The Uncertainty Of Climate Sensitivity

…and its implications for the Paris “negotiations”:

In my previous post Climate sensitivity: lopping off the fat tail, I argued that it is becoming increasingly difficult to defend high values of ECS. However, the uncertainty is sufficiently large that we can’t really identify a meaningful ‘best value’ of sensitivity, or rule out really high values.

A key issue is that emerging estimates of aerosol forcing are considerably lower than what was used in the AR5 determinations of ECS, implying lower values of ECS than was determined by the AR5.

This uncertainty in ECS makes emission targets rather meaningless. It will be interesting to see how this uncertainty is factored into the Paris negotiations

Note, there are other papers on this general topic that are in the review process, I expect a spate of such papers to appear during the next month.

Paris is doomed to failure, thankfully.

18 thoughts on “The Uncertainty Of Climate Sensitivity”

  1. Clearly, my own belief always has been that the climate studies have been a political power lever rather than actual concern with climate. I would guess that the politicians involved will not consider this issue to be much of a molehill, they will proceed with the power grab and leave the details to the bureaucrats

  2. It will be interesting to see how this uncertainty is factored into the Paris negotiations

    It won’t be, because it’s not really about “climate change” mitigation or prevention, of course.

    (Or to the extent it is, it’s about fantasies about it, rather than what science is actually properly practiced on the issue.)

  3. It’s zero. There is no observable direct cause and effect relationship between CO2 concentration and temperature. Between temperature and CO2 concentration, on the other hand, there is a very obvious relationship.

    It is a positive relationship. If there were a net positive effect of CO2 concentration on temperature, that would comprise a positive feedback loop, which could not be stabilized even with T^4 radiation.

    All things being equal, an increase in atmospheric CO2 should lead to an increase in surface temperatures. But, all things are not equal. In the present climate state, it is not possible for CO2, in the aggregate response, to significantly affect global temperatures.

    1. I sent Wood-for-Trees an e-mail. I wanted to know which temperature series was best, but I received no reply.

      Murry Salby uses that correlation between global temperature and net CO2 emissions to claim that only a small fraction of the increase in atmospheric CO2 that has everyone all worried can be attributed to human activity. I have posted on this subject before from the perspective of atmospheric carbon isotopes, and I have been working on this since this spring.

      Salby uses a linear model for ocean uptake of CO2. I am using a non-linear model of ocean uptake based on the chemistry of rates-of-reaction of the carbonate buffer system due to Revelle. I also use a two-compartment surface ocean-deep ocean model to account for the difference in radiocarbon age of those two ocean layers.

      Long-story-short, Salby’s mantra is “CO2 absorption (by the terrestrial biosphere) is proportional to concentration, CO2 emission (from gunk in swamps) is proportional to temperature.” Including only the absorption side of this (the terrestrial biosphere is a net absorber of roughly half the emitted CO2 that does not appear in the atmosphere, the ocean absorbing the other half), I get a really good match to the curve of the increasing atmospheric CO2 over time, from about 280 ppm in the 19th century to over 400 ppm as of 2015. In other words, if you smoothed out the wiggles in Wood-for-Trees curve, we are to blame for the increase in CO2.

      But, when I include an emission-proportional-to-temperature term (my guess is that as it warms, fallen trees, leaf litter, swamp gunk decays faster as the bacteria are happier), I get a different result. If I adjust the emission term to match the wiggles Wood-for-Trees is showing at your link, and if I increase the absorption term for plants taking up CO2 so averaged net uptake in the terrestrial biosphere stays the same, only half the increase in CO2 from 280 to 400 ppm can by blamed on humanity, with the other half the result of the warming in the 20th century.

      How about that, it has gotten warmer, which has driven half of the CO2 increase — how is that for “splitting the difference”?

      1. Which temperature set are you using?

        GISS and HADCRUT4 are pretty close now that they’ve both gone to the dark side, and use dubious infilling and extrapolation of bucket measurements for ocean temperatures. The interesting thing is when you compare the Northern hemisphere and the Southern.

        For over 100 years, NH and SH tracked in the surface data. Then, in about 2000, NH diverged radically from SH, according to their products. Yet, the SH still agrees with the satellites.

        Upshot is: I think it is excruciatingly apparent that the NH temperatures are being made up out of whole cloth.

        So, I would recommend SH only, as representative of what the satellites likely would have shown if they went back that far.

        1. I have to check, but I think I went with what Wood-for-Trees is using, which I believe is a HADCRU product. I e-mailed Wood-for-Trees on this question.

          Where is a good source for satellite temps?

          1. I think WFT has up to date data. I just think you should use only the SH product, as I think NH has been corrupted. Or, use satellite data (though, it does not go back as far).

      2. The relationship in the plot I first gave is pretty precise, much more so than I would have expected. It integrates pretty closely, too:

        You really do not need human inputs at all to get this very high fidelity agreement. What I believe is likely happening is that CO2 rich waters are upwelling, and the increase temperature is preventing it downwelling again in proportion. The imbalance gives you a steady, inexorable, and temperature dependent rate of accumulation in the surface waters, and thence to the atmosphere, according to the differential equation

        dT/dt = k*(T – T0)

        which, no doubt, is a linearized and transient response for a complicated nonlinear and long term dynamical relationship.

        1. I will take another look at that Wood-for-Trees plot.

          That “you don’t need human inputs at all to get this very high fidelity agreement” is pretty much what Murry Salby is saying.

          So at one end, you work with a linear model for ocean absorption of CO2 as aqueuous carbonates, and everything CO2 is explained as thermal emission with no need for the human inputs — in the absence of the temperature rise, the human inputs are swallowed up in the ocean.

          At the other end, you ignore the year-to-year changes in net atmospheric emissions, which Salby and Wood-for_Trees explains is large and otherwise unexplained, and the rise in CO2 is all the (unabsorbed by ocean and biosphere) human input.

          In the middle, using nonlinear ocean absorption, the year-to-year swings are temperature driven but the longer-term rise works out to half human input and half thermal emission. I am favoring the thermal emission coming from rotting vegetation in soils as better accounting for the carbon isotope ratios. The deep ocean has an old radiocarbon age, and substantial emissions from there throws the isotopes out of kilter.

          But this is a work-in-progress, and I will take a look at Southern Hemisphere and satellite temperature records I can find.

          1. Yeah, the thing is, people try to get a superficial match with the total CO2. But, total CO2 can fit any low order polynomial. There’s just not a lot of information to key off of there.

            Matching the derivative, however… that’s a lot harder, as there is a lot more going on than just the rising trend.


            That’s with the SH record. The match is even better with the satellite data:


            The remarkable thing is, when you estimate a scale factor to match the variational components, you pretty much match the trend as well.

            So, the problem with accounting part human, part natural is that you have to reduce the scale factor, and now your variational components in the derivative do not match so well.

            One could select a range of ratios of human to artificial that would preserve a decent match with the variational components. But, the simplest explanation, and the more probable, is that human inputs are having negligible effect, and it is almost all naturally driven.

    1. If they did, they would know that the purported tight regulation of CO2 prior to the Industrial Revolution is contradictory to the extreme sensitivity posited for human inputs. Wide bandwidth for the former, low for the latter – these are mutually exclusive.

      1. The stability in the face of significant step/impulse changes is the part that boggles me.

        The word “Catastrophic” is the only potentially troublesome aspect … and the stability makes that contraindicated.

      2. I am beginning to come around to regarding ice cores as doing heavy filtering of the swings. The low bandwidth for “the latter” is the effect of smoothing of the ice core layers “bleeding” gas into each other.

        1. I think the ice cores are total bunk, myself. There really is no way to validate them, as the overlap with modern records is A) short and B) a bit circular in logic, as they calibrate the ice core data using the modern records.

          1. Engineers as a class, I think, are likely to be less credulous about unvalidated data than scientists working in ivory towers. We frequently encounter the natural world’s non-compliance with models, even when the models are thought to be extremely well grounded. The Explorer 1 engineers never expected it to go into a flat spin. The Tacoma Narrows Bridge engineers never expected it to collapse because of mere wind.

  4. Well, clearly there is a need for a paper which generates a precise CO2 sensitivity figure near 3C per doubling of CO2. Might give the game away, if Mann coauthors it though.

Comments are closed.