Scroll To Top

Friday, October 4, 2013

Wondering Willis Eschenbach is uncertainly sensitive at WUWT

Sou | 12:40 AM Go to the first of 9 comments. Add a comment

Wondering Willis Eschenbach's sensitive side

Wondering Willis Eschenbach has returned from his hot but remote airports and brings great news (archived here).  He has done an about face and is no longer an insensitive lout, but extremely sensitive when it comes to climate.

No more figuring out feedbacks.  We no longer need to wonder what impact the disappearing Arctic sea ice will have over what time frame, or whether clouds will have a net positive or negative feedback effect.  Wondering Willis has pronounced (in a convoluted post using a circular argument, in which he misses the point of a six year old paper) that climate sensitivity is as follows:

Climate Sensitivity = Climate Sensitivity

No need to worry any more.  Problem solved.  Willis says all the climate models can be dismantled.  There is no need for climate modellers to puzzle any more.  Just ask Willis.  Don't ask Jeffrey Kiehl.  Though Willis does give Dr Kiehl a pat on the head for effort:
Note that Kiehl’s misidentification of the cause of the variations is understandable. .... But as a first cut at solving the paradox, as well as being the first person to write about it, I give high marks to Dr. Kiehl.
(Kiehl attributed differences between the models in regard to climate sensitivity to uncertainty in aerosol forcing.  Willis argued Kiehl was wrong.  As far as I can gather, Willis attributed the difference between those same models in regard to climate sensitivity to differences in climate sensitivity!  These days differences between models in regard to climate sensitivity is attributed to uncertainty in cloud feedback.)

Kiehl, Jeffrey T. "Twentieth century climate model response and climate sensitivity." Geophysical Research Letters 34.22 (2007).

Willis is uncertain about uncertainty (very high confidence)

Despite being certain about climate sensitivity, I can say with very high confidence that Willis is uncertain about uncertainty (same article).  So much so that he "laughed because crying is too depressing".  He thought that when the IPCC report stated:
The model spread in equilibrium climate sensitivity ranges from 2.1°C to 4.7°C and is very similar to the assessment in the AR4. There is very high confidence that the primary factor contributing to the spread in equilibrium climate sensitivity continues to be the cloud feedback. This applies to both the modern climate and the last glacial maximum.
...that it contradicted the fact that there is a degree of uncertainty in cloud response.  But of course he got it all wrong (again).  Willis foolishly writes:
How can they have “very high confidence” (95%) that the cause is “cloud feedback”, when they admit they don’t even understand the effects of the clouds?
What was meant in the report was that there was very high confidence that the difference in estimates of climate sensitivity can be attributed to different estimates of the effect clouds will have on the radiation balance. The authors have high confidence that there is large uncertainty in regard to cloud feedbacks. Higher sensitivity would mean that clouds exert a stronger positive feedback, while lower climate sensitivity would be expected if changes in clouds exerted a less positive or maybe dampen the forcing with a slightly negative feedback.   This is from page TS-54 of the WG1 Technical Summary (my paras and bold italics):
The water vapour/lapse rate, albedo and cloud feedbacks are the principal determinants of equilibrium climate sensitivity (ECS, the equilibrium change in annual mean global surface temperature following a doubling of the atmospheric CO2 concentration). All of these feedbacks are assessed to be positive, but with different levels of likelihood assigned ranging from likely to extremely likely. Therefore, there is very high confidence that the net feedback is strongly positive and the black body response of the climate to a forcing will therefore be amplified.
Cloud feedbacks continue to be the largest uncertainty. The net feedback from water vapour and lapse rate changes together is extremely likely positive and approximately doubles the black body response. The mean value and spread of these two processes in climate models are essentially unchanged from AR4, but are now supported by stronger observational evidence and better process understanding of what determines relative humidity distributions.. Clouds respond to climate forcing mechanisms in multiple ways and individual cloud feedbacks can be positive or negative.
Key issues include the representation of both deep and shallow cumulus convection, microphysical processes in ice clouds, and partial cloudiness that results from small-scale variations of cloud-producing and cloud-dissipating processes. New approaches to diagnosing cloud feedback in GCMs have clarified robust cloud responses, while continuing to implicate low cloud cover as the most important source of intermodel spread in simulated cloud feedbacks.
The net radiative feedback due to all cloud types is likely positive. This conclusion is reached by considering a plausible range for unknown contributions by processes yet to be accounted for, in addition to those occurring in current climate models. Observations alone do not currently provide a robust, direct constraint, but multiple lines of evidence now indicate positive feedback contributions from changes in both the height of high clouds and the horizontal distribution of clouds. The additional feedback from low cloud amount is also positive in most climate models, but that result is not well understood, nor effectively constrained by observations, so confidence in it is low.
It all goes to show that no matter how much effort one takes to clarify meaning, there will always be someone who gets it all wrong.


  1. I've just written a post about observational constraints which (thanks to Karsten) I've realised is mostly wrong. It was, however, motivated by the Kiehl paper which has a plot showing that the ECS is larger when the adjusted forcing is smaller. This all seemed rather strange to me but I think is simply because the analysis assumes that the current heat uptake of the system (energy imbalance) is always the same and hence the method (given that it depends inversely on the adjusted forcing) produces a higher ECS for a lower adjusted forcing. So, the result in Kiehl seems to be strongly influence by that one assumptions which doesn't seem obviously well founded (i.e., if the adjusted forcing was actually lower you might expect the current energy imbalance to also be lower).

    Anyway, Willis seems to end all his posts with a comment along the lines of

    I think the whole concept of “climate sensitivity” is meaningless in the context of a naturally thermoregulated system such as the climate.

    which means, as far as I can tell, that Willis thinks our climate is controlled by something magical (or something as yet unknown) so I've taken to largely ignoring what he writes.

    1. The reason that paper found that result was because it was predefined in the models he looked at. In other words, he looked at papers that showed different climate sensitivity over the same historical period, when obviously the temperature rose the same amount and so did CO2.

      He then looked at what might have contributed to the difference in climate sensitivity and settled on differences in aerosols. (He also wrote: This strongly suggests that the scatter among the models is mostly due to the range in modeled change in ocean heat storage.)

      I don't know if he looked at other areas of uncertainty eg cloud cover but I didn't see anything like that in the paper.

    2. I'll add that my head started to spin a bit when I tried to unravel all this, so I could have Kiehl's paper wrong too :)

    3. I think you're right about the models, although there is quite a lot of scatter. I was mainly referring to his curves that just seemed a little odd (although they kind of fit the model results but I suspect the fit is not that good, statistically). Maybe the assumption is right but, mathematically, his curves come from fixing the term representing the energy imbalance.

    4. Yes, I think I got it fairly right. From Knutti (2008)

      Kiehl [2007] recently showed a correlation of climate sensitivity and total radiative forcing across an older set of models, suggesting that models with high sensitivity (strong feedbacks) avoid simulating too much warming by using a small net forcing (large negative aerosol forcing), and models with weak feedbacks can still simulate the observed warming with a larger forcing (weak aerosol forcing).

    5. Sorry, I wasn't commenting on your post or criticising your analysis of Kiehl as such (which I should probably have made clearer :-)). I was simply making a comment about how I found the curves in Kiehl (2005) a little strange (in that a lower adjusted forcing produced a higher ECS). What Knutti (2008) says makes sense although it just seems to be suggesting that you can get the same net warming (over some time interval) using models with different sensitivities but that have different net forcings (through variations in aerosol feedbacks for example). It's not clear how Kiehl concludes this because the function he uses to fit the data depends only on the adjusted forcing and the heat uptake of the system (H) so any aerosol influence is implicit and the reason he seems to get the curves he gets is simply because it depends inversely on Delta T_2x and everything else is fixed.

      Admittedly, I've been getting most things wrong today, so maybe I'm wrong about this too :-)

    6. Actually, I should have read Knutti more closely in that it says

      There is no correlation between the climate sensitivities of the CMIP3 models and their respective heat uptake efficiencies (the heat flux into the ocean per unit global surface warming at the point of CO2 doubling in a1%/yr CO2 increase scenario)

      So, maybe fixing H is kind of consistent with the models anyway. That would seem to imply something with respect to other forcings and feedbacks though (I would think) but maybe I should just give up for the day as I seem to be getting myself more and more confused :-)

    7. I think I understand what you are saying but I'm not 100% sure.

      If two models show the same temperature rise over the same period for the same CO2 forcing but predict different climate sensitivities then an inverse correlation between total forcing and sensitivity is implied, isn't it? The two notions are one and the same. The thing is to find out why - and Kiehl attributed it to differences in aerosol forcing.

      Model A predicts high sensitivity. Model B predicts low sensitivity. Both Model A and Model B show the the same temperature rise over the same period while the CO2 forcing was identical. Therefore either Model A must include some effect that is counteracting the (higher) CO2 forcing (eg incorporating a larger negative aerosol forcing) or Model B must be including something that is adding to/subtracting less from the CO2 forcing for the period in question (eg incorporating a much lower negative aerosol forcing).

      Remember aerosol from smog or volcanoes is a forcing not a feedback and is negative. It reflects incoming radiation. So it counteracts the positive forcing of CO2.

      Like I said, this can do your head in :D

    8. We may be talking at cross purposes a little here :-)

      Consider Kiehl's equations 3 and 4. Equation 3 is

      Delta Q = (Delta T Delta Q_2x)/Delta T_2x _+ H

      Delta Q is the change in forcing. Delta T is the change in temp. H is the ocean heat uptake rate. Delta Q_2x is the change in forcing for a doubling of CO2. If I understand this correctly, from observations Delta T = 0.6 degrees. H = 0.7 W/m^2. Delta Q_2x = 3.7 W/m^2. You can plug in the constants and you get Equation 4

      Delta Q = 2.22/Delta T_2x + 0.7

      Kiehl then uses this to produce the curves in Figure 1 and it shows an inverse relationship between change in forcing (Delta Q) and ECS (Delta T_2x). The point that I was trying to make (maybe not very clearly) is that what Kiehl has done here doesn't seem (as far as I can tell) to provide any physical mechanism for this relationship. It really just comes from the observational constraints (fixed Delta T, fixed H and Delta Q_2x). So, as I think you're saying, you need some other forcing that changes from model to model (aerosols) that can result in those models with different changes in forcings (Delta Q) and climate sensitivities (Delta T_2x) having the same change in temperature (Delta T).

      Anyway, I think you're quite right and I was just trying to make an observation about the relationship that Kiehl found between Delta Q and Delta T_2x.


Instead of commenting as "Anonymous", please comment using "Name/URL" and your name, initials or pseudonym or whatever. You can leave the "URL" box blank. This isn't mandatory. You can also sign in using your Google ID, Wordpress ID etc as indicated. NOTE: Some Wordpress users are having trouble signing in. If that's you, try signing in using Name/URL or OpenID. Details here.

Click here to read the HotWhopper comment policy.