Uncertainty doesn’t imply nothing is known or nothing should be done

October 17, 2014 3:39 pm3 comments

While Judith Curry brings up some interesting notions in ‘Climate Sensitivity Uncertainy‘, I don’t agree with many of her arguments and conclusions.

I’m pleased that Dr. Curry acknowledges that “uncertainty in itself is not a reason for inaction” (see also ‘Why Climate Uncertainty is No Excuse for Doing Nothing‘ on TheConversation.com). I do find that conclusion slightly at odds with her frequent calls to put less effort in mitigation. Curry says that “deep uncertainties remain”, while at the same time apparently basing her anti-mitigation viewpoint on the assumption that climate sensitivity (ECS) is known to be low. If this deep uncertainty however extends to ECS, one would think that the risk of substantial warming entails a substantial risk that is worth hedging against. Is she so sure that ECS is low and impacts benign? Don’t uncertainties cut both ways? In short, I sense some inconsistencies in her approach to uncertainty.

Dr. Bart Verheggen

Dr. Bart Verheggen

There is no science without uncertainty, and climate science is no different. The climate system is highly complex and non-linear, so indeed systemic uncertainties abound. However, there are very different degrees of the level of (un)certainty surrounding different aspects of climate change. Many of the big picture aspects are reasonably well known, as very well said by Dr Dessler at the beginning of this year. It appears to me that Dr Curry is at times inflating the uncertainty to the point of creating the appearance of ignorance. I think that does a disservice to the prospect for “a more meaningful dialogue on how to address the complex challenges of climate variability and change”, which is a goal she frequently expresses to strive towards. As an example of inflated uncertainty, in contrast to what Judith Curry implies, it is “extremely likely” (as assessed by the IPCC) that the warming since 1950 is predominantly anthropogenic, and likewise is the projection that the warming will continue with continuing emissions very robust. There are uncertainties and ranges of probability, but the impression that this is totally up in the air is mistaken, to my mind. Perhaps in her (to my mind mistaken) belief that uncertainties are frequently ignored, she started over-compensating in the other direction?

Dr Curry writes “Even if CO2 mitigation strategies are successful and climate model projections are correct, an impact on the climate would not be expected until the latter part of this century.” That is very similar to what I wrote on CCNF earlier. But whereas Dr Curry implies that this lowers the need to reduce our emissions, I drew the opposite conclusion:

Therefore, preventive emission reduction measures (if desired) would need to be taken before the full extent of the consequences becomes apparent. This means that the longer we wait, the harder it will be to address the consequences of global warming, since by then we have committed ourselves to more warming. As such, those who caused the problem are in the best position to solve it, but since the full consequences will not materialize until much later, they have the least incentive to do so.

Image of LNG Super Tanker.  Source:  http://knowledge.allianz.com/

Given the inertia of the climate system, if we want to change course, it’s important to start steering the wheel in the desired direction in time. Source

In the same post, I also frame climate change as a wicked problem, but again, draw different conclusions and interpretations from those of Dr Curry.

Her recent article with Nic Lewis arrived at smaller ECS estimates than many other studies. They basically followed the same energy balance approach as Nic Lewis has advocated over the past few years. The current paper uses slightly different values for some key input parameters than Otto et al did, but is otherwise very similar. They did go through great length in making a detailed uncertainty analysis and as such can definitely be regarded as a worthwhile contribution to the literature. As Curry rightly points out, this study will of course hardly be the last word on the topic though, and it’s not a game changer. The pros and cons of the energy balance method are in both cases its simplicity. To their credit, they also discuss in their paper the extent to which this method really gives an estimate of ECS or merely an underestimate (namely, the effective sensitivity).

For policy considerations, it is of course important to not consider only one study, but consider (an assessment of) the scientific literature as a whole, such as IPCC does in a most authoritative way.

A big question for ECS research is why certain methods (based on energy balance models) seem to arrive at much smaller estimates than most other methods (based on paleoclimate, GCM’s or climatological constraints)? In that respect, recent studies such as those by Shindell et al and Kummer and Dessler are worth pointing out, as they tried to do exactly that. They concluded that the efficacy of aerosols may be greater than expected, which could explain the apparent discrepancies. I say “apparent” because the error bars around the estimates are still very wide and thus overlapping to a great extent, even if the best estimates are relatively far apart.

Source: Cowtan and Way

Source: Cowtan and Way

As Robert Way pointed out in the comments under her post, Lewis and Curry used a global temperature dataset which has a known cool bias because large portions of the Arctic are excluded. Using the temperature dataset by Cowtan and Way would bring up their estimates by a small amount. Likewise, more recent estimates of ocean heat uptake seem to point to larger values than used by Lewis and Curry. Lewis and Curry report median values as their central estimate, whereas many other studies report(ed) mean values. The former is usually smaller than the latter for typical ECS probability distributions. Taking the above into account would bring their estimates closer in line with those of most other studies (though a difference remains).

The paleo-estimates are interesting in the sense that from a variety of time periods and from a variety of studies and methods, ECS appears to be in the range between 2.2 and 4.8 degrees C. It thus seems that when someone advocates for a value lower than that, they have some explaining to do as to why such large temperature swings occurred in the (deep) past?

Finally, the policy relevance is probably a lot smaller than Dr. Curry makes it out to be, because even if ECS turns out to be as low as she thinks it is (1.64 degrees C per doubling of CO2), if we continue on a business as usual type pathway, we will still commit ourselves to a warming of over 3 degrees. Or, as Richard Millar wrote at RealClimate.org: this “might just be the difference between a achievable rate of emissions reduction and an impossible one”.

Climate Forcing and Climate Sensitivity with ranges of Lewis and Crok paper and IPCC AR5

Expected (transient or equilibrium) warming as a function of (transient or equilibrium) sensitivity. Figure adapted from Guido van der Werf. Source: Klimaatverandering.wordpress.com

Perhaps interesting to note that Nic Lewis and James Annan and John Fasullo discussed the topic of climate sensitivity to great lengths at ClimateDialogue.org. See also other responses to Judith Curry’s similar essay in the Wall Street Journal e.g. in the Huffington Post, in the Carbon Brief, and at the Union of Concerned Scientists.

THE FORUM'S COMMENT THREAD

  • Dr. Verheggen,

    Regarding your climate sensitivity graph: I have a few quick questions on equilibrium warming:

    A.) If I wanted to get a rough estimate of the equilibrium warming response to a tripling of the preindustrial atmospheric concentration of CO2eq (so 3 x 280 CO2eq ppm), I would just take my best sensitivity-per-doubling estimate on the bottom bar and multiply it by 1.5?

    And for quadrupling (4 x 280 CO2eq ppm), which seems quite possible if not likely in the long-term absent significant mitigation (unless we run out of coal first), I would take my best ECS-per-doubling estimate and multiply by 2 right?

    So as a hypothetical, let’s say that equilibrium warming per doubling is 2.4 C (which is in between Dr. Curry’s “best estimate” and the IPCC’s middle-of-the-road estimate of 3 C). For a rough measure, I would get:

    Tripling (CO2eq peaks at 840 ppm): 2.4 x 1.5 = Equilibrium warming of 3.6 C.

    Quadrupling (CO2eq peaks at 1120 ppm): 2.4 x 2 = Equilibrium warming of 4.8 C.

    Is this correct? If not, what am I not understanding?

    I find these Qs often pop up in my #SciComm when the topic of business-as-usual emissions is being discussed.

    B.) Also, am I correct in identifying these greenhouse gas levels as “CO2eq” (which means “carbon dioxide equivalent”)? This means that the concentrations of methane and other greenhouse gases are accounted for right?

    Gracious Thanks,

    MQ

    • Hi Michael,

      Indeed, for doubling of CO2 the eventually expected warming would be equal to the ECS value (as per the definition of ECS).

      For quadrupling it would be twice the ECS value.

      For intermediate values it’s slightly more complicated as pointed out by Andreas, since the relation is logarithmic rather than linear. You can use df = 5.35*ln(C/C0) to find the radiative forcing, and use e.g. the figure above to find the expected warming for a given ECS value.

      Alternatively, you can use the above equation to scale the radiative forcing to that of a doubling of CO2.

      A doubling of CO2 corresponds to ~3.7 W/m2 (as per the given equation). Tripling CO2 corresponds to ~1.59 times that. So it would eventually lead to 1.59 times the ECS value of value.

      The figure can be used if you know the radiative forcing, and is handy in light of e.g. the RCP scenarios which are given in terms of expected radiative forcing.

  • Hi Michael,

    your calculations are not quite correct.

    The temperature change dt = S*df, where S is the climate sensitivity in units of temperature change over forcing and df is the forcing in units of W/m^2. But the forcing for CO2 is not linear but logarithmic: df = 5.35*ln(C/C0), where C and C0 are the final and initial CO2 concentrations.

    Curry’s estimate of S = 1.5/3.7 = 0.41 K/(W/m^2).

    To get the temperature change for a tripling of CO2 the forcing is df = 5.35*ln(3) = 5.9 W/m^2 and dt = 0.41*5.9 = 2.4 K.

    For a quadrupling you’d get dt = 0.41*7.4 = 3.0 K.

    Andreas

Leave a Reply

You must be logged in to post a comment.

PUBLIC COMMENT THREAD

  • http://ClimateChangeNationalForum.org Michael Quirke

    Dr. Verheggen,

    Regarding your climate sensitivity graph: I have a few quick questions on equilibrium warming:

    A.) If I wanted to get a rough estimate of the equilibrium warming response to a tripling of the preindustrial atmospheric concentration of CO2eq (so 3 x 280 CO2eq ppm), I would just take my best sensitivity-per-doubling estimate on the bottom bar and multiply it by 1.5?

    And for quadrupling (4 x 280 CO2eq ppm), which seems quite possible if not likely in the long-term absent significant mitigation (unless we run out of coal first), I would take my best ECS-per-doubling estimate and multiply by 2 right?

    So as a hypothetical, let’s say that equilibrium warming per doubling is 2.4 C (which is in between Dr. Curry’s “best estimate” and the IPCC’s middle-of-the-road estimate of 3 C). For a rough measure, I would get:

    Tripling (CO2eq peaks at 840 ppm): 2.4 x 1.5 = Equilibrium warming of 3.6 C.

    Quadrupling (CO2eq peaks at 1120 ppm): 2.4 x 2 = Equilibrium warming of 4.8 C.

    Is this correct? If not, what am I not understanding?

    I find these Qs often pop up in my #SciComm when the topic of business-as-usual emissions is being discussed.

    B.) Also, am I correct in identifying these greenhouse gas levels as “CO2eq” (which means “carbon dioxide equivalent”)? This means that the concentrations of methane and other greenhouse gases are accounted for right?

    Gracious Thanks,

    MQ

    • http://ourchangingclimate.wordpress.com/ Bart Verheggen

      Hi Michael,

      Indeed, for doubling of CO2 the eventually expected warming would be equal to the ECS value (as per the definition of ECS).

      For quadrupling it would be twice the ECS value.

      For intermediate values it’s slightly more complicated as pointed out by Andreas, since the relation is logarithmic rather than linear. You can use df = 5.35*ln(C/C0) to find the radiative forcing, and use e.g. the figure above to find the expected warming for a given ECS value.

      Alternatively, you can use the above equation to scale the radiative forcing to that of a doubling of CO2.

      A doubling of CO2 corresponds to ~3.7 W/m2 (as per the given equation). Tripling CO2 corresponds to ~1.59 times that. So it would eventually lead to 1.59 times the ECS value of value.

      The figure can be used if you know the radiative forcing, and is handy in light of e.g. the RCP scenarios which are given in terms of expected radiative forcing.

  • http://ceoas.oregonstate.edu/profile/schmittner/ Andreas Schmittner

    Hi Michael,

    your calculations are not quite correct.

    The temperature change dt = S*df, where S is the climate sensitivity in units of temperature change over forcing and df is the forcing in units of W/m^2. But the forcing for CO2 is not linear but logarithmic: df = 5.35*ln(C/C0), where C and C0 are the final and initial CO2 concentrations.

    Curry’s estimate of S = 1.5/3.7 = 0.41 K/(W/m^2).

    To get the temperature change for a tripling of CO2 the forcing is df = 5.35*ln(3) = 5.9 W/m^2 and dt = 0.41*5.9 = 2.4 K.

    For a quadrupling you’d get dt = 0.41*7.4 = 3.0 K.

    Andreas

  • Pingback: The New Climate Pact between U.S. and China, details and reactions across media

  • Pingback: CCNF Columnist-members elect new board for 2015