Gavin Schmidt (@climateofgavin) has been defending the detection and attribution chapter in AR5 against criticisms made by Judith Curry and others. His technique is to dismiss such criticism either because it covers a slightly different time scale or else because the authors fail to understand the IPCC ‘fingerprinting’ studies. He avoids addressing the growing evidence of a 60 year natural oceanic heat cycle or that its effects on warming are likely to be non zero between 1950 – 2005. This is despite the fact that a downturn in this cycle can also explain the 16 year long pause in warming since 1998. The IPCC fingerprinting studies described in chapter 10 of AR5 seem to be rather opaque and even a bit like ‘black magic’. This is because at its core the attribution always fall back on ‘expert’ assessments, whose quantitative reasoning is not documented. In particular, why is it that expert assessments of ‘natural variation’ in the temperature data are found to be essentially zero? Why then is the observed warming from 1940 to 2000 only 0.42C when compared to 0.6C from 1951 to 2005 on which the attribution studies are based?
Key examples of Gavin’s logic to any counter-argument are as follows:
– In reply to Judith Curry’s post 50-50 attribution.
Watch the pea under the thimble here. The IPCC statements were from a relatively long period (i.e. 1950 to 2005/2010). Judith jumps to assessing shorter trends (i.e. from 1980) and shorter periods obviously have the potential to have a higher component of internal variability. The whole point about looking at longer periods is that internal oscillations have a smaller contribution. Since she is arguing that the AMO/PDO have potentially multi-decadal periods, then she should be supportive of using multi-decadal periods (i.e. 50, 60 years or more) for the attribution.
Well I am not sure she was actually saying that. She was using a hypothetical short time period just to make her point clearer. Gavin jumped on that to then dismiss the rest of her points. Her main argument was indeed for the 1951-2010 period.
Here however, Gavin appears to be clutching at straws to avoid the conclusion of Tsung and Zhou that the AMO oscillation is real.
What is the evidence that all 60-80yr variability is natural? Variations in forcings (in particularly aerosols, and maybe solar) can easily project onto this timescale and so any separation of forced vs. internal variability is really difficult based on statistical arguments alone (see also Mann et al, 2014). Indeed, it is the attribution exercise that helps you conclude what the magnitude of any internal oscillations might be. Note that if we were only looking at the global mean temperature, there would be quite a lot of wiggle room for different contributions. Looking deeper into different variables and spatial patterns is what allows for a more precise result. –
The assessment that internal variability is zero is just based on “expert assessment”. So if you or I can see a clear 60 year oscillation in the temperature data, we are simply deluded because we are not “experts”. Only experts can interpret the ‘fingerprint’ of global warming.
– Second example: does not the following seem to be more like downright fiddling?
Judith’s argument misstates how forcing fingerprints from GCMs are used in attribution studies. Notably, they are scaled to get the best fit to the observations (along with the other terms). If the models all had sensitivities of either 1ºC or 6ºC, the attribution to anthropogenic changes would be the same as long as the pattern of change was robust. What would change would be the scaling – less than one would imply a better fit with a lower sensitivity (or smaller forcing), and vice versa (see figure 10.4).
So the ‘experts’ can scale up or down the anthropogenic forcings in the models so as to exactly match the data. Therefore because the models use stochastic internal variation while excluding multi-decadal variation, the ‘expert opinion’ of the modellers will be that natural variation averages to zero. In other words when you now scale zero by any factor whatsoever you still always get zero !
Third example – Now lets look at Fig 10.4
Both myself and Paul Mathews queried in Oct 2013 how it was possible for the ANT forcing to have small error bars while its component parts were much larger. Namely GHG has an error of 0.4 C and OA (aerosols) has an error of 0.35 C, while ANT has an error of just 0.1 C. The normal way to combine the sum of errors is to take their RMS summed value which would have resulted in an ANT error of 0.5 C.
This was Gavin’s response.
gavin says: 14 Oct 2013 at 1:01 PM
“Just for completeness, and to preempt any confusion, this post from Paul Matthews, is a typical example of what happens when people are so convinced by their prior beliefs that they stop paying attention to what is actually being done. Specifically, Matthews is confusing the estimates of radiative forcing since 1750 with a completely separate calculation of the best fits to the response for 1951-2010. Even if the time periods were commensurate, it still wouldn’t be correct because (as explained above), the attribution statements are based on fingerprint matching of the anthropogenic pattern in toto, not the somewhat overlapping patterns for GHGs and aerosols independently. Here is a simply example of how it works. Let’s say that models predict that the response to greenhouse gases is A+b and to aerosols is -(A+c). The “A” part is a common response to both, while the ‘b’ and ‘c’ components are smaller in magnitude and reflect the specific details of the physics and response. The aerosol pattern is negative (i.e. cooling). The total response is expected to be roughly X*(A+b)-Y*(A+c) (i.e. some factor X for the GHGs and some factor Y for the aerosols). This is equivalent to (X-Y)*A + some smaller terms. Thus if the real world pattern is d*A + e (again with ‘e’ being smaller in magnitude), an attribution study will conclude that (X-Y) ~= d. Now since ‘b’ and ‘c’ and ‘e’ are relatively small, the errors in determining X and Y independently are higher. This is completely different to the situation where you try and determine X and Y from the bottom up without going via the fingerprints (A+b or A+c) or observations (A+d) at all.
This is a remarkable statement (apart from the time-scale rebuff). In the above argument Gavin claims that Anthropogenic indeed means GHG + Aerosols, but that the two are inversely proportional. So apart from some minor terms, Aerosols = -const*GHG . That really is the same thing as saying
ANT = GHG – fudge factor*GHG
ANT = (1- (fudge factor))*GHG = all ‘observed’ warming – eliminating the need for any natural component at all.
Fudge Factor is the year to year tuning needed to make CMIP5 hindcasts agree with the temperature data.
Now look at what Chapter 7 of AR5 actually has to say about aerosols.
‘Aerosols dominate the uncertainty in the total anthropogenic radiative forcing. A complete understanding of past and future climate change requires a thorough assessment of aerosol-cloud-radiation interactions.
Or at what a real aerosol expert has to say
‘The IPCC is effectively saying that the cooling influence from aerosols is slightly weaker than previously estimated and that our understanding has improved.
Aerosol radiative forcing estimates.
The other values refer to different ways of calculating the impact and it is these numbers that inform the overall value in the report. The satellite based value refers to studies where satellite measurements of aerosol properties are used in conjunction with climate models; they are not wholly measurement based. In terms of studies using climate models on their own, the IPCC used a subset of climate models for their radiative forcing assessment, choosing those that had a “more complete and consistent treatment of aerosol-cloud interactions”.
The satellite based central value of -0.85 W/m2 is less negative than the central value from the climate models, which means the models indicate more cooling than the satellite based estimate. Compared to the subset of climate models that the IPCC used for their radiative forcing judgement, there is little overlap between their ranges also.
If we examine more details from the climate models, we see that there are large differences between models in terms of what types of aerosol it considers to be important. For example some models say that dust is a major contributor to the global aerosol burden, while others disagree. These are important details as climate models can sometimes broadly agree in terms of the radiative forcing estimate they provide but for very different reasons. Black carbon is another species that can contribute to varying degrees in different models, which is important as it warms the atmosphere; how a model represents black carbon is going to have a strong influence on the reported cooling. Nitrate is a potentially important species that often isn’t even included in climate models.
The current state of understanding of aerosols suggests that they’ve exerted a cooling influence on our climate, which has offset some of the warming expected from the increase in greenhouse gases in our atmosphere. Improving this understanding will be crucial for assessing both past and future climate change.
So these experts emphasize the uncertainties in the net forcings of aerosols and clouds. There is little evidence of any proportional anthropogenic cooling with CO2 emissions. Furthermore the models are still exaggerating the cooling effects of aerosols.
Finally here is a classic Gavin put down .
I have tried to follow the proposed logic of Judith’s points here, but unfortunately each one of these arguments is either based on a misunderstanding, an unfamiliarity with what is actually being done or is a red herring associated with shorter-term variability. If Judith is interested in why her arguments are not convincing to others, perhaps this can give her some clues.
This simply dismisses any evidence of a short term variation that is visibly present in the data, because it goes against the ‘expert fingerprint assessment’. Nor is the AR5 attribution time-scale anyway near long enough that natural variation averages to zero. This can simply be seen in the figure below.
Looks very much like a 60 year oscillation to me – but then again I am no ‘expert’ !
quotes above taken from these posts
- Natural versus Anthropogenic
- IPCC attribution statements redux: A response to Judith Curry
- The 50 50 argument
- The IPCC AR5 Attribution Statement