan 26, 2010 at 11:40 AM | Permalink I thought that one of the points of the paper was that the “models” did show a warming in the very area studied and the observations were of the opposite sign. If that is true then there would be an affect on the models overal AGW projection based on the percentage of the overall “warming” affected by new understanding of this area limited area. 44. vg Posted Jan 19, 2010 at 1:40 AM | Permalink How in earth could anyone think that they would be able to model climate? you would have to be able to predict the suns behaviour Sunspots, geomagnetism), our surrounding universe (cosmic rays). Earths wobble etc… One example: Its been raining cats and dogs here in Australia “drought is over” even with a significant El Nino. The most we could hope for is short term forecasting and even then it doesn’t seem to work. For factual data (correct me if I am wrong) the only models that seem to predict significant events correctly such as the recent cold spell in Europe, based on sun activity are those put out by Farmers Journal is it Piers Corbyn? 45. Geoff Sherrington Posted Jan 19, 2010 at 1:44 AM | Permalink There are several experimental designs that might answer some questions sought. (What are the questions that the models are supposed to answer? They were never clearly expressed in any case). One experimental design starts with a small but stong anomalous result in a certain place/time, whose authenticity is accepted, then attempts to model and explain it. At the other end, we have the teraflop approach which seems to be heading to an answer of 42 +/- 42. In a sense, the teraflop approach needs to explain the annoying discrete anomalies as part of its validation. Sure, one cannot predict the date of the next big volcanic eruption and factor that in. But one should be able to model an abrupt temperature shift from causes as yet undescribed. That is why I am working away at the 1998 hot year, trying to find a mechanism that can explain matters like (a) did the hot time start at an identifiable locus? (b) can we measure how far it spread? (c) how fast it spread? (d) how high? (e) generally, any systematics that might be used in future prediction. (Yes, I know it was El Nino time, but cause and effect are blurred). Here are 2 graphs. Source data are at http://vortex.nsstc.uah.edu/data/msu/t2lt/uahncdc.lt They use UAH satellite temperature data over oceans then land, for the lower troposphere in the usual way. For a simple analysis, I subtracted the 13 months before the hot spot from the 13 months of the hot spot. If anyone can tell me that there are useful systematics in this, then I’ll be thankful. It goes back to the electronics black box approach above. I see nothing wrong in looking at input/output balance in a defined region, even if the processes in the black boxes above and below are very complex. The motor car moves because of a complex series of sub-actions, but its speed can be read from one dial. Still, it’s nice to know the Carnot cycle is involved. I have great difficulty in attribution of any part played by GHGs. The global tropical troposphere does not light up for a year, or even a few months, then cool again, by any GHG mechanisms I can describe. The graphs follow: Note: Above, Geoff Sherrington Posted Jan 18, 2010 at 8:18 PM, was in the wrong sense. The 1998-1997 change was +1 deg C over the tropical oceans as the graph shows. • Judith Curry Posted Jan 19, 2010 at 8:00 AM | Permalink Geoff, the 1998 year was El Nino, the biggest in a century. It was overhyped as the warmest year with an attribution to global warming. Those who hyped it in this way now need to answer why there have been some relatively cool years recently (and they’re not doing a very good job, “travesty” and all that). It doesn’t make scientific sense to focus on individual years and attribute a warm or cold spike (which are almost certainly associated with El Nino La Nina) to AGW. • EddieO Posted Jan 19, 2010 at 1:47 PM | Permalink Judith, I detect a healthy dose of scepticism in your comments. But then all scientists should be sceptical, either by nature or by training, and be proud to wear there scepticism as a badge of honour. My mantra for dealing with over enthusiastic claims of imminent catastrophic climate change is simply “prove it”. It’s the scientific approach! • Geoff Sherrington Posted Jan 20, 2010 at 1:04 AM | Permalink Judith, To the contrary, it makes eminent sense to get the micros right as a prelude to getting the macros right. El Nino has been put forward as an explanation for the robustly hot 1998 year, but it explains nothing. What has to be explained is the source of energy that ended up as heat over the circum-global equatorial region for a few months. Mere cross-Pacific heat redistribution that we know to accompany an El Nino is not enough. Where did the excess heat come from and where did it go in 1998? • EdeF Posted Jan 19, 2010 at 11:12 AM | Permalink I have access to a very reliable analog computer that can model the world’s climate for any region. It runs at 1.0000000 real time. • WillR Posted Jan 19, 2010 at 2:11 PM | Permalink Unfortunately it recently did a cold boot. Perhaps your questions can be answered when it warms up… • Posted Jan 19, 2010 at 2:14 PM | Permalink At least that may have killed off some of the viruses (with which it is said to be teeming). • Charlie A Posted Jan 19, 2010 at 2:26 PM | Permalink Unfortunately, although I have the same model of analog computer, it isn’t very well instrumented and it’s difficult to get a good readout. • Posted Jan 19, 2010 at 3:03 PM | Permalink Re: Geoff Sherrington (Jan 19 01:44), Geoff, I’m sure I saw something that spoke re 1998 along the celestial harmonics line – try people analysing cycles. Oceans have the rogue wave, rare enough that it was once thought to be a myth. 46. Colin_d Posted Jan 19, 2010 at 1:54 AM | Permalink Thanks Judith and CA for the review and discussion. A pebble in my shoe in all of this is how it contrasts starkly with the “Science is Settled” mantra. It is far from settled and it is bordering on obscene that far-reaching changes are attempted based on one possible interpretation of this ‘unsettled science’. • Judith Curry Posted Jan 19, 2010 at 7:56 AM | Permalink Colin, the “science is settled” is political statement, not a scientific statement! Decision making under uncertainty is often done facing conditions of much greater uncertainty than say the IPCC report. Even the IPCC report gives probabilities for its conclusions (none reach 100%, although they do say that warming in the latter half of the 20th century is “unequivocal”) • Kenneth Fritsch Posted Jan 19, 2010 at 10:23 AM | Permalink From this link and excerpts from it below we have the following development from the IPCC on “the science is settled”. That that statement in quotes is a political one and not a scientific one is rather evident from this new development. It further impugnes the science bent of the IPCC versus what many here have seen as a political one. Please note that the probability of this predicted occurrence was agreed by “experts” to be better than 90%. Since the IPCC only reveals in very general terms how the experts establish these probabilities we can only guess that it might have the average of the guesses of experts and on a subject that they were obviously oblivios. Two years ago the Intergovernmental Panel on Climate Change (IPCC) issued a benchmark report that was claimed to incorporate the latest and most detailed research into the impact of global warming. A central claim was the world’s glaciers were melting so fast that those in the Himalayas could vanish by 2035. In the past few days the scientists behind the warning have admitted that it was based on a news story in the New Scientist, a popular science journal, published eight years before the IPCC’s 2007 report. ..Hasnain has since admitted that the claim was “speculation” and was not supported by any formal research. If confirmed it would be one of the most serious failures yet seen in climate research. The IPCC was set up precisely to ensure that world leaders had the best possible scientific advice on climate change. …When finally published, the IPCC report did give its source as the WWF study but went further, suggesting the likelihood of the glaciers melting was “very high”. The IPCC defines this as having a probability of greater than 90%. The report read: “Glaciers in the Himalaya are receding faster than in any other part of the world and, if the present rate continues, the likelihood of them disappearing by the year 2035 and perhaps sooner is very high if the Earth keeps warming at the current rate.” • Judith Curry Posted Jan 19, 2010 at 6:25 PM | Permalink I am astonished and disgusted by this. I am surprised that something like this snuck all the way through the IPCC review process. This is a much bigger black eye to the IPCC than hockeystick stuff • Dave Dardinger Posted Jan 19, 2010 at 6:59 PM | Permalink Dr. Curry, I have a question which I’d not normally ask here, but since Steve Mc is off enjoying the other sort of curry, I’ll throw it out. We have a fair number of people who come here and say that maybe the the hockeystick is bent and some of the team members aren’t paragons of virtue when it comes to upholding the Scientific Method, what about all the other things which point to AGW? Inevitably they point to melting glaciers at the top of their list of AGW proofs. Given the recent hit or two on the glacier front, could you give us the top 3 or 5 or whatever research areas which are clear indications that AGW exists at a measurable level? • Posted Jan 20, 2010 at 1:20 AM | Permalink Yes, it’s always said – it was said from the moment Climategate broke until now, it is said once again about the ridiculous Himalayan glacier situation – that this particular bad stuff doesn’t for one moment affect the other crucial evidence for AGW, evidence so significant, it is implied, as to warrant exactly the same worldwide reductions in carbon emissions as were being proposed already. But what precisely are those other lines of evidence? Not just evidence for warming, but for the fact that increased CO2 in the atmosphere is causing the warming and that this might lead to catastrophic results in the future. Polar bears? Sea levels? Where’s the anthropogenic stamp? As Dave says, what are the top 3-5 lines that are unaffected by what we’ve seen just these last two months? • Steven Mosher Posted Jan 19, 2010 at 3:23 PM | Permalink Thanks for that Dr. Curry, you’re a closet Lukewarmer I bet. As you note “the science is settled” is a political or sociological statement. All science has uncertainty and the vast majority of our behavior is carried out under the conditions of uncertainty, and even where its not, as in math and logic it’s good practice to check your work twice. • Judith Curry Posted Jan 19, 2010 at 6:21 PM | Permalink I am a “warmer” in the sense that I think the best evidence supports a warming of 1.5-4.5C for a doubling of CO2, all other things being equal. All other things are unlikely to be the same as in the 20th century such as solar variability, volcanic eruptions, ocean oscillations, so what the actual temperature change will be in the 21st century, i have no idea. I am a lukewarmer in the sense that I don’t see catastrophic impacts (except at the higher end of the warming). • Steven Mosher Posted Jan 19, 2010 at 6:54 PM | Permalink Basically, a Lukewarmer doesnt deny radiative physics, so GHGs warm the planet. We push for open access. We tend to hover around the 1.5C-2.0C figures. So. FUD F= Fear and alarmism. U= Uncertainity and the next right action. D= Denial and inaction We are in the U camp. • Posted Jan 20, 2010 at 1:10 AM | Permalink I’m in the U camp. But I’m not sure the camp agrees what the ‘next right action’ is. This may not the place for the full discussion but it’s a neat triangulation, thanks. • Carrick Posted Jan 22, 2010 at 2:28 AM | Permalink Steven Mosher: We tend to hover around the 1.5C-2.0C figures. Well, dang, I guess I’m not a luke warmer after all! Based on the known physics, I put climate sensitivity at around 2.5°C (range of 1.8 to 3.5C, I use geometric mean since that’s more appropriate for a log-normal variable like CO2 climate sensitivity). I allow that unknown physics could push this number lower, but I’m a firm believer that one makes one’s estimates based on what is known now, not speculation of what we’ll discover in the future. I guess I’m a catastrophe denier. Like Judith, I don’t see catastrophic consequences for humanity unless you get to the high end of allowed climate sensitivities. • See - owe to Rich Posted Jan 20, 2010 at 3:42 PM | Permalink Judith, I’m afraid they’re not going to let you into the lukewarmer club if your range goes as high as 4.5C. That really could mean serious heating by 2100AD. Mosher has set a range (1.5,2.0), but my personal lukewarmness has a range (1.0,2.0). It’s at the upper end if I believe HadCRUT3 has properly taken account of UHI – but I don’t really. I can’t quite explain why, but my faith in HadCRUT3 has taken a bit of a battering in the last few months. Rich. • Judith Curry Posted Jan 20, 2010 at 5:20 PM | Permalink The possibility for the upper end >3C is some tipping point like massive release of carbon from the tundra. This could be caused by combination CO2 and increasing soot on the surface in high latitudes. I would say that my range 1.5-4.5C is likely (>66% probability). Note IPCC said very unlikely that it is less than 2C. I can’t agree with that. • WillR Posted Jan 20, 2010 at 5:34 PM | Permalink Judith Said… The possibility for the upper end >3C is some tipping point like massive release of carbon from the tundra. That should be easy to test on Canadian tundra. Check the temperature records, move some tundra into a “greenhouse” increase average temperature and watch the CO2 Off gas… Better still, I guess you could even put the structure in place the summer before — measure the following summer. You could do a number of structures — and maintain a control even… Considering the$ at risk — I would think it would be easy to promote this experiment. Cheap experiment — maybe there’s some funding…???

• Tolz
Posted Jan 20, 2010 at 4:16 PM | Permalink

Dr. Curry, doesn’t this “best evidence” have to include GCM predictions? And doesn’t even the low range of sensitivity require, basically, a positive cloud feedback?

• Judith Curry
Posted Jan 20, 2010 at 5:20 PM | Permalink

I think you lukewarmer guys are “uncertainty deniers”, with such a narrow range 🙂

• Colin_d
Posted Jan 19, 2010 at 3:27 PM | Permalink

Judith, a political slogan like that does not survive in isolation and the problem for science (and maybe its credibility will take a hit as a result) is that this one has been sustained by scientists. According to the IPCC it has been sustained by thousands of scientists and you don’t hear any protests from them that they are being mis-represented. As a scientist a lot of my work is at the interface between politics and science and it can be difficult to dampen the political desire for settled science.

• Judith Curry
Posted Jan 19, 2010 at 6:23 PM | Permalink

Other than a few scientists that are actively advocating policies, i don’t think any scientists publicly state that the science is settled.

Posted Jan 19, 2010 at 8:26 PM | Permalink

Dr. Curry, it is these “political” scientists that get the press and set the tone for all of the discourse on CO2 and its relationship to climate. Thus we find ourselves where we are with even good scientists like yourself referring to those of that don’t agree with the political “science” as “deniers”.

• bender
Posted Jan 19, 2010 at 6:59 PM | Permalink

The arguemnt is not that “the science is settled”. That is a straw man well-loved in the denialosphere. The actual argument is that “the science is settled ENOUGH” … enough to move forward with mitigation and/or adaptation based on a precautionary principle. Very, very different things.

• Posted Jan 19, 2010 at 7:55 PM | Permalink

I agree and disagree. “The science is settled” has been proclaimed by not a few activists, so it’s not just a straw man set up by those opposed to them. All on the basis of 2000 of the “world’s top climate scientists” said to be full agreement with the IPCC. I’ve even seen people liken the situation to that of Einstein’s General Relativity – it’s that proven. But they never quite say what is that proven and that’s the catch.

Of course you’re right about how the IPCC presents it. 90% confidence and all that (apart from the Himalayan glaciers disappearing by 2035 as it turns out). And I’m sure that’s the way Professor Curry would couch things, though with her own numbers.

So, settled enough to make sure Africa’s coal never provides the cheap electricity it could for those that currently have none? How settled does the science need to be for that, for example, to be a legitimate consequence? (I know that’s out of bounds generally on CA and I’m sorry. Yet the idea of “settled enough” inevitably brings in all the myriad consequences, on both sides. I especially respect Freeman Dyson’s fervent championing of coal to power the poor out of poverty. You have to get your hands dirty with such questions in this area, it seems to me. Hard. Important.)

• bender
Posted Jan 19, 2010 at 9:28 PM | Permalink

The science is settled” has been proclaimed by not a few activists

.
Do you really care what the activists proclaim? If so, why? Are we here to probe the science, or to correct activist misconceptions? I’m not interested in the latter. Plenty of blogospheric bandwidth devoted to that already. I don’t see that as CA’s function at all.

• Posted Jan 20, 2010 at 12:57 AM | Permalink

What about an activist called James Hansen and his comment about trains delivering coal to power stations being like “death trains”? Of course I care about what that very influential activist says, because of the potential impact on every human being of that way of thinking – especially those that do not already have electricity. And I also agree with you therefore – all the more so – that we should attend to the science. It was the phrase “settled enough” that did for me. You’re completely right to suggest that’s the right phrase. It’s incorrect though to say that the claim of settled science is only a straw man. It’s much more difficult than that.

• WillR
Posted Jan 20, 2010 at 5:37 PM | Permalink

Steve: snip – OT

• Posted Jan 20, 2010 at 5:58 PM | Permalink

Nah, I’ve seen the films, it’s got to be a beautiful woman tied up on the line and they always get released in the nick of time. I’m on the roof dueling with Hansen and sure enough, the tunnel just came into sight.

Meanwhile, your second para was just getting interesting when the last clause caused my parser to fail. Is the bug my end or yours?

47. Posted Jan 19, 2010 at 2:26 AM | Permalink

There is an interesting article by an Hungarian planetary scientist IDŐJÁRÁS published in: Quarterly Journal of the Hungarian Meteorological Service Vol. 111, No. 1, January–March 2007, pp. 1–40 “Greenhouse effect in semi-transparent planetary atmospheres”
This is a theoretical derivation of the Greenhouse effect from basic physics. The article points out that the boundary conditions arising from the assumption of a semi infinite atmosphere leads to problems in the conservation of energy and to an assumption of a temperature discontinuity at the surface.
Today’s models are all based on the those erranous assumtions. These two problems ultimately lead to a significant over estimate of the greenhouse sensitivity. The article estimates that the expected global temperature rise from a doubling of CO2 is of the order of 0.24 deg C. Another interesting point is that the article points out that basic physics (gravitational potential … atmospheric pressure) should lead to an automatic global scale temperature control through changes in cloudiness. If I understand the situation correctly this should be an obvious support for the Lindzen’s “Iris” idea.
The article is fairly heavy reading but really interesting. The article can be found at:

Click to access vol111001_01.pdf

• HankHenry
Posted Jan 19, 2010 at 3:06 PM | Permalink

Oh wow. I’d never heard of Emiliana huxleyi, but I’m glad I have now. I’ve been looking for a sculpture for my front lawn and this may be it. Better than most pollen grains I’ve looked at. I also see that it inspired Lovelock’s Gaia idea. It seems the AGW crowd has put an implicit gag order on any further discussion of that lovely thesis. My view is that just as CO2 fertilizes plants on land, so must it fertilize the biochemistry of carbonates. There is a reason there is so little CO2 in the atmosphere and so much is built up into carbonate rock columns. It’s availability to living things is a veritable force of nature.

• HankHenry
Posted Jan 19, 2010 at 3:10 PM | Permalink

darn, I meant to post this after a post by Julian Flood about two posts post of here.

• Bernie
Posted Jan 19, 2010 at 6:33 PM | Permalink

Lars:
It would be useful to directly alert Lubos to this interesting paper.

• Posted Jan 20, 2010 at 12:16 PM | Permalink

Everyone here needs to read this paper. I found the math easy to follow and the implications, particularly for GCMs, profound.

Posted Jan 19, 2010 at 3:17 AM | Permalink

This is obviously a “science” that has to return to basics!No castle is built on shaky grounds.

• Posted Jan 19, 2010 at 12:49 PM | Permalink

“They said I was daft to build a castle in the swamp, but I built it all the same, just to show ’em. It sank into the swamp. So, I built a second one! That sank into the swamp. So I built a third one! That burned down, fell over, then sank into the swamp, but the fourth one stayed up!”

49. Julian Flood
Posted Jan 19, 2010 at 5:09 AM | Permalink

* an increase in surface reflectivity (impossible since tropical ocean surface reflectivity doesn’t change with surface temperature)

Do coccolithophores respond to temperature with increased growth? If so they would substantially increase ocean albedo. Do higher SSTs lead to increased surface wind? If so, increased emissivity during darkness would provide a feedback in high latitudes. There is an interesting blip in windspeed to go with the SST blip from ’39 to ’45 that Wigley (?) complains about as inexplicable in the CRU emails.

* increase in aerosols that are making clouds more reflective with relatively small impact on IR emission (no major volcanoes during this period, but Saharan dust and biomass burning could have the desired effect. However, no known relationship between surface temperature and dust/biomass aerosols)

Does DMS production increase with increased temperature? Does the ocean flora manipulate the sea surface to increase salt particle production when temperatures threaten to rise too high? Does dust and dissolved silica alter the relative success of diatoms (relatively poor at DMS production) and coccolithophores? Does excess nitrate runoff lead to reduced Emiliana huxleyi populations as they succumb to viral infection?

* increase in low cloud amount (that reflects sunlight while emitting IR at nearly the same temperature of the surface). This works only if the increase in low cloud amount is not obscured by high clouds (which dominate the radiation fluxes at the top of the atmosphere when they are present). For this to be a significant effect, we would need to see a decrease in deep convective clouds in the tropics, which to my knowledge hasn’t been observed.

I find this surprising: my experience of low level oceanic stratocumulus is that it frequently occurs without higher cloud masking it.

Even more surprising is the limited attention given by climate science to the biology of the ocean surface. If it all turns out to be Emiliana huxleyi, it’ll be enough to make a diatom laugh.

JF

• Judith Curry
Posted Jan 19, 2010 at 7:52 AM | Permalink

Julian, good questions

surface albedo: no idea about the cocco….. wind doesn’t influence the surface albedo in SW unless the sun angle is very low (however wind is a big deal in terms of microwave reflectivity which is important for remote sensing of the sea surface but not for energy balance)

re aerosol particles and ocean biology: good question, an open topic for research

re low cloud amount: abundant low clouds (without overlying high clouds) are found in the subtropics (not the tropics)

The issue of upper ocean biology being involved in feedbacks is interesting. I personally have very little knowledge about this. My understanding is that over the small temperature changes we are talking about, the issue is in terms of productivity is nutrients rather than temperature. I would appreciate hearing more about your ideas on this.

• Julian Flood
Posted Jan 21, 2010 at 7:32 AM | Permalink

Re surface albedo: I think emissivity is the crucial factor here. Ruffled water cools more quickly. If the water is smoothed for any reason, it will stay warmer during darkness. I’ve got a graph of it somewhere and I’ll try to find the reference.

Re: aerosols: no sign of VOCALS yet.

Re: stratocu: I didn’t know that.

I now reveal the system of the world… well, part one.
One of the more telling arguments for human involvement in climate change is the isotope record of C, but elevated C12 is essentially only a biological signal which says nothing about attribution. The assumption is sometimes — perhaps ‘often’ is not too strong here — made that the signal is from the past in the form of burnt fossil fuels. It is entirely possible that the isotopic signal is from the present.

There is a nice graph in ‘Evidence for pre-industrial variations in the marine surface water carbonate system from coralline sponges, Böhm ea.’ (it can be found on Ferdinand Engelbeen’s site) which shows changes in the isotope ratio since 1400. This has been pointed to as evidence for fossil fuel burning: it is nothing of the sort. The ratio begins to change in 1700 and the fall correlates poorly with fossil fuel consumption (eyeball only which needs to be mathematically verified). There are other possible explanations for that graph, but research would be needed to eliminate those which do not match reality.

The easy way for plants to fix carbon is the C3 process, a process which discriminates against the heavier carbon isotopes. C3 plants, land and sea, will preferentially pull C12 from the atmosphere. Any disturbance of this steady state will increase the C12/C13 ratio and will show up in the atmosphere as increased C12. In certain circumstances — low nutrient levels, lack of certain trace elements such as zinc and chromium, low CO2 levels — the C4 pathway is easier. Some phytos are obligate C3, some C4 and some can switch from one process to the other. C4 is much less discriminating between C isotopes, as is the Crassula Acid Mechanism (CAM).

Diatoms use a C4-like process. They are ferocious competitors for nutrients, so much so that only when diatoms exhaust their own limiting nutrient, silica, do the calcareous phytos bloom. If dissolved silica in the productive zones around the continents has been increased by agricultural run-off and dust then the calcareous phytos will be at a disadvantage, switching when starved to C4, and the isotope signal will change accordingly. The populations of various diatom species will vary. The spectacular coccolithophores will be fewer and the oceanic albedo will drop. Google for images of Emiliania huxleyi blooms off Newfoundland to see how this coccolithophore reflects light upwards as it sinks — benefitting whatever plant is growing above it. It would be interesting to check the literature for the feeding habits of recently hatched codlings: an ideal ultimate theory of climate change would include the collapse of the cod fishery off Newfoundland*, and if silica increase changes the ecosystem to starve minute codlings it would be a gratifying bonus. I can find no measurements of the light levels inside and outside of phyto blooms.

Pulses of dissolved silica and of inputs such as volcanic dust rich in zinc and chromium might show up in the record as changes in C12/C13. Aeolian dust, if it can be shown to dissolve, might be correlated with minor isotope variation. Global temperature changes silicate weathering rates and if that silicate goes into the deep oceans it will upwell in the southern ocean after some hundreds of years, changing the flora and with cause and effect separated by that lag. It might be possible to tie current silica levels to the MWP.

The ooze should show the associated changes in phyto/diatom populations. A separate check on radiolarian populations might be useful as giving some handle on the levels of silica in blue water ocean areas. Silica levels in deep water should be interesting.

It should be possible to test whether the flora of the ocean surface has been changing over the last 200 years. If C4, C4-like and CAM metabolism has become more common then we can speculate that there will be other implications, not merely changes in isotope fluxes. Obvious examples are DMS production, carbon export to the deep ocean, albedo and surface effects.

If biology is providing climate feedbacks, the isotope signal will give us clues about how the biology is changing and how this will alter its ability to drive those feedbacks.

I will now go away and think.

JF

*any truly global theory should be overly ambitious.

• willis Eschenbach
Posted Jan 21, 2010 at 11:09 PM | Permalink

Judith Curry, thanks as always for your comments and insights. You say:

Julian, good questions

surface albedo: no idea about the cocco….. wind doesn’t influence the surface albedo in SW unless the sun angle is very low (however wind is a big deal in terms of microwave reflectivity which is important for remote sensing of the sea surface but not for energy balance)

While wind does not change the surface albedo in SW, it definitely changes the surface area. This is because as soon as you get spray, you have a huge number of tiny droplets, each of which has a surface which is radiating …

re low cloud amount: abundant low clouds (without overlying high clouds) are found in the subtropics (not the tropics)

As someone who has lived in the tropics for years, I can assure you that you are wrong. The tropics generally have low clouds from about 11AM onwards. If the ocean is warm enough, some of them turn into thunderstorms and high clouds … but the rest are still there.

So if the (low) clouds form at say 10:00 rather than 11:00, this makes a very large and significant change in the energy balance. Despite Julian’s claim, this can easily occur with no change in the amount of high clouds, I see it all the time.

This is a very large unrecognized problem in climate science, which is that averages aren’t enough. For example, generally tropical clouds are around from about 11 AM to 11 PM. My figures show that if this shifts forward by one hour, it makes about a 4 W/m2 difference in the radiation balance … but the average amount of clouds hasn’t changed at all. The only thing changed is the timing.

The issue of upper ocean biology being involved in feedbacks is interesting. I personally have very little knowledge about this. My understanding is that over the small temperature changes we are talking about, the issue is in terms of productivity is nutrients rather than temperature. I would appreciate hearing more about your ideas on this.

One of the more fascinating finds in the last few years is that plankton emit gas with sulfur containing compounds when they get heat stressed. These sulfur compounds serve as cloud nuclei, and thus tend to cool the plankton down … how big is that effect? No one knows, but given the area of the ocean, it could be quite large. Haven’t noticed any models incorporating it, though … and it’s a negative feedback. Coincidence? You be the judge …

w.

• Posted Jan 22, 2010 at 2:25 AM | Permalink

Willis,

I made no claim about tropical stratocu — it was a repeat of Dr Curry’s comment above. VOCALS experiments should be complete by now and no doubt they are analysing the results of flying through the continent-sized low level cloud off Chile.

Yes, thanks for emphasising that emissivity is the point of wind effects on the ocean surface — a smoothed ocean will cool at a slower rate than a ruffled one during darkness.

In answer to Dr Curry’s request, I posted a rather long part one reply which seems to have vanished into the ether, Knowing our host’s dislike of speculation, I am reluctant to post again with part two ramblings added — one does so dislike one’s pearls being snipped. The dilemma is, of course, having to choose between rudeness to our host or his guest.

In a nutshell — manipulation of cloud salt particle size by changes to the ocean ‘skin’ with surfactants and lipids, manipulation of the number of CCNs by DMS, manipulation of the physics of cloud droplets by oil or surfactant. This gives the ocean flora the ability to increase or decrease cloud cover, controlling temperature, UV and light levels. Underwater control of light can be achieved by use of coccolithophore platelets. Long term changes in the populations by e.g. replacement of calcareous phytos by silicaceous diatoms (hence the low C isotope signal) will show up as a changed steady state position. Radiolarian shells (silica) have been atrophying for 3 million years and perhaps we have reversed that decline.

JF

• Posted Jan 22, 2010 at 12:24 PM | Permalink

Interesting, my long post came, went, recame and has now gone again. Is there a Connelly in the house?

JF

• Judith Curry
Posted Jan 24, 2010 at 7:47 AM | Permalink

Willis, a comment about the low clouds. In the tropics, these are shallow cumulus, typically with an areal coverage that is typically less than 20%. this dilutes the radiative impact of these clouds relative to say the extensive sheets of low level stratocumulus in the subtropics.
You make an interesting point about the diurnal cycle. The ISCCP dataset does include diurnal cycle information on clouds (with 3 hour resolution), and could be use to look at any impacts from the diurnal cycle of cloudiness.

50. Ivan
Posted Jan 19, 2010 at 8:34 AM | Permalink

Dear dr Curry,

you still did no explain: why warming in the polar region of the Northern Hemisphere is an AGW “fingerprint”, while warming in the Southern polar region is not?

• Patrik
Posted Jan 19, 2010 at 9:57 AM | Permalink

Yes, that is a central question.

• Lance
Posted Jan 19, 2010 at 10:25 AM | Permalink

Patrik,

I can’t speak for Dr. Curry, nor am I near as qualified as she, but I believe the standard response to why the Antarctic isn’t showing the “fingerprint” of AGW is that the Antarctic, unlike the Arctic, is surrounded by ocean and thus the warming “signal” isn’t as pronounced or is being “masked”.

I think this is a hand waving non-answer but it is the response I have heard from RC and other warmist sources.

• Posted Jan 19, 2010 at 1:20 PM | Permalink

The effects of CFCs are another post-hoc “classic” in explaining why Antarctica is not warming as fast as the Arctic. Still, the atmosphere over the Antarctic continent is the coldest and thus driest in the world. I would grreatly appreciate it if anyone would care to give a logical, easy-to-follow explanation of why we are not expected to see an enhanced CO2-induced warming in a place with such a small amount of water vapour.

And, once again, if anyone has a link to the last LC response mentioned by Lubos, it would come in very handy for this interesting discussion.

Thanks,
Mikel

• Judith Curry
Posted Jan 19, 2010 at 6:17 PM | Permalink

Ivan i will get to this one, its important, it requires more time to reply, i’m taking care of some of the quicker replies first

51. Ivan
Posted Jan 19, 2010 at 9:57 AM | Permalink

Curry: “Colin, the “science is settled” is political statement, not a scientific statement! Decision making under uncertainty is often done facing conditions of much greater uncertainty than say the IPCC report. Even the IPCC report gives probabilities for its conclusions (none reach 100%, although they do say that warming in the latter half of the 20th century is “unequivocal”)”

That warming was “unequivocal” since 1975-76 until about 2000 nobody would deny. However, nobody would deny that there was unequivocal cooling 1945-1975 either.

When AGW alarmists (IPCC included) say that “science is settled’ what they mean by that is not that nobody denies that there was a warming in recent decades, but that nobody denies that man was a cause of that warming, and that warming will get worse due to the man in the future, and that this will have catastrophic consequences. So, we don’t have here only a non-controversial factual statement as you suggest, but a very complex and entirely unsubstantiated chain of reasoning which is sold and advertised to the public as “rigorous argument” using the authority of governmental and intergovernmental bodies. The debate on feedbacks here just highlights the problems in one of the many unsubstantiated leaps in that “argument”.

In that regard, your response to Colin was a classic red herring.

• Judith Curry
Posted Jan 19, 2010 at 6:16 PM | Permalink

The IPCC statement says that “most of the warming in the latter half of the 20th century is very likely caused by humans”. “most” implies greater than 50%, and “very likely” means > 90% as defined by IPCC. Do I agree with the combination of 50% and 90% in this statement? No. At least some of the warming is caused by anthropogenic CO2 (no one argues with that), exactly how much requires sorting out all the causes of natural variability. In the 90’s, the warm temperatures got a boost from warm phases of both AMO and PDO (the same thing that i think caused the bump ca 1940).

• geo
Posted Jan 19, 2010 at 6:31 PM | Permalink

Yes, I think a lot of us sense that IPCC took “worst case” and passed it off as “middle of the road” and then tried to shout everybody else down.

Re AMO and PDO. . . would you agree that we’re likely going to have a relatively firm grip in say, about 10 years, as to what the real future temperature trend line is liable to look like?

• Judith Curry
Posted Jan 19, 2010 at 8:02 PM | Permalink

10 more years of data (plus model improvements) will certainly help us sort all this out.

52. blueice2hotsea
Posted Jan 19, 2010 at 1:29 PM | Permalink

May I?

Winter NH warming would be an expected fingerprint of CO2 induced warming because both water vapor and methane, which are abundant GHGs in the summer, plummet in the winter, leaving CO2 dominant.

So, a change in CO2 will have a much more noticeable effect on surface temperature than say in the summer ocean tropics where as JC has already noted, “virtually no IR emitted by the surface reaches space”.

Also, temp swings over land are more pronounced than over the oceans; the NH has more land than the SH.

• Judith Curry
Posted Jan 19, 2010 at 8:21 PM | Permalink

A quick reply to the Arctic vs Antarctic situation. A new report has been put out by the British Antarctic Survey that is relevant here
http://www.antarctica.ac.uk/press/press_releases/press_release.php?id=1065

• Lance
Posted Jan 19, 2010 at 9:47 PM | Permalink

Dr. Curry,

Doesn’t the idea of using one unproven climate theory (the CFC’s depleting the ozone hole) to explain the failing of another (the lack of AGW warming in the Antarctic) seem a bit precious?

Talk about layering uncertainties onto uncertainties!

Wouldn’t it be more consistent and honest just to admit that the lack of warming in the Antarctic is inconsistent with AGW and say that we don’t understand why?

Using the alleged build up of CFC’s(which are supposed to be anthropogenic greenhouse warming gases anyway) in the stratosphere as an ad hoc excuse for the lack of expected warming looks like a desperate attempt to explain a blatant empirical inconsistency with AGW theory and its attendant “polar amplification”.

The best scientists I know are not afraid to say “Gee, I don’t really know what happened there. Maybe we should rethink things a bit.”

• Judith Curry
Posted Jan 20, 2010 at 5:16 PM | Permalink

Lance, regardless of what is causing the antarctic ozone to be depleted, its depletion is a robust observation, and the depletion does impact the greenhouse effect in this region, particularly because there is low water vapor in this region

• schnoerkelman
Posted Jan 21, 2010 at 7:55 AM | Permalink

Re: Judith Curry (Jan 20 17:16),
I’m sorry if this is OT but since it was being discussed…

As I understand it ozone depletion takes place when temperatures are very low, that is Antarctic winter with virtually no insolation. Given the low temperatures I would think that IR emission would also be reduced and thus the role of GHGs reduced in turn.

Could you perhaps clarify this or provide a link?
Thanks, bob

• DeWitt Payne
Posted Jan 22, 2010 at 3:35 PM | Permalink

…the [ozone] depletion does impact the greenhouse effect in this region….

Indeed. Quick and dirty MODTRAN calculation using SubArctic Winter conditions. Setting stratospheric ozone scale to zero increases LW TOA emission by about 4 W/m2, or about twice the radiative effect at that latitude of doubling CO2, and in the opposite direction. The surface temperature offset required to restore balance is about -1 C. What’s not included in MODTRAN, however, is the change in stratospheric temperature, which in the absence of sunlight, would warm because less energy is being radiated away by the ozone.

In reply to another comment above:

Upper tropospheric warming in the tropics is not a fingerprint only of enhanced greenhouse warming. Any warming of the surface should, to a first order, cause enhanced warming of the upper troposphere compared to the surface. Over a short term, it does. A lack of observed enhanced upper troposphere warming over the long term in the tropics, however, may be a test of model validity, the surface temperature record or the upper tropospheric temperature record or some combination of all three.

• blueice2hotsea
Posted Jan 22, 2010 at 11:15 AM | Permalink

Dr. Curry, thanks for that link.

Antarctic cooling caused by CFC reduction(!), a good reason not to expect a CO2 fingerprint there.

But, now I am bothered by the image that we are like the heat-stressed plankton mentioned by W.Eisenbach, which blithely moderate a more favorable climate through gas emissions. Remember when global cooling and looming food shortages were big concerns? We coincidently reduced CFC’s, scrubbed power plant SO2 and expanded methane producing wetlands. And now, a warmer, greener planet.

53. Ivan
Posted Jan 19, 2010 at 2:50 PM | Permalink

bluice,

if you look at 1930s and 1940s warming “blip” (Wigley, Cilmate gate, 2009) largely thought to be natural (by IPCC), the warming in Arctic was even larger than today, while also mostly absent in the Antarctica. How then Arctic warming could be a specific fingerprint of GHG warming?

• blueice2hotsea
Posted Jan 19, 2010 at 3:37 PM | Permalink

Hi Ivan-

You would look for anomalous warming in NH winters, for example, during a period of steady or falling temperatures around the rest of the world.

Of course, it might be less inflammatory to refer to this as a prediction rather than a fingerprint. However, unless a competing theory also predicts exactly this sort of thing, then fingerprint is OK.

54. Posted Jan 19, 2010 at 3:23 PM | Permalink

Thanks Dr. Curry for participating so actively here. As a member of the peanut gallery it’s always tiring and discouraging to see how much segregation of thought there is between points of view here and at RealClimate. You are helping to fix that in a very academically enlightened way. Bravo!

• Judith Curry
Posted Jan 19, 2010 at 8:00 PM | Permalink

55. Steven Mosher
Posted Jan 19, 2010 at 3:55 PM | Permalink

bender bender bender.

Nutshelling.

1. Best approach to determining if the models have issues with modelling
low level clouds in tropics on short to intermediate time scale.

Seems to me that you would want to run a GCM and take the issues that
JC raises ( flux in and out of the tropics ) off the table. So that the
question ISNT was LC’s approach justified, but rather what’s the best
approach?

2. Can such an approach ( looking at this time scale) even tell us anything much
of interest? Which I suppose goes to the question about the linkage between
short time scale net negative feedback processes and longer time scale
net positive.

Did I nutshell this thing correctly? Oh pool party this saturday. sorry for the late notice.

• bender
Posted Jan 19, 2010 at 6:53 PM | Permalink

Yes, I agree. Of course, if I take him at his word, Lindzen would probably never take this approach … because he has stated in that lengthy interview of his that he doesn’t run models. It’s not his thing. Indeed, he seems remarkably unopinionated on the structure and function of GCMs. I think this is part of what JC is pointing at: you can’t do a proper job of this kind of paper unless you go the extra mile to compare to data apples to modeling apples. And if you’re not going to do that, should you really be making the kind of extratropical extrapolations that L&C dare to do? It does seem a little cavalier, to say the least.

56. Ivan
Posted Jan 19, 2010 at 4:14 PM | Permalink

“You would look for anomalous warming in NH winters, for example, during a period of steady or falling temperatures around the rest of the world.”

Are you denying global warming since 1977? Is your suggestion that we had just NH polar winters warmer since than while global temperature did not risen? I thought AGW predicts GLOBAL warming…

“Of course, it might be less inflammatory to refer to this as a prediction rather than a fingerprint. However, unless a competing theory also predicts exactly this sort of thing, then fingerprint is OK.”

But, the problem here is that the competing theory “predicts exactly this sort of thing”. And moreover, the same pattern and similar magnitude of warming already happened in 1920-1940 as a consequence of the NATURAL warming. How then the same pattern of warming could be the “fingerprint” or even just “specific prediction” of AGW? It is not a specific prediction. On the other hand, the anomalous tropical tropospheric amplification of warming is a specific prediction of AGW.

• blueice2hotsea
Posted Jan 19, 2010 at 5:30 PM | Permalink

Hi Ivan-

I agree with Judith Curry that perhaps the best place to look for the ‘fingerprint’ is in Northern latitude winters. Not saying we’ll find it.

Where would you look for it?

Given the sparsity of contiguous measurements, I hope not the Antarctic.

• Ivan
Posted Jan 20, 2010 at 9:04 AM | Permalink

Hmm, interesting I thought that Arctic weather network was also very sparse, wasn’t it?

The problem is that the Antarctica is actually an ideal place to look if you want to identify the polar amplification; an enormous dry continent covered with vast ice sheets and with minimal influence of the surrounding ocean.

Since polar amplification is predicted in the case of natural causes as well, the real fingerprint is tropical tropospheric warming. And, needless to say – it does not exist in the data.

57. Waterloo
Posted Jan 19, 2010 at 5:31 PM | Permalink

First, let me state that Dr. Curry is clearly a person of science. Science relies on addressing counter claims, devising experiments and tests, and producing verifiable results, and Dr. Curry clearly shares that philosophy.

Now, for my question As a control (feedback) Engineer, one of the must troubling things that I see is the use of the term “climate sensitivity” to mean the sensitivity of climate to a change in CO2, and, in fact, taking it a step further and using it as the linear response to a doubling of CO2. From my point of view, this is a horrible misleading use of the term sensitivity, which should be the change in the output to an infinitesimal change in the parameter (ie, the norm of the partial derivative of the model to the parameter in question). This, IMHO, missuse of the term sensitivity makes me extremely concerned if proper sensitivity analysis are being done on the other parameters used in climate models. So Dr. Curry, why was the term sensitivity used in this way, and are climate models subjected to proper sensitivity analysis?

• RB
Posted Jan 19, 2010 at 6:08 PM | Permalink

Actually, doesn’t the sensitivity as known to climate science go as 4.3 ln(C/C0) – definitely not linear.

• Henry
Posted Jan 19, 2010 at 7:47 PM | Permalink

There is no certainty over the 4.3 factor (which corresponds to a 3°C increase for a doubling of the CO2 concentration). It might instead correspond to 1°C, 2°C or 4°C; this uncertainty would have a dramatic impact on the possible cost of adaptation.

There does appear to be a consensus that the relationship is logarithmic (so the marginal sensistivity falls inversely as concentrations increase), even though concentrations of other greenhouse gases such as methane are thought to have different non-linear relationships with temperature. To me the lack of a single relationship function across different gases could cast doubt on the certainty of each of the asserted relationships.

• Judith Curry
Posted Jan 19, 2010 at 6:10 PM | Permalink

Waterloo, I agree totally! this is a misuse of the term sensitivity in terms of its meaning in the context of control theory. Back in the 1970’s, a few scientists applied linear control theory to a simple global energy balance model, and derived a sensitivity parameter that is essentially delta radiative flux at the top of atmosphere/delta surface temperature. Most of my essay is about why i don’t think this is really “sensitivity” in the context of a real feedback analysis. Your suggestions on a better way of doing things would be appreciated!

• RB
Posted Jan 21, 2010 at 4:39 PM | Permalink

This was the book chapter I was referring to in the above.

• bstewart
Posted Jan 21, 2010 at 4:44 PM | Permalink

BLouis79 might be referring to the fact that in any ensemble of GCM runs that ought to reach a steady state, some do and some run away to unrealistic conditions. (By runs that ought to reach steady state, I mean for example no CO2 increase, or increase CO2 until it doubles and then hold CO2 at the new level.) It is (or used to be) common practice to toss the realizations that run away, without any understanding of why those particular realizations misbehaved and others didn’t. Flux corrections aim to reduce these misbehaviors, but don’t entirely eliminate them. Do I have that right?

58. Gerald Browning
Posted Jan 19, 2010 at 5:58 PM | Permalink

Judith Curry,

Can you now give as careful a review of the problems with climate models?

No mathematician or numerical analyst would believe that a finite difference approximation of the atmospheric and oceanic equations with the real atmospheric and oceanic dissipation rates is accurate for more than a few
days. In this regard see the spread of error from an inaccurate approximation of the surface boundary layer to higher altitudes in a few days in Sylvie Gravel’s manuscript on this site. Note that this error growth is due to only one parameterization error. It is not hard to imagine the
problems due to other parameterization errors, especially in the tropics where the total (net) heating is dominant, but very poorly understood.

If a long time scale is crucial to understanding climate sensitivity, climate models have no mathematical or numerical justification for determining that sensitivity. They are basically a WPA project for
scientists that are dependent on them for publications without the use of rigorous mathematical analysis.

As one final note, can you please provide a rigorous mathematical explanation for convective adjustment, i.e., the artificial conversion of a small columnar feature (instability)into a large scale hydrostatic feature?
In particular, what impact does that adjustment have on the numerical
accuracy of the finite difference approximation of the continuous dynanical equations?

Jerry

• Judith Curry
Posted Jan 19, 2010 at 6:06 PM | Permalink

Jerry, this is a big topic for another time.

• Gerald Browning
Posted Jan 19, 2010 at 8:52 PM | Permalink

Judith,

Isn’t it interesting that when I ask these very specific mathematical and numerical analytic questions, Real Climate will not post them and all meteorologists avoid the issues just as you are doing. Is there something to hide or are you just protecting the club? I think both. You are not that naive.

Jerry

• Skip Smith
Posted Jan 19, 2010 at 9:15 PM | Permalink

… or maybe it’s impossible to discuss complex issues like this though blog comments.

• Judith Curry
Posted Jan 19, 2010 at 9:18 PM | Permalink

Two points:

I have a limited amount of time to spend on blogging, and i prefer to post on topics that have the broadest interest to the CA community and best intersect with my personal expertise

Some months ago, i did review a paper that you requested, the CA regulars seemed to agree that I won the most quatloos, but you were dissatisfied with my arguments.

• bender
Posted Jan 19, 2010 at 9:24 PM | Permalink

• Steven Mosher
Posted Jan 19, 2010 at 11:14 PM | Permalink

The cage match between judith ( Dr. J), Lucia and Dr. B ( jerry).

• bender
Posted Jan 19, 2010 at 11:54 PM | Permalink

• Steven Mosher
Posted Jan 20, 2010 at 2:57 PM | Permalink

dude you were the referee

• Lance
Posted Jan 19, 2010 at 9:51 PM | Permalink

Even if I am the ten thousandth person to say it, I really do respect your openness and willingness to come here and answer the questions which are not always gently asked of you.

Kudos!

• Gerald Browning
Posted Jan 20, 2010 at 12:03 AM | Permalink

Judith,

It appears you have plenty of time to denigrate the Lindzen Choi manuscript that seeks an alternative to global doom and to answer any praise from ignorant
bloggers.

Why not answer questions posed by a mathematician and numerical analyst that knows the flaws in climate models
from personal experience, mathematical analysis, and numerical analytic theory.

BTW you did not win the argument on the other thread. You refused to read the manuscript that indicates the ill posedness of the hydrostatic system (now demonstrated with numerical convergence tests on this site) and to answer any rigorous scientific questions I posed. As usual your arguments contained plenty of rhetoric and no scientific content. I await your answer to the questions I posed above. Either you are too ignorant to answer them or too afraid to offend your club.

For starters, please explain the mathematical basis for convective adjustment. That is a straight forward single
mathematical question. Try to answer it without a smokescreen and continual excuses.

Jerry

Jerry

• bender
Posted Jan 20, 2010 at 9:44 AM | Permalink

May I add that explaining this convective adjustment would probably appeal to many, not just the marginal, at CA?
.
And may I add that I find it annoying how alarmists (and even lukewarmers) seem to have many excuses ready at hand when it comes time to talking math with Gerry Browning or chaos with Tom Vonk or the statistics of long-term persistence with Dimitris Koutsoyiannis. If these guys are so wrong, why don’t you just refute them and be done with it? Gav?

• Posted Jan 20, 2010 at 10:46 AM | Permalink

Having battled through some of Vonk in the past (thanks for the suggestion of soundtrack!) I was wondering why you didn’t mention Koutsoyiannis at the same point above. Are they broadly making the same point? If the answer is “Read Climate Audit” then I know the answer and don’t worry. If there’s a simple way to give some clues, even better.

• Steven Mosher
Posted Jan 20, 2010 at 2:00 PM | Permalink

Dr. K had a paper reviewed here a while back.
you will also find a piece on Mandelbrot.

bender is right, WRT the lukewarmers. Let me see if I can express it. If Mandelbrot is correct and Jerry and Tom V are correct, then it would appear that we have no hope of predicting the magnitude of the response of the climate to added GHGs..I like hope.

maybe Vonk or Jerry will qualify that characterization of their position. Further, I wonder whether anything they say has any bearing on the fundamental proposition that an atmosphere with more GHGs will have a higher temperature than one without. The threshhold question
is that one. I liken it to a problem in forebody vortices in aircraft design. You could not predict the vortex at a high AOA, but you damn well knew it could form and you knew it would form on one side or the other, couldnt predict which, and you could know that the side force exerted would lead to a nose slice and bad news for the pilot.

So ya, there is a lot of arm waving when Tom and Jerry
show up because no one wants to tackle it directly.

• Posted Jan 20, 2010 at 2:23 PM | Permalink

Tom and Jerry meet Dr K. Isn’t that the latest from Pixar? You did ask for pictures …

I saw the Koutsoyiannis stuff a while back and liked the fact it came out of hydrology and Hirst, among other things (with help from Kolgomorov, who was evidently a rare kind of genius). Real world and all that.

I’d picked up that long-term persistence makes it very possible to think you’ve seen a trend when you haven’t. But how does LTP relate to chaos theory? I think that was my question.

Don’t care what the lukewarmers make of it for now, the analogy with aircraft design may be helpful but the climate system is one big ball of fun compared even to that. The best model of the universe is the universe and all that. At what subset of the universe does that begin not to be true?

But thanks – and thanks for those who have come on this thread to help enlighten us (and to Judy). I confess coccolithophores were new to me too. What fun.

• Steven Mosher
Posted Jan 21, 2010 at 2:01 PM | Permalink

LTP and chaos? ( BTW I like the hurst stuff as well )
hmm. Well conceptually I would imagine that you’ve got some sort of orbit with a long time scale periodicity. I think Darth Vonk has discussed a system with 5 attractors..I’ll shut up

vonk vonk vonk

( see if he appears)

Steve mc: No more chaos discussion on this thread. It’s got nothing to do with Lindzen and Choi .

• RB
Posted Jan 20, 2010 at 2:44 PM | Permalink

Modelers seem to call decadal prediction midway between initial value problems and boundary condition problems.

Projections of how anticipated changes in greenhouse
gases and aerosols will influence climate over
time scales of several decades to centuries (dec–cen)
can be considered primarily as “boundary condition
problems” (Fig. 2). Such model-based projections seek
to describe climate trends, not the details of individual days, seasons, or years. In contrast, daily weather
forecasts and shorter-term SI climate predictions [e.g.,
El Niño–Southern Oscillation (ENSO) forecasts] can
be thought of as “initial value problems,” for which
detailed knowledge of the observed current conditions
are crucially needed to define the starting point
(the initial conditions). Lorenz (1963) demonstrated
how, even if one possessed a hypothetically perfect
numerical model representing all of the physical
processes completely and without error, unavoidable
uncertainties in the initial conditions will invariably
grow and contaminate the numerical simulation of
transient weather systems. This sensitivity to initial
conditions (sometimes referred to as the “butterfly
effect”) limits to about 2 weeks the time period
over which even a perfect model could yield skillful
weather forecasts. When considering El Niño, a quasioscillatory
phenomenon that evolves more slowly than
synoptic weather systems, skillful numerical forecasts
of monthly mean or seasonal mean conditions (Shukla
(Kirtman et al. 2002).

• RB
Posted Jan 20, 2010 at 10:48 PM | Permalink

I don’t confess to have understood this but the gist I got out of it is that models used in IPCC AR4 are reliable for estimating sensitivity.

• bender
Posted Jan 20, 2010 at 10:06 PM | Permalink

But you didn’t wonder why I didn’t mention Tsonis? Or Wunsch? Well, I mention them now.

• Posted Jan 20, 2010 at 10:47 PM | Permalink

Bender, to be clear, in my question I was referring to this response of yours to mpaul:

The standard counter argument is that at a certain time-scale the “stochastics” (i.e. determinstic nonperidic flow) are irrelevant compared to the deterministic forcings. This is a logical argument. That doesn’t mean it’s true, however. Read everything Tom Vonk has written at CA.

I thought at once of Dr K as I read that, as I thought he was also challenging the ‘standard counter argument’ about stochastics and I was wondering about the relationship to Vonk. But I said not to bother if the best answer was to “read the blog”. I don’t want to waste your time. Although I’ve heard of Tsonis and have some idea of Wunsch they didn’t come to mind here. Seriously, don’t worry if there’s nothing it seems worth adding. I’m happy to plug away in my own time!

• bender
Posted Jan 20, 2010 at 11:25 PM | Permalink

The link between all of them is what we like to suppose are “stochastics”. By supposing we know something about these “stochastics” it allows us to be cateogorical in our statisics and unambiguous in our notions of that which is weather and that which is climate. But then the realities of atmospheric and ocean convection strike, and these stochastics – do the old assumptions hold? No? Then what is the consequence? You tell me: is earth’s climate ergodic? This is your link.

• Posted Jan 21, 2010 at 6:48 AM | Permalink

I take it that earth’s climate is non-ergodic. I have some new areas to explore. Thanks.

• H Hak
Posted Jan 20, 2010 at 10:16 AM | Permalink

Jerry, have you tried posting on klimazwiebel.blogspot.com? Von Storch and Zorida might take on your challenge.

• bender
Posted Jan 20, 2010 at 10:26 AM | Permalink

And they are welcome to comment here! So why not just alert them to Gerry’s question and direct them here?

• Posted Jan 20, 2010 at 10:48 AM | Permalink

Seems the right place to me too. Couldn’t have a friendlier audience. Well, we’re working on it.

• H Hak
Posted Jan 22, 2010 at 2:50 PM | Permalink

Gerald, I put your post here on Klimazwiebel under “Schiermeier in nature” with a request for von Storch and Zorita to comment.
Will see what happens.

59. philh
Posted Jan 19, 2010 at 9:30 PM | Permalink

As an old lawyer I have always found that some of the best work by a law firm is done around the lunch table discussing particular legal issues. This thread, thanks to Dr. Curry and the other participants, reminds me of those sessions. What a pleasure to see this.

• Jimmy Haigh
Posted Jan 21, 2010 at 4:58 AM | Permalink

It would be an interesting excercise to see what the response would be were Dr Curry to post this on Real Climate?…

• Judith Curry
Posted Jan 24, 2010 at 7:39 AM | Permalink

Philh, climate researchers need to make the case in a way that makes sense to lawyers. At the the annual meeting of the American Meteorological Society last week, Jane Lubchenco (undersecretary of NOAA) made a speech that focused on the need for better communication of climate research, but she seemed to think that a dumbing down of the science was needed so as to be comprehensible by the masses. I agree that better communication is needed, but not the dumbing down part. Climate researchers need arguments that will convince technically educated people outside the climate research field, and that make sense to lawyers. I think the technical blogs are a great venue for this kind of communication.

60. H Hak
Posted Jan 19, 2010 at 9:33 PM | Permalink

A bit OT, but interesting quote
“Looking back at the worst times, it always seems that they were times in which there were people who believed with absolute faith and absolute dogmatism in something. And they were so serious in this matter that they insisted that the rest of the world agree with them. And then they would do things that were directly inconsistent with their own beliefs in order to maintain that what they said was true.” —The Meaning Of It All: Thoughts of a Citizen-Scientist, 1998
Richard Feynman

61. H Hak
Posted Jan 19, 2010 at 9:42 PM | Permalink

Dr Curry, this quote was not meant for you, but might apply to some scientists today. Your contribution to CA here is very much appreciated

62. H. Tuuri
Posted Jan 19, 2010 at 9:42 PM | Permalink

I analyzed the ERBE tropics data with a free statistics program called OpenStat and posted the results as a comment on realclimate.org. I used a different smoothing approach than LC09, but I was able to repeat the main result of Lindzen and Choi: the climate sensitivity of the tropics is lower than for a black body.

OpenStat also tells us that there is essentially no correlation of tropical sea surface temperature anomaly to the heat transfer from the tropics to higher latitudes. That suggests that we cannot use El Nino to measure the climate sensitivity of higher latitudes. But, since the tropics comprise roughly 1/3 of the surface area of Earth, their impact to the global climate sensitivity is significant.

As to the impact of a doubling of CO2 in high Northern latitudes, I think the temperature rise here has stopped since 2001. Maybe the fast rise in 1980-2000 was just a fluke, and the climate sensitivity is not that high, after all?

63. Chris D
Posted Jan 19, 2010 at 11:02 PM | Permalink

This comment is for Team members who certainly read this blog and these words I type, but for their own reasons choose not to engage as Dr. Curry has done today.

I suspect you’re eager to repair the political damage done to your cause in recent weeks. In that light, you might find one layman’s reaction to Dr. Curry’s post quite useful.

I am a layman who in following CA for 2-3 years, has grown to distrust you and your science. I’m not saying your science is wrong. I’m saying I’ve grown to strongly distrust you.

In one single post, in one single day, with simple measured direct thoughtful tones, from an evident posture of mutual respect, Dr. Curry has immeasurably opened my mind to potentially trusting your science again. One knowledgeable person, one post, mutual respect->open mind.

Yes, this is an emotional reaction. Yes, the science is broad and has multiple lines of evidence blah blah. No I don’t understand the science. But yours is a political problem, so emotions are in play, and your science isn’t as relevant. This Average Joe’s reaction to this post should intrigue you as a way out of your political mess, especially since you’re so confident in your settled science. I’m telling you, Dr. Curry is giving you a live training session on the fastest, most effective way to overcome the political problem you’ve gotten yourself into. That should interest you.

Engage here. You’re smart guys. You know it would work.

Dr. Curry- thank you for inspiring me. I hope you enjoy respect among your colleagues for what you’re doing here, and I hope it encourages others to do the same.

• Steven Mosher
Posted Jan 19, 2010 at 11:18 PM | Permalink

it’s funny when I argue that sharing code and data and opening will OPEN MINDS, some people tell me I’m crazy and that everyone who opposes the science is some sort of irrational denialist who wont listen to reason.

• Posted Jan 20, 2010 at 1:05 AM | Permalink

Yes, wanting the full evidence = “denying the science”! But the movement to open all climate source code and data is surely unstoppable now. All credit to Steve, Ross, you and others that have banged that drum for so long

• Judith Curry
Posted Jan 20, 2010 at 7:23 AM | Permalink

Chris, thank YOU for your comments! I also wish more climate researchers participated in the blogosphere. This thread has been a credit to CA in a several of ways: the thoughtful and often insightful replies by the posters (and lack of snarkiness, for the most part) and Steve’s willingness to host this thread. I have not seen other blogs mentioning this discussion, but hope that other climate researchers will see that the process isn’t too painful and there can be some substantial appreciation and possibly changing of minds in the process.

• kim
Posted Jan 20, 2010 at 11:30 AM | Permalink

Hear, hear.
==========

64. DeWitt Payne
Posted Jan 19, 2010 at 11:34 PM | Permalink

The forcing from 2X CO2 is higher in the tropics than at high latitudes as this figure from Hansen, et.al., 2005 shows. The reason is that the altitude of the tropopause is higher in the tropics because the lower atmosphere is warmer and less dense so the temperature difference between the surface and the altitude where the optical depth of the CO2 band becomes less than one is larger. That significantly increases the total forcing from CO2 in the tropics compared to high latitudes. Since the total forcing from CO2 is lower at high latitudes, the magnitude of the change in forcing from doubling is smaller too. You can see the difference for yourself by playing with the numbers on Archer’s MODTRAN page.

• Gerald Browning
Posted Jan 20, 2010 at 1:15 AM | Permalink

Dewitt,

If I understand you correctly, your arguments support the neglect of the extratropics in the Lindzen Choi manuscript?

I find it amusing that a 6 page manuscript with some simple mathematical equations and observational data has caused such an uproar among the AGW proponents. A bit embarassing for them, especially if Lindzen addresses the
criticisms in his followon manuscript.

Jerry

• DeWitt Payne
Posted Jan 20, 2010 at 11:10 AM | Permalink

Jerry,

Several comments in this thread repeated the argument that warming from CO2 is greater at high latitudes because there’s less water vapor. That’s wrong. It also has nothing to do with whether restricting the analysis to the tropics is justifiable. While forcing is highest in the tropics, sensitivity, in terms of surface temperature, is very likely lowest. Much of the excess heat will be transferred to the higher latitudes and, assuming that relative humidity remains approximately constant in the planetary boundary layer in the tropics, most of the retained heat should be latent not sensible.

The global map in my link above also shows why expansion of equatorial and mid-latitude desert area may be self limiting in a warmer world. If you look at Africa, there’s what amounts to a hole in the forcing over the Sahara. Very low humidity means more efficient radiation of heat from the surface to space.

• DeWitt Payne
Posted Jan 20, 2010 at 11:22 AM | Permalink

Oops. I should have looked at the chart again before I posted. I remembered the hole in Africa, but the location was incorrect. It seems to be related to the ITCZ and regions of high rates of upwelling.

• Gerald Browning
Posted Jan 20, 2010 at 1:33 PM | Permalink

Dewitt,

The latent heat forcing in the tropics (and hence SST that generates the water vapor in the atmosphere) is a dominant
component of the forcing in the tropics and the convective heat towers rise to much higher altitudes than in the midlatitudes. The dominance of the net heating of this type is restricted to the equatorial region and smaller scales in the modlatitudes.

Jerry

• Judith Curry
Posted Jan 20, 2010 at 7:05 AM | Permalink

DeWitt, forcing is not sensitivity (nor is it feedback). the sensitivity of interest is the change of global surface temperature to a change in external forcing (e.g. increasing CO2). What is going on in the upper tropical troposphere is obviously important for the top of atmosphere radiative forcing, but it is not particularly important for the global sensitivity.

• Gord Richens
Posted Jan 20, 2010 at 7:52 AM | Permalink

Hmm. So the two need not be consistent?

• bender
Posted Jan 20, 2010 at 10:05 AM | Permalink

What is going on in the upper tropical troposphere is obviously important for the top of atmosphere radiative forcing, but it is not particularly important for the global sensitivity.

First, this is an assertion. Do you plan on defending it?
.
Second, I think DWP’s point could be rephrased as follows. If one were going to bias a study such as L&C’s by focusing on just one part of the world, you’d choose exactly the part of the world that they chose: tropical. Other parts of the world (extratropical, polar) would lead to an even larger extrapolation problem of the sort JC has in mind. Sure, an unbiased analysis would be better. Sure, it would be nice to replicate it in different parts of the globe. Sure, it would be preferable if L&C didn’t go out on a limb with speculative conclusions.
.
Third, tropical is not entirely divorced from extra-tropical. So I would be careful how hard I pushed that argument. That ENSO and PDO are “orthogonal” is an algebraic trick and does not imply functional (dynamic) independence. I’ll see if I can find that map that showed up a few months ago which suggested a ~2 year delay between ENSO & PDO warm anomalies. There was not a lot of commentary on it at the time. Just David Smith and a couple of others. But it made the point that tropical heat can be lost to the atmosphere directly (within the tropics) or moved poleward with some time delay and THEN lost to the atmosphere. Again, this does not mean that L&C’s extyrapolation is justified. It just mitigates some of the damage inflicted by JC’s extrapolation argument.

• See - owe to Rich
Posted Jan 20, 2010 at 3:28 PM | Permalink

Hi Bender,

In your #3 you forgot my name. Still, I won’t take it personally, and help you by pointing to
https://climateaudit.org/2009/05/12/unthreaded-n2/#comment-183534 which I had bookmarked, following which there were comments by you and me on estimating the time lag. Unfortunately, the graph itself isn’t actually there, but there is a pointer to Bob Tisdale’s.

HTH,
Rich.

• bender
Posted Jan 20, 2010 at 10:13 PM | Permalink

Thank you. That was it. You were in my “et al”, but I couldn’t put my finger on exactly who it was. It’s your handle. I don’t think of you as “Rich”. You’re so much more!

• Judith Curry
Posted Jan 20, 2010 at 5:12 PM | Permalink

The importance of the tropical upper troposphere for TOA radiative forcing is indicated in the Hansen 2005 graph provided earlier.

If sensitivity is defined as delta surface temperature/delta CO2 concentration, then the plots of observed and modeled surface temperature change reflect the local sensitivity

Regional sensitivity is a dodgy thing to try to determine. See figure 12.5 from my book which can be found at http://curry.eas.gatech.edu/Courses/6140/lectures.html (look at lecture 12) which describes the average zonal energy transport

• ianl8888
Posted Jan 21, 2010 at 1:48 PM | Permalink

Steve – snip. Too general an issue. Insufficiently connected to post specifics.

• D. Patterson
Posted Jan 22, 2010 at 5:19 AM | Permalink

What is going on in the upper tropical troposphere is obviously important for the top of atmosphere radiative forcing, but it is not particularly important for the global sensitivity.
[….]
The importance of the tropical upper troposphere for TOA radiative forcing is indicated in the Hansen 2005 graph provided earlier.

Absent from those sources is a recognition of the asymmetric forcings associated with the tropical Hadley circulation and their effects upon extratropical sensitivities as described in the modeling experiments by Yoshimori, Broccoli, and others preceding them. Although the authors acknowledge Chiang and Bitz [2005] were unable to find a confirmation of the modeled results in the observational record of the interhemispheric records, this problem once more highlights the potential perils of relying upon parameterizations and GCM results which are not adequately confirmed by the observational records.

Have you considered the results of Yoshimori and Broccoli [2009] when asserting TOA is not particularly important for the global sensitivity?

Yoshimori, M., and A. J. Broccoli (2009), On the link between Hadley circulation changes and radiative feedback processes, Geophys. Res. Lett., 36, L20703, doi:10.1029/2009GL040488.

• Ryan O
Posted Jan 21, 2010 at 3:06 PM | Permalink

Bender,

I bet you’ve already seen this, but for other interested readers, this would provide background for your #3. As you pointed out, many of the observed “teleconnections” and “orthogonal” modes interpreted from temperature or pressure time series are likely to be statistical artifacts. That means the observed independence of ENSO/PDO are also artifacts, and ENSO/PDO (and NAM/SAM) do not necessarily have physical meaning. I agree wholeheartedly that this weakens the extrapolation argument. However, L&C still have the burden of proof to show that the extrapolation is valid, as they are the ones making the claim that their analysis can be used to determine global sensitivity.

Anyway, as far as the artifacting goes, one of the best references I’ve found is here:

Click to access ijc-compagnuccirichman.pdf

A must read for anyone who believes these indices are physical phenomenon. Most will fail the test in the above (and fail badly).

65. Leo G
Posted Jan 20, 2010 at 1:31 AM | Permalink

Where does one look to find the global average temp if our atmosphere contained no CO2?

I am trying to help my son with a geography project.

(You don’t want the old man to look stoopid do you?)

🙂

• oneuniverse
Posted Jan 21, 2010 at 6:12 AM | Permalink

Hi Leo,

The figure I’ve heard most often is that the greenhouse effect as it stands now contributes roughly 30 degrees of warming to the climate. I don’t think that figure is derived from first principles, though.

There’s an article at Roy Spencer’s site, discussing how things might be on Earth if there were no greenhouse gases present. I’m not sure if it answers your question about temperature, though – haven’t read it entirely.
http://www.drroyspencer.com/2009/12/what-if-there-was-no-greenhouse-effect/

66. ianl8888
Posted Jan 20, 2010 at 5:05 AM | Permalink

Judith Curry

I’ve been looking forward to this for some time – your courage in participating in open debate is to be admired, because the closed Peer Review system is simply too difficult for non-climate scientists to access in any timely or sequential fashion

I’m only a geologist, so although core science is OK for me, understanding the nuances in this thread will take me some time … I do not at all begrudge this time, however

But there is one irritation that is unhappily far too repetitively familiar. Apparently Lindzen & Choi have published (?) an extension, addition, rebuttal or whatever. The irritation (and way too much of a constant) is that we do not have this to inform us further. So we simply go on shadow boxing. If you have access, could you please post a link ?

• Judith Curry
Posted Jan 20, 2010 at 6:58 AM | Permalink

Ian, I haven’t seen his new paper yet either, unfortunately, apparently it is under review and it is not posted at Lindzen’s web site. I agree that peer review process is too slow, but i also understand why scientists don’t publicly post their papers until they are accepted for publication. Assuming that LC again submitted their paper to GRL, GRL is pretty quick at getting papers reviewed and published.

• ianl8888
Posted Jan 20, 2010 at 7:10 AM | Permalink

Shadow boxing … at least the delay may give me time to absorb the different points of view up to this time

67. Posted Jan 20, 2010 at 7:38 AM | Permalink

DeWitt Payne: “Since the total forcing from CO2 is lower at high latitudes, the magnitude of the change in forcing from doubling is smaller too”.

The total forcing from CO2 is lower at high altitudes but we expect the CO2 forcing to warm them more than rest of the planet (at the surface)? Now I´m at a loss.

• bender
Posted Jan 20, 2010 at 9:47 AM | Permalink

Maybe you just confused “altitudes” with “latitudes”? Just a thought …

• Posted Jan 20, 2010 at 12:01 PM | Permalink

Thanks Bender, I did mistype latitude. The question remains though. CO2 forcing is lower at higher latitudes but that´s where we expect to see the highest warming?? (at least at the surface level)

68. TG
Posted Jan 20, 2010 at 9:24 AM | Permalink

JC said: […i also understand why scientists don’t publicly post their papers until they are accepted for publication…]

It is interesting, however, that physics, math and some other communities have no problem with posting preprints on http://arXiv.org; I believe the case is even stronger for climate science community to do likewise.

• Posted Jan 20, 2010 at 9:44 AM | Permalink

Another aspect to the openness appropriate to the very unusual situation of climate science. IT acting as blessing as well as curse (as it is in over-hyped GCMs, for example) – very fitting. I totally agree.

• Judith Curry
Posted Jan 23, 2010 at 7:55 AM | Permalink

TG I totally agree. But when a topic is highly politicized, scientists are worried about a massive letter writing campaign that might result in a withdrawal from publication of an in press article. Think M&M, Soon and Baliunas, etc.

• Posted Jan 23, 2010 at 9:17 AM | Permalink

You are talking in the two examples given of a letter writing campaign from those trying to protect AGW ‘orthodoxy’ by getting certain papers withdrawn from publication. But if preprints had been available on http://arXiv.org – if this was an accepted part of the culture in climate science, as it is in maths and physics – surely this would have served to deter such a letter writing campaign? Is there a downside therefore? Isn’t this something to campaign for as part of an Open Climate Initiative (as I will call it until my elders and betters tell me different) – putting right at least some of what went wrong as revealed in the CRUtape letters?

• Judith Curry
Posted Jan 24, 2010 at 7:31 AM | Permalink

Richard, I personally agree with you. If everything was open, progress would be more rapid and letter writing campaigns would be public for all to assess. How to change the culture is not simple. If any of you are AGU members, consider voting for Peter Webster for the AGU section president on Atmospheric Science, who is campaigning for more transparency and openness in climate research
http://www.agu.org/elections/ballot/atmos.shtml#pwebster

• Posted Jan 24, 2010 at 11:01 AM | Permalink

Thanks very much for that link to Peter Webster. I don’t have a vote myself so the reply is mostly to highlight this for AGU members that might miss it. Go for it, my geoscience friends. A group of you put me, a little IT guy, onto a more intelligent path on AGW in Bristol around 1992. Now, please sort it.

69. David Longinotti
Posted Jan 20, 2010 at 9:30 AM | Permalink

I admire the expertise displayed here on the question of the sensitivity of the climate to CO2 but, if the goal is to understand the reason for the apparent rise in late 20th century global temperature, the effort seems to be misspent.

A good scientific theory accounts for past observations and has predictive value. The hypothesis that the climate is quite sensitive to CO2 seems to be a failed theory on both counts. It fails to explain the MWP, and it’s previous predictions of significant warming due to increases in atmospheric CO2 have proved false over the last 15 years.

If there is doubt about the MWP evidence (as summarized in papers like that of Mr. Loehle), or if there are hypotheses as to why CO2 effects have been ‘masked’ for the past 15 years, then I think that’s where the debate needs to begin. Without agreement on the main data that needs to be explained by the model, how can different models be reasonably evaluated?

70. H. Tuuri
Posted Jan 20, 2010 at 9:51 AM | Permalink

Judith Curry,

“3. Given that LC focus their analysis on the tropical oceans, the results from their analysis of ERBE data seems very implausible: a strong negative feedback in the shortwave (SWR), with a small positive feedback in the infrared (IR). The negative SWR feedback is basically an increase in the planetary albedo with increasing temperature, without a correspondingly large decrease in outgoing IR. How could this possibly be?”

In my calculation (see my comment above), most of the change, maybe 75 %, in the outgoing radiation comes from the longwave radiation. Shortwave radiation, that is, reflected sunlight, contributes at most 25 %.

The concept of a ‘negative/positive feedback’ cannot be applied separately to the changes in the SW and LW flux. What matters is the sum of these changes. If the flux change per Kelvin is more than for an idealized black body, we say that there is a ‘negative feedback’. Otherwise, there is a ‘positive feedback’. A negative feedback means a ‘low climate sensitivity’ while a positive feedback means a ‘high climate sensitivity’.

I have used the OpenStat program and its ‘equally weighted theta autocorrelation moving average’ over 12 months to smooth the flux and SST graphs. Since the precise number that we get for delta-NET flux / delta-SST depends on the smoothing algorithm, I need to try different smoothing algorithms to see if the LC09 result can be repeated with most of them.

• Judith Curry
Posted Jan 23, 2010 at 7:53 AM | Permalink

By looking at the SW and IR feedbacks separately, you can identify the fingerpring of what is going on. For example, if you changed the surface albedo, you would see this effect in the SW and not the IR.

71. Gerald Browning
Posted Jan 20, 2010 at 3:16 PM | Permalink

snip

Steve: Jerry, you’ve already had this argument. It’s my editorial decision that it’s OT for this thread.

72. Jim
Posted Jan 20, 2010 at 8:02 PM | Permalink

Judith,

you said “an increase in surface reflectivity (impossible since tropical ocean surface reflectivity doesn’t change with surface temperature)”. Do you know this for a real-world fact? Surface reflectivity can be affected by bubbles or coatings, such as oil or protein. The sea contains life, and, temperature may alter the contents of the ocean resulting in a change in reflectivity. So, is that an observation or an inference?

• Judith Curry
Posted Jan 21, 2010 at 6:56 AM | Permalink

Could it change? Conceivably. Has it changed over the period that LC have looked at the data? No evidence to support this.

• Gerald Browning
Posted Jan 21, 2010 at 4:02 PM | Permalink

Judith,

Please cite reference that there has been no change instead of guessing.

Jerry

• Judith Curry
Posted Jan 23, 2010 at 7:51 AM | Permalink

Jerry, go to google scholar, google ocean surface reflectance temperature and see what pops up. Alot of stuff about remote sensing of the ocean surface in the microwave. My statement “There is no evidence to support . . .” means exactly that. No one is writing papers on this, there is no obvious physical mechanism for this. If you would like to do the analysis yourself, more than 25 years of global ocean surface albedo data (from satellite) is available from ISCCP
http://isccp.giss.nasa.gov/products/products.html

• BLouis79
Posted Jan 21, 2010 at 2:39 PM | Permalink

snip – OT for Lindzen and Choi

73. BLouis79
Posted Jan 21, 2010 at 2:36 AM | Permalink

OK, so Dr Curry has no confidence in L&C.

In my reading to date as a science graduate, I have no confidence in:
* land based temperature records
* computer model output claiming lengthy forward predictions on short-term data
* long-term weather predictions beyond 3 days (as studied by researchers)
* Kiehl/Trenberth’s 0.9W/m2 net positive radiative forcing based largely on Hansen’s computer model output (given net out of balance error based on TOA data satellite is near an order of magnitude larger)
* claims of long equilibrium periods of years when land can heat and cool within 24hours (climate researchers need to look at equilibrium within dirurnal and seasonal cycles rather than equilibrium of the observed totality)
* IPCC reports when IPCC dissenters have expressed their views on the science
http://environmentalresearchweb.org/cws/article/opinion/35820

I have not yet seen:
* a robust rejection of the null hypothesis “that earth’s climate is controlled by negatitve feedback processes and is fundamentally stable”
* a robust rejection of Gerlich and Tscheuschner’s physics falsification of radiative greenhouse effects
* a robust relationship between radiative forcing and temperature based in energy physics (mass, specific heat, heat diffusion) and mathematical proof
* a robust rejection of Nordell and Gervet’s analysis of heat from thermal pollution accounting for a large part of observed warming
http://www.ltu.se/shb/2.1492/1.5035?l=en
* a clear statement of the various GCMs assumptions and input data so we can see what is included/excluded

It is a well known and fundamental human cognitive bias to assume we have control over things we don’t have any control of.

74. Frank
Posted Jan 21, 2010 at 5:03 PM | Permalink

Judith and Steve: If we are ever fortunate enough to have commentary on another controversial paper from Judith, could you consider breaking the commentary up into multiple posts on narrower topics? I have returned to this post, the RC post, the LC paper and some of the references several times since it appeared digesting the arguments and hoping to formulate an intelligent question or comment that reflects something more than pre-conceived bias. Multiple posts on narrower topics might convey more information and promote more intelligent discussion.

75. Frank
Posted Jan 21, 2010 at 6:16 PM | Permalink

Judith: In the summary of your 2003 conference on sensitivity and feedbacks, I noticed a revealing comment in bold:

“the basic questions of how to make progress on
quantifying climate feedbacks and VERIFYING

Unlike the person who wrote this comment (and apparently most climate scientists), Professor Lindzen (and many readers of this blog) are interested in demonstrating that climate models are incorrect – not verifying them. In the traditional scientific method, theories are never proven to be correct. Instead theories USUALLY become accepted by withstanding challenges from the most demanding experiments that scientists can devise. The present political environment makes it difficult for this type of rigorous science to be done.

If one were to test Newton’s theory of gravity by observing the motion of planets, one could waste time verifying that Newton’s theory is consistent with observation for the outermost planets which move very slowly across the sky and are difficult to measure accurately. Or one could focus on the one planet that is subject to the strongest gravity and has the most accurate measurements. This leads to the discovery that Mercury’s motion is inconsistent with Newton’s theory.

It seems to me that LC09 has taken the second approach. We know that there is a controversial disagreement between the observed and predicted warming in the upper tropical troposphere, so it makes sense focus on that region (and not dilute our analysis with uncertain data from the rest of the planet). The flux of energy from the top of the atmosphere is greatest in the tropics, so our measurements are relatively more accurate there. Our ability to accurately measure changes in temperature and radiation diminishes over long periods of time, so LC has focused on short time periods with large temperature change. Water vapor and clouds are most critical feedbacks and they can be measured by changes in SWR and LW flux. So these choices appear to represent brilliant science, not grounds for criticism. (Longer term studies in the arctic, on the other hand are unlikely to produce conflicts with models and have the additional complication of large ice-snow-albedo feedbacks.)

Professor Lindzen’s paper assumes that water vapor and cloud feedbacks associated with large rapid changes in SST’s should take place within several months and be accurately represented by AMIP models. The key question is: “Are these assumptions correct?” (Even if Trenberth et al have shown different behavior in longer time periods.

(I do think LC took their analysis to far by claiming that the disagreements they observed between models and observation can be converted to disagreements about climate sensitivity. Climate sensitivity involves multiple time scales.)

• Patrik
Posted Jan 22, 2010 at 9:42 AM | Permalink

OT

• RB
Posted Jan 22, 2010 at 9:54 AM | Permalink

REsponse to OT

• Judith Curry
Posted Jan 22, 2010 at 10:59 AM | Permalink

Frank I especially agree with your last statement in parenthesis. Their “metric” is useful for comparing climate models and observations, but it is unconvincing as a measure of climate sensitivity

I prefer the word “evaluation” rather than “verification.” The challenge with climate models is the broad range of space/time scales involved. A global climate model is designed to reproduce the global transfers of energy and momentum. It is not expected to reproduce accurately individual weather systems, nor is it expected to replicate the climates of small regions.

Coming up with useful observational tests for climate models is very important, and i wish Lindzen and Douglass et al. would look at the globe rather than just the tropics, i think this is a better test and would provide more insights.

Climate modelers are interested in evaluation tests that can be used to point to specific areas for improvements. Everyone knows that climate models are imperfect, especially climate modelers themselves. If LC could take their analysis that extra step to identify whether it is clouds or water vapor, low clouds or deep convective clouds, etc that is causing the problem, this would be very valuable. I understand that they are starting to use cloudsat data (a cloud radar that provides vertical resolution of cloud charateristics), this should provide some insightful results.

• D. Patterson
Posted Jan 23, 2010 at 12:42 AM | Permalink

Isn’t that the very problem Chiang and Bitz [2005] and Yoshimori and Broccoli [2009] were unable to solve so far with their resources? If so, should Lindzen and Choi be criticized for not undertaking the effort with their current resources, or is there some reason to believe Lindzen and Choi may have the resources necessary to improve on those earlier experiments?

• Frank
Posted Jan 23, 2010 at 3:49 AM | Permalink

Thank you for your kind response. In my limited experience, the best way to improve a model is to look closely where the most robust data and a model disagree (the tropical upper troposphere). Assuming the model can be tuned to perform better in this region, then we can see what impact the improvements imply for the whole planet; preferably without over-publicizing the intermediate steps. Minimizing the statistical disagreement (Santer et al) or assuming the disagreement will disappear as old data is repeatedly reanalyzed and observations become more reliable (the NRC report?) may serve only to direct attention away from the observational data most likely to improve models.

I’m still stuck what seems be a key question: LC assumes that water vapor and cloud feedbacks associated with large rapid changes in SST’s should take place within several months and be reasonably represented by AMIP models. Are these assumptions correct (or at least reasonable)? If not, LC09 appears to be without merit.

• Judith Curry
Posted Jan 23, 2010 at 7:41 AM | Permalink

Frank, two points re your question. The water vapor and cloud properties change very rapidly in response to weather systems (time scale of hours). If there is something like an El Nino, this will impact the envelope of weather systems (not to mention the sea surface temperature), and so impact the water vapor and cloud properties. Changes in water vapor and clouds have an instantaneous impact on both the top of atmosphere and surface radiative fluxes. Large scale systematic changes in water vapor and clouds will influence surface temperature which in turn will influence water vapor and clouds. So there are a range of time scales involved for cloud and water vapor feedbacks. Note the LC methodology does not isolate cloud and water vapor feedbacks, but includes all potential feedbacks in the tropics during the time window of interest (in any event, this is likely to be primarily cloud and radiation feedbacks).

Re the AMIP models, Roy Spencer has found that the CMIP models agree much better with the ERBE data than the AMIP models. If you are trying to determine feedback with sea surface temperature being the main variable, you are not going to get the right answer if the model has fixed surface temperature (i.e. the sea surface temperature is not allowed to feed back into the model)

• Phil C
Posted Apr 25, 2010 at 6:34 PM | Permalink

There is one metric that captures all the possible climate effects of radiation absorption from whatever source- the heat-capacity, mass-weighted change in the global atmosphere/hydrosphere temperature- the total increase in energy in the system. Ideally this would include all the mass above the surface, but just getting good figures for the ocean and atmosphere would probably be adequate.

Unfortunately, we’ve only really just gotten started on that.

76. David Weisman
Posted Jan 21, 2010 at 6:29 PM | Permalink

I’ve been having some trouble understanding the premises of this paper – it could even be this is addressed, and I read that section without understanding, though I don’t believe so.

I take it we all agree there was an ice age thousands of years ago. Does that mean that either there is some stimulus to which climate must be fairly sensitive, or there must have been a dramatic change in some kind of forcing between the last glacial maximum 20000 years ago and now? If the pull towards equilibrium stability is so powerful, it would need to be overwhelming.

There are many explanations for alternating ice ages and interglacials, but this hypothesis would seem to create problems for them all, by arguing that any warming for any reason would be quickly countered by increased radiation, and vica versa for cooling.

77. Steve McIntyre
Posted Jan 22, 2010 at 1:33 AM | Permalink

I’ve placed quite a few comments into moderation and will look at them over the next day or two. Most of them pertain very tangentially to Lindzen and Choi and many try to debate AGW from first principles in short bites – an enterprise against blog editorial policies.

I haven’t moderated this thread as much as I should have. Unless you’re fairly experienced in the issues, I’d prefer it if you listened in to the discussion, rather than trying to raise issues even if they seem important to you. Feel free to participate in other threads.

78. stephen richards
Posted Jan 22, 2010 at 9:32 AM | Permalink

Frank

Well done. Most succinctly put. In software testing we do VV&T. Verification, Validation and testing.

Not done on climate models to my knowledge.

• Judith Curry
Posted Jan 23, 2010 at 7:46 AM | Permalink

Stephen, very substantial VVT is done on the climate models (well the best ones anyways such as ECHAM, GFDL, NCAR). I spoke at length on this on a previous thread, the one referred to above with myself, Gerald Browning, and Lucia (no link to that thread was actually identified tho).

• Posted Jan 23, 2010 at 9:06 AM | Permalink

Would that be Curry Reviews Jablonowski and Williamson in February 2008?

I totally accept that there must have been lots of VVT – or stuff called VVT – done with climate models. I have a lot of questions in this area and I think Dr Curry you would be a very good person to begin to answer them but I’ve been snipped just a little here and assume this isn’t central to the LC paper either.

• Judith Curry
Posted Jan 24, 2010 at 7:25 AM | Permalink

Richard, thanks much for finding the link. I am certainly not the most qualified person to address the VVT issue for climate models, but I can address some of the broad issues. perhaps a topic for another time/thread.

• Posted Jan 24, 2010 at 5:23 PM | Permalink

Thank you. I will read that thread with care before I try.

79. Dave F
Posted Jan 22, 2010 at 11:57 AM | Permalink

Isn’t it possible that climate sensitivity isn’t a static number?

• Dave F
Posted Jan 22, 2010 at 12:15 PM | Permalink

I mean that something, like cloud cover, could actually affect the sensitivity number instead of being represented as its own forcing. Or is this a silly question?

• Posted Jan 22, 2010 at 1:48 PM | Permalink

snip – enough chaos

80. Gerald Browning
Posted Jan 23, 2010 at 1:50 AM | Permalink

let us compare the science of Judith Curry (hurricane
intensity increased by AGW with minimal data to support
the claim) to that of Richard Lindzen who has developed a
reasonable theory backed by observational data and compared to climate models. I think we should ask Richard to review Judith”s work?

Jerry

• Judith Curry
Posted Jan 23, 2010 at 7:26 AM | Permalink

If Lindzen would like to take on such an assignment, it would certainly be very interesting (but i doubt he is interested). I would be happy to write an update on the hurricane story. By the way there is an important new paper out on this topic from the GFDL group by Bender et al. published last week in Science. Here is the link for GFDL’s press release with the citation (i haven’t found an online version of the paper yet.)

http://www.gfdl.noaa.gov/21st-century-projections-of-intense-hurricanes

• Gerald Browning
Posted Jan 23, 2010 at 3:40 PM | Permalink

Judith,

I am sure that Lindzen wouldn’t
waste his time on your work, but it would be amusing if he would.

As for your citation (an unverifiable modeling result), the quote below from
the abstract points out the futility of using the observational data . Isn’t that exactly what you did?

Jerry

Our view is that potential data problems need to be addressed before one can have much confidence in quantitative trends on time scales long enough to isolate the greenhouse-gas induced trend. In the absence of the clear detection of a trend, we cannot use the observations of recent decades to either confirm or refute the trends projected by our model.

• Judith Curry
Posted Jan 24, 2010 at 7:22 AM | Permalink

Jerry,

if you are interested in convective adjustment, here is what i suggest you do. Write a blog post that includes a description of convective adjustment along with a literature survey of its applications. Then conduct a mathematical analysis of this in the context of the relevant dynamical equations included in an atmospheric model. Then conduct a series of simulations using a regional mesoscale model: 100 km resolution and 10 km resolution, hydrostatic and nonhydrostatic, with nonhydrostatic simulations run both with and without convective adjustment. Then analyze the simulations.

Once you do all that, I would be happy to comment on your analysis. The bottom line is that someone has to do work. I don’t intend to waste any more of my time on topics that you suggest unless you yourself do some significant work on the topic and write a coherent blog post, since you advertise yourself as being capable of doing this kind of analysis.

• Gerald Browning
Posted Jan 24, 2010 at 2:56 PM | Permalink

Steve: snip – Jerry, please observe blog rules on civility

• Jimchip
Posted Jan 25, 2010 at 7:05 AM | Permalink

Professor Curry,

I read that paper and thought it was a pretty balanced approach. The last paragraph:

“Our results suggest that a significant anthropogenic increase in the frequency of very intense Atlantic hurricanes may emerge from the background climate variability in the latter half of the 21st century” followed by “Future studies should reassess our findings using both updated climate model projections and improved hurricane simulation models.” Sounds like real science to me.

You can imagine my astonishment… Bender wrote a Science paper 🙂

81. Alexander Harvey
Posted Jan 23, 2010 at 4:14 PM | Permalink

Judith,

The LC argument appears to me to fail at Equation 1:

deltaT0 = GO*deltaQ

as this is only true when (deltaFlux = 0).

It only holds at equilibrium where by definition the net Flux is zero.

This could be corrected for provided a similar correction is made to Equation 2, to get one at least as far as Equation 3. But at the cost of explicitly including deltaFlux in the intermediate workings and this prevents identifying the “The feedback parameter F” with deltaFlux/deltaSST.

Rather “the feedback factor” f, would at best be equated to 1-(deltaFlux/deltaSST)*Q (when there is no (external) forcing). This however would imply that the models (with negetive values of deltaFlux/deltaSST) displayed a value of f > 1.

In summary, I do not think one can analyse the data in this way, even if one tries to correct for the fact that Equation 1 as it stands does not hold for non-equilibrium states. As it stands one must strike out Equations 1&2, immediately that non-zero values of deltaFlux are introduced into the argument. Having done so the following arguments are without foundation.

To be honest I find the whole of [11] through [13] quite baffling at many levels but mostly I find these sections incorrect or and the conclusions baseless. However this does not in itself prevent the ERBE data from indicating negetive feedback, but it is not I think it is shown to be the case by the argument put forward. The case for the model data is much more worrisome as I do not think the arguments put forward demonstrate anything particularly meaningful.

If someone can demonstrate to me how one can derive meaningful insights from such dynamic data based on equations that refer to equilibrium conditions, I will be interested to listen.

Best Wishes

Alex

• Alexander Harvey
Posted Jan 23, 2010 at 6:20 PM | Permalink

In the above 1-(deltaFlux/deltaSST)*Q should I expect have read 1-(deltaFlux/deltaSST)*G0.

Alex

• Jim D
Posted Jan 23, 2010 at 8:01 PM | Permalink

Alex,I had a similar concern. In (2), it appears LC put their inferred feedback from a local ocean warming pulse, such as El Nino, with the atmospheric forcing from global CO2 doubling, a century-long process, as though the feedback to these two processes would be the same. The assumption is that the tropical black box will have the same feedback no matter what the nature of the processes leading to a given deltaQ. Of course, in El Nino there are all kinds of things like stronger convection in areas that might lead to a somewhat special feedback. I am skeptical that two such different processes can be equated in (2).

• Judith Curry
Posted Jan 24, 2010 at 7:14 AM | Permalink

Alexander, I agree with you (also with Jim D). I don’t think this kind of analysis can support determination of climate sensitivity particularly for short time periods.

• Alexander Harvey
Posted Jan 24, 2010 at 8:56 AM | Permalink

Judith, and Jim,

I wish I could get as far as you can and consider whether or not this type of analysis is useful at a conceptual level, but I just cannot get passed three paragraphs where their choice of equations simply baffles me.

As best as I can judge, putting all the other limitations to one side, their determination of f (the feedback factor) is out by 1 in all cases. For the ERBE data it should read f= -0.1 not f= -1.1 and I cannot see how anything else can be the case.

The statement that seems completely awry is their “The feedback
parameter F is deltaFlux/deltaSST, assuming the same incoming
radiation in the system.”. I do not know where this comes from, nor how the caveat about incoming radiation has been interpreted but as it stands it is wrong, or I am not having one of my more lucid days/weeks.

Now this is a shame, as it looks like the error is made in the same way in Figure 3, implying that if one just remembers to add 1 to f throughout, their conclusions may logically follow, and I would be able to consider loftier things in the manner that you do. As it stands I cannot get over the initial hurdle of picking my way through the baffling, I think think erroneous workings out.

I am not singling out this paper for special treatment I seem to find many papers using simple models baffling or just simply wrong at a logical level, not just a conceptual one.

Alex

• Jim D
Posted Jan 24, 2010 at 12:18 PM | Permalink

Alex, you should look at the guest commentary over at RC “Lindzen and Choi Unraveled” by Fasullo, Trenberth and O’Dell. They point out that it should be -0.125, not -1.1, and take this as a simple error that they don’t focus on. Furthermore other comments on RC have shown that LC got the wrong sign in their shortwave correlation, which would then make f positive. In LC, if you compare the first Figure in 1a (SST) with first in 2a (ERBE shortwave) you can visually see the correlation should be negative, not positive as LC got. If these criticisms pan out, LC is fast becoming shoddy, in my opinion.

• Jim D
Posted Jan 24, 2010 at 12:25 PM | Permalink

My 2a should be 1b, sorry.

82. Judith Curry
Posted Jan 24, 2010 at 8:25 AM | Permalink

I’ve gone back through the comments and posted a number of new comments to many of the threads. I’m not sure how to view the most recently posted comments, but i will continue to browse the threads for a few more days to see if there is further interest. Thanks to all of you for your comments!

83. Steve McIntyre
Posted Jan 24, 2010 at 10:11 AM | Permalink

Judith, the idea of “dumbing down” the science as a useful form of communication is one of the most pernicious attitudes of climate science. It results in an almost total absence of expositions that present results at levels between an elementary school cartoon and presenting results of a GCM run.

Scientists from other fields are starving for insightful expositions. As readers know, I’ve asked unsuccessfully for “engineering quality” derivations of 3 deg C from doubled CO2 for a few years – which is what I believe is required right now. Or for the “best” exposition that can be presented – not a cataloguing of GCM runs, but an exposition. IPCC documents don’t do this unfortunately.

Recently, Jerry North sneered at me in a interview with a Texas newspaper for making such a request and presented a totally fictitious account of a thread on his suggested reading that I presented at CA. This is the sort of attitude that has become all too prevalent in the “community”. I asked for this sort of presentation not because I don’t understand the basic idea of the “greenhouse effect” but because I wanted to discuss at CA what he thought was the best exposition of doubling CO2. North’s sneering was totally unjustified.

As someone who has now acquired considerable experience in discussing science with the public, my recommendation to the “community” is exactly the opposite: don’t “dumb down” the presentations. Try to engage scientists from other fields as a target market.

And going to “communications specialists” isn’t really what’s needed. You get into the Susan Hassool “re-framing” policies that seem in vogue in the “community”. I would recommend the opposite: as Lucia wrote well, climate scientists should try to answer the question that’s asked, rather than “re-framing”. They should clearly disclose adverse results. They should totally renounce disputes over data availability. They should renounce disputes over methodological transparency, providing source code where necessary to resolve any questions.

Most of all, they should stop blaming the audience or oil companies or fossil fuel interests if the audience doesn’t get their argument and look into the mirror and ask themselves how they could have made a more effective presentation. If every climate scientist was suspended for 6 months and had to make a living selling something, they’d come back to their job as a better climate scientist, both because they had to talk nicely to people for a while and because they’d be grateful for the opportunity that they have to be paid for doing an interesting job.

• Judith Curry
Posted Jan 24, 2010 at 10:44 AM | Permalink

Steve, I TOTALLY agree with you here. Unfortunately, the people who are “trying” to communicate with the public either think they have done enough (e.g. the IPCC reports) or are working to further simplify the science and arguments (mainly resorting to the “impacts” part of the argument), and then they bemoan the scientific ignorance of the general public.

I have been making the argument to my peers (not sure i am getting anywhere with this) that we need to satisfy the engineers on VV&T of the climate models, the statisticians on our uncertainty analysis, and the lawyers on the overall logic of the argument.

In the April issue of Discover Magazine, they are writing an article on the ipcc, climategate, uncertainty, etc. They seem especially interested in the uncertainty angle and how it is portrayed/communicated. Will be interesting to see how the article turns out, if they use some of my quotes maybe some of these ideas will get a wider audience.

The issue with the lawyers may be the one that gets the most traction in the short term, apparently climategate is motivating a lot of lawyerly activity re the IPCC and the overall global warming arguments.

84. Willis Eschenbach
Posted Jan 24, 2010 at 3:12 PM | Permalink

Judith, I was sailing along, agreeing with what you wrote, when suddenly you said:

Everyone knows that climate models are imperfect, especially climate modelers themselves.

No. Not. Nicht. Nein. Couldn’t disagree more. I fear that once again the lovely virtual world of models has claimed another victim

If the modelers “know that climate models are imperfect”, then why on earth are they using those POS models to forecast the future climate a century from now? To do that, you need a model that’s damn near perfect. I’ve built a host of computer models, and I know which ones that I’ve built are imperfect.

And I can assure you I don’t use those for hundred-year forecasts.

You can’t have it both ways, Judith. Either they think their models are imperfect, or they think they’re good enough for hundred year forecasts. Since each and every model claims to be accurate at the hundred year timeframe, I have to say that you are wrong. They don’t think they’re imperfect, they think they’re good for a hundred years.

Thanks as always for your contributions,

w.

• Judith Curry
Posted Jan 24, 2010 at 6:50 PM | Permalink

Willis, nowhere in the IPCC do they sell the 21st century simulations as “forecasts”. These are scenario simulations. The IPCC has been very careful in its wording on this. Popular use of the IPCC results may refer to forecasts, but if you go to the IPCC report, you will see that these simulations are referred to as scenario simulations. Let me know if you spot any wording to the contrary in this regard.

• Willis Eschenbach
Posted Jan 24, 2010 at 7:27 PM | Permalink

Judith, you say that the model results are not forecasts, but “scenario simulations”. I fear that this is a difference that makes no difference, which I will return to in a moment.

First, a few examples. IPCC FAR Chap. 9:

In a further study, Lee et al. (2006) assess whether anthropogenic forcing has enhanced the predictability of decadal global-scale temperature changes; a forcing-related enhancement in predictability would give a further indication of its role in the evolution of the 20th-century climate. Using an ensemble of simulations of the 20th century with GS forcing, they use Bayesian tools similar to those of Lee et al. (2005) to produce, for each decade beginning with 1930 to 1939, a forecast of the probability of above-normal temperatures where ‘normal’ is defined as the mean temperature of the preceding three decades. These hindcasts become skilful during the last two decades of the 20th century as indicated both by their Brier skill scores, a standard measure of the skill of probabilistic forecasts, and by the confidence bounds on hindcasts of global mean temperature anomalies (Figure 9.10). This indicates that greenhouse gas forcing contributes to predictability of decadal temperature changes during the latter part of the 20th century.

From Fig. 9.10

A forecast for the decadal global mean anomaly for the decade 2000 to 2009, relative to the 1970 to 1999 climatology, based on simulations performed with the Canadian Centre for Climate Modelling and Analysis Coupled Global Climate Model (CGCM2) is also displayed.

and

Constraints on the upper end of the likely range of climate sensitivities are also important, particularly for probabilistic forecasts of future climate with constant radiative forcing.

In the references for that chapter:

Allen, M.R., J.A. Kettleborough, and D.A. Stainforth, 2002: Model error in weather and climate forecasting. In: ECMWF Predictability of Weather and Climate Seminar [Palmer, T.N. (ed.)]. European Centre for Medium Range Weather Forecasts, Reading, UK, http://www.ecmwf.int/publications/library/do/references/list/209.

From Chapter 3

Climate models are being subjected to more comprehensive tests, including, for example, evaluations of forecasts on time scales from days to a year. This more diverse set of tests increases confidence in the fidelity with which models represent processes that affect climate projections.

So I’d say that they do in fact sometimes call them “forecasts” … go figure.

In any case, my question for you is, how does a “scenario simulation” differ from a “forecast”? What makes one a “simulation” and the other a “forecast”? If I put a bunch of information about celestial movements and the changes in the Earth’s orbit over time, and I say “The sun is going to rise at at 06:17AM on March 17th, 2100”, is that a “forecast” or a “scenario simulation”?

Finally, regardless of whether you call it a “forecast” or a “simulation”, I still say that the modelers are vastly overestimating the power of their creations. They are not saying that they are imperfect. They are saying that we can trust their century long “scenarios” enough to bet trillions of dollars on them … I don’t. Do you?

The idea that we can simulate the climate for centuries flies in the face of real-world experience using results from climate models.

w.

• Judith Curry
Posted Jan 24, 2010 at 7:48 PM | Permalink

Willis, a scenario simulation basically tests the sensitivity of the atmosphere to various scenarios of increased CO2. It does not pretend to predict the 21st century solar variations and volcanic activity, which would be needed for an actual “forecast.”

With regards to the quotes you identify in the ipcc report that use the word “forecast”, i would say that only the text re Fig 9.10 uses the word forecast inappropriately. Climate models can make seasonal climate forecasts which is what the Allen et al. ref refers to and also the text from chapter 3. The use of the word “forecast” in the context of the years 1930-1939 is a bit weird since this is the past, but does not refer to a forecast for future centuries.

Now search for “simulation” in the same chapters, and you will see that this is the word that is predominantly used wrt the 21st century climate model runs

• oneuniverse
Posted Jan 24, 2010 at 9:13 PM | Permalink

Dear Dr. Curry,

The Technical Summary contains at least one instance of ‘prediction’ and ‘projection’ being used more or less interchangably.

“The broad predictions of earlier climate models, of increasing global temperatures in response to increasing greenhouse gases, have been borne out by subsequent observations. This strengthens confidence in near-term climate projections and understanding of related climate change commitments.”
IPCC AR4 WG1 Technical Summary Box TS.7, p.59

The reference to ‘predictions of earlier climate models’ cannot be to the TAR, as that also referred to its model results as projections rather than predictions. (I cannot check for the SAR as it is no longer on the IPCC website, or I couldn’t find it elsewhere on the web).

Furthermore, I find it difficult to differentiate between the AR4’s definition of ‘climate projection’ (with respect to climate models) and ‘climate prediction’. I’ve placed the full glossary definitions at the foot of this comment, but excerpted :

“Climate projections are distinguished from climate predictions in order to emphasize that climate projections depend upon the emission/concentration/radiative forcing scenario used, which are based on assumptions concerning, for example, future socioeconomic and technological developments that may or may not be realised and are therefore subject to substantial uncertainty.”

Climate model predictions also depend in exactly the same way upon the “emission/concentration/radiative” forcing scenario used. The same uncertainties apply.

Would you be able to provide a concrete example demonstrating the difference between a climate projection and a climate prediction, based on the AR4 definitions?

From the IPCC 4AR WG1 Glossary :

Climate Model: “[..] Climate models are applied as a research tool to study and simulate the climate, and for operational purposes, including monthly, seasonal and interannual climate predictions.”

Climate Prediction: “A climate prediction or climate forecast is the result of an attempt to produce an estimate of the actual evolution of the climate in the future, for example, at seasonal, interannual or long-term time scales. Since the future evolution of the climate system may be highly sensitive to initial conditions, such predictions are usually probabilistic in nature. See also Climate projection; Climate scenario; Predictability.”

Climate Projection: “A projection of the response of the climate system to emission or concentration scenarios of greenhouse gases and aerosols, or radiative forcing scenarios, often based
upon simulations by climate models. Climate projections are distinguished from climate predictions in order to emphasize that climate projections depend upon the emission/concentration/radiative forcing scenario used, which are based on assumptions concerning, for example, future socioeconomic and technological developments that may or may not be realised and are therefore subject to substantial uncertainty.”

• oneuniverse
Posted Jan 24, 2010 at 9:34 PM | Permalink

My reading of this is that the IPCC considers a climate projection to be a climate prediction if the forcings scenario used is an attempt at predicting the real future forcings.

• Judith Curry
Posted Jan 26, 2010 at 10:03 AM | Permalink

This is correct. a climate prediction would include best estimate “predictions” of solar and volcanic activity, whereas a projection would not. But the line between projection and prediction is far too fine, and the 21st century runs should be referred to as scenario simulations IMHO

85. Rob Mitchell
Posted Jan 24, 2010 at 3:27 PM | Permalink

Hi Judith, I’m Not sure if anyone mentioned it in comments but surely the mechanism of shortwave negative feedback is the timing of thunderstorm development during the day. Higher SST result in earlier thunderstorms resulting in increased albedo. Of course this is a tropics only mechanism.
Cheers

• Judith Curry
Posted Jan 24, 2010 at 6:53 PM | Permalink

Rob, thanks for your comment. Willis Eschenbach did mention the issue of cloud diurnal cycle, this is a possible mechanism.

86. Willis Eschenbach
Posted Jan 25, 2010 at 12:00 AM | Permalink

Willis, a scenario simulation basically tests the sensitivity of the atmosphere to various scenarios of increased CO2. It does not pretend to predict the 21st century solar variations and volcanic activity, which would be needed for an actual “forecast.”

Correct me if I am wrong, but it seems that you are saying that a scenario simulation (a “projection”) is a forecast which is contingent on another forecast. The projection says “Here is a forecast of CO2 emissions for the next century. Assuming that this CO2 forecast is correct, ceteris paribus, here is what the temperature will be.” How is that not a forecast? It just contains two forecasts rather than one.

But again, I see no difference between that and a forecast of the time of the sunrise on Easter in the year 69000. Both of them are dependent on other forecasts, like the forecasts of the decreasing mass of the sun and the slowing of the earth’s rotation. Both of them contain the “ceteris paribus” clause. How is one a forecast and the other a “scenario”? Heck, even the Frame paper you cite above is called “Constraining climate forecasts: The role of prior assumptions” …

The IPCC TAR made the distinction as follows:

Climate Prediction
A climate prediction or climate forecast is the result of an attempt to produce a most likely description or estimate of the actual evolution of the climate in the future (e.g., at seasonal, interannual, or long-term time scales. See also climate projection and climate scenario.

Climate Projection
A projection of the response of the climate system to emission or concentration scenarios of greenhouse gases and aerosols, or radiative forcing scenarios, often based upon simulations by climate models. Climate projections are distinguished from climate predictions in order to emphasize that climate projections depend upon the emission/concentration/radiative forcing scenario used, which are based on assumptions, concerning, for example, future socioeconomic and technological developments that may or may not be realized and are therefore subject to substantial uncertainty.

So the only difference between a forecast and a projection is that the assumptions in a projection “may not be realized”. But again, the assumptions in a forecast may not be realized either … so what’s the difference?

Things also get quite sketchy when we ask the question, is the climate sensitivity a forecast or a projection? This question is quite relevant to the LC paper.

More to the point, however, is that they are treated as forecasts by both activist scientists and activist lay people. They either take one of the extreme scenarios or the median scenario, and use it to say “The sky is falling! We must spend trillions of dollars!” A quick look on Google finds thousands and thousands of web pages with the phrase “ipcc forecasts”. Even Google Scholar finds a hundred and fifty examples of the phrase …

Which is why I said it is a difference that doesn’t make a difference. When both scientists and lay people call them forecasts, and treat them as forecasts, at that point it’s walking like a duck and quacking like a duck …

• Judith Curry
Posted Jan 25, 2010 at 7:11 PM | Permalink

willis, this is a very unfortunate public perception about the models, which is not helped by loose use of terminology by scientists. For what its worth, i’ll add this to my list of things that need to be cleaned up

87. Alex Heyworth
Posted Jan 25, 2010 at 2:22 AM | Permalink

Judith Curry (Jan 19 18:47), Judith, you said you found George White’s web paper interesting and were going to look at it in detail. Be very interested to know your view once you’ve had that opportunity.

88. Kenneth Fritsch
Posted Jan 26, 2010 at 11:20 AM | Permalink

“Climate projections are distinguished from climate predictions in order to emphasize that climate projections depend upon the emission/concentration/radiative forcing scenario used, which are based on assumptions concerning, for example, future socioeconomic and technological developments that may or may not be realised and are therefore subject to substantial uncertainty.”

“Climate model predictions also depend in exactly the same way upon the “emission/concentration/radiative” forcing scenario used. The same uncertainties apply.”

I think we are making the distinctions between projections and predictions more complicated than need be. Projections include the uncertainties of the models predictions given the conditions (scenarios) and the uncertainties of these given scenarios coming to fruition. It would appear logical to attempt to separate these uncertainties and particularly from a science point of view because a scientist (climate modeler) would be inclined to think that she could eventually understand the science sufficiently to drastically reduce the CIs. Anyone who knows the fallibilities of economic forecasts with the elements of human action so intimately entwined would see the reasons for separating the hard (eventually hard and maybe less soft at the moment) science from the soft(er) one.

In a nutshell, it tells me that when all scientific studies are presented in making projections (and with all the uncertainties contained therein) we still have the very uncertain process of selecting the proper scenario. We can talk about volcano occurrence uncertainties but those should be temporary glitches over the long haul.

As for formal and proper usage of the terminology of projections, forecasts and predictions, it is only necessary to know the full considerations of what the user is talking about and then they can use whatever term as long as the reader/listener understands those considerations – and those considerations should be spelled out regardless of the terms used.

• Posted Jan 26, 2010 at 11:46 AM | Permalink

We can talk about volcano occurrence uncertainties but those should be temporary glitches over the long haul.

Do we know that for a fact? Does the long haul mean a billion years or something shorter?

• Kenneth Fritsch
Posted Jan 26, 2010 at 6:13 PM | Permalink

Richard, I am thinking in terms of the current regime of volcanism. I recollect that paleoclimatology in geological time shows sufficient rates of volcanism to alter the climate for long periods of time. The sulfate aerosols from volcanoes in current times that are attributed to changing albedo and cooling the globe are washed out of the atmosphere in at most years’ periods of time. I an assuming that without some rather drastic feedback effects during the aerosol residence time that the earth’s climate goes back to doing what it did before the eruption.

Since my references are primarily recollections from the History Channel, I stand ready to be corrected here. I think that volcanism has some credibility for causing the dinosaur extinction 60 million years ago either in whole or conjunction with the meteor impact.

That brings up an interesting point since eruptions of a sufficiently high level would have some effect on CO2 levels in the atmosphere. Could a high level of volcanism cause an immediate global cooling and then after the sulfates wash out of the atmosphere the increased CO2 levels would cause global warming. Or are there some conflicting feedback effects that operate to neutralize one or the other main effect? Has there been any modeling on this condition?

• Posted Jan 26, 2010 at 8:01 PM | Permalink

OT

89. oneuniverse
Posted Jan 27, 2010 at 7:42 PM | Permalink

Professor Curry: “Willis, nowhere in the IPCC do they sell the 21st century simulations as “forecasts””

IPCC AR4 WG1 Summary for Policy Makers p.15 : “Projected warming in the 21st century shows scenario-independent geographical patterns similar to those observed over the past several decades. Warming is expected to be greatest over land and at most high northern latitudes, and least over the Southern Ocean and parts of the North Atlantic Ocean”

It is clear from the above paragraph that the IPCC is interpreting their analysis of the results of the 21st century projections as a prediction of warming.

90. David Weisman
Posted Aug 5, 2010 at 8:57 PM | Permalink

I’ve been rereading the abstract and some of the discussion of Lindzen and Choi 2009. While many people have discussed the statistics itself, I’m still struggling with what they wanted to prove about climate sensitivity. Many scientists now believe the ice ages coincide with tiny orbital variations, but it took awhile to come up with the idea because the change in energy received by the Earth between ice ages and interglacials is minute. It seems the global temperature must be fairly sensitive to some stimuli over longer time scales – unless someone has posited a much more drastic forcing to explain the alternation between ice ages and interglacials.

Others have suggested changes in the sun could explain terrestrial temperature variation – but the actual energy variation in the sun seems to be small, so we have the same problem – if we try and postulate a low climate sensitivity.

It seems we may need a low sensitivity to certain things combined with a high sensitivity to others. Any ideas how this might work?

91. David Weisman
Posted Sep 20, 2010 at 7:12 PM | Permalink

Not much current interest in this paper on climate audit, although there was initial eagerness to have Judith Curry respond to it, and a lot of follow up discussion. I’d love to hear what some of the professional statisticians here think of the revisions.

Has anyone read the sequel, and looked at the statistical methods used? The abstract is on CiteSeerX.:

This work was subject to significant criticism by Trenberth et al. [2009], much of which was appropriate. The present paper is an expansion of the earlier paper in which the various criticisms are addressed and corrected. In this paper we supplement the ERBE data for 1985-1999 with data from CERES for 2000-2008. Our present analysis accounts for the 36 day precession period for the ERBE satellite in a more appropriate manner than in the earlier paper which simply used what may have been undue smoothing. The present analysis also distinguishes noise in the outgoing radiation as well as radiation changes that are forcing SST changes from those radiation changes that constitute feedbacks to changes in SST. Finally, a more reasonable approach to the zero-feedback flux is taken here.

• ianl8888
Posted Sep 21, 2010 at 3:09 AM | Permalink

Thanks for this link – occasionally I revisit this thread to reinforce (or remove) some point I think I vaguely remember

The problem with this link of course is that the full paper is behind a paywall … pity

92. Posted Jun 19, 2011 at 7:47 AM | Permalink

The debate between Dessler and Lindzen at Texas A&M in October 2010 can be viewed at: http://www.globalclimatescam.com/2010/10/richard-lindzen-of-mit-debates-andrew-dessler-of-texas-am-on-global-warming/ does round-up the issue on LC and the update.

Dessler made a good try at ridiculing Lindzen, at around 1:11:20 – 1:17:58 and Richard’s punch back after 1:21