My recent post The Coming Climate included a description of the orbital factors inducing natural cycles of warming and cooling far larger than any possible effect from CO2.

Yesterday commenter Alberto Zaragoza Comendador took this further into a discussion of the uncertainties in paleoclimatology. He started by referring to a paper by Lindzen, which focused on only one of these dynamics: fluxes of equator-to-pole heat transport.

Lindzen et al. 1993 concluded:
The present note shows the importance of aspects of the forcing which lead to changes in meridional (i.e., tropics to higher latitudes) heat fluxes. These aspects are seasonal, and involve the distribution of heating; they do not necessarily involve changes in globally and/or annually averaged insolation. Thus, simple, commonly used notions of climate sensitivity as employed in Houghton et al. (1990) are not relevant. Indeed, the present mechanism can readily produce major changes in climate (including, as a by product, changes in the globally averaged temperature) in systems which are profoundly insensitive to a doubling of CO2. To assume (as was done in Hoffert and Covey 1992, for example) that major climate changes necessarily require high sensitivity to such changes in gross averaged forcing is clearly inappropriate.

The full text of comment by Comendador is below from Climate Etc. (here)

Alberto Zaragoza Comendador | June 11, 2016 at 3:19 pm |

Somebody mentioned Milankovitch which reminded me of other thing.

http://www-eaps.mit.edu/faculty/lindzen/171nocephf.pdf (Lindzen et al 1993)

Check out the last paragraph. It explains why every paleo estimate of sensitivity is hopeless: even if you knew the temperatures, which you don’t really but even if you did, you’d have no idea what caused them. Lindzen uses the example of equator-to-pole heat transport but there are many more things that can cause climate to change, and we mostly know nothing about how they affected climate in the past.

We have records of methane and CO2, but we don’t have records of cloud albedo, ozone, water vapor, vegetation… someone might quibble that we do have some records of vegetation and dust for example, but unlike CH4 and CO2 you cannot assume these dust or vegetation ‘levels’ applied globally.

(Hell, until recently the greening trend of the last half-century was in dispute, even though we can look at the world’s plants and trees with seven billion pairs of eyes plus a few billion cameras, including some mounted on satellites. To assume we can now estimate greening or browning trends from twenty thousand years ago is preposterous.)

And we have estimates of how much area was covered in ice, but we have no idea how much dust that ice was covered with, or if the radiative forcing one could expect from the ice was really ‘apples to apples’ (i.e. of the same efficacy) as that of CO2. Now we know that as sea ice recedes the Arctic gets cloudier so the overall effect is about 1/3 of what you would expect simply looking at the decline in ice; we have no idea if the same thing would happen upon the disappearance of an ice sheet because we have never observed such a thing. Until recently we thought Greenland was reflecting less sunlight because it was getting dustier; it then turned out that, rather, the satellites’ sensors were getting degraded.

We still have little idea how aerosols affect climate… yet some guys are trying to model the aerosols of the past. And they’re not even the same kind (today the main agent is sulphuric acid, before dust).

(It’s funny that the ‘forcing efficacy’ issue has been raised about instrumental studies, and not about paleo papers that would be devastated if efficacy really changed much between forcing agents. For example, only about 20% of the forcing in LGM reconstructions is CO2).

You cannot simply assume that whatever change in GHG concentration (or other ‘forcings’) took place at the time of these temperature changes was responsible for said changes. In fact, by excluding other factors (which you know nothing about) you will systematically overestimate sensitivity. That’s why paleo sensitivity disagrees with both energy budget and inter-annual (ERBE/CERES) estimates. It also explains in part why the range of sensitivity in paleo is so wide.

Admittedly this is also a strike against sensitivity estimates using the instrumental record, but less so – because we have observations that allow us to rule out a good many ‘natural’ causes of climate change. We know that in the last 150 years the AMOC hasn’t shut down and there hasn’t been a massive change in equator-to-pole heat transfer. We know that since 1980 the amount of cloud cover has remained more or less the same, within a 5% band; the warming trend since then would be very difficult to explain from changes in cloud cover alone. And so on.

Whenever one mentions natural climate change the response from a certain side is something like ‘the ocean cannot create heat’ or ‘the clouds cannot change by themselves’. While technically true these statements are meaningless and reveal at best ignorance and at worst deception. All climate changes will involve a radiative change at some point, but said radiative change (forcing, feedback, whatever you call it) does NOT have to originate from a radiative source. It can all start with a change in air currents, or ocean currents, or tree cover, or sea ice, or methane emissions from bacteria, or…

Asking ‘yeah but what caused the clouds to change?’ is like asking why are there planets.

The best example of non-radiative climate change is in fact Milankovitch cycles, which affect not the amount but the distribution of sunlight (well eccentricity changes the amount of sunlight, but precession and tilt don’t). Technically speaking, the forcing is zero; paleo estimates consider GHGs, ice sheets and vegetation/dust forcings but if one is strict they should be considered feedbacks. Sensitivity, calculated the way it’s done for observational estimates, would be infinite.

You can also see how simply switching one of these radiative ‘things’ from forcing to feedback, or viceversa, can allow a researcher to arrive at a radically different sensitivity number. It’s all meaningless.

The one advantage of the paleo method is that since there is enough time for the ocean to reach equilibrium you avoid that source of uncertainty. But that also means you cannot use it to estimate TCR.

Anyway, as time goes on the estimates of aerosol forcing and heat uptake will get better and better. The instrumental studies will arrive at a number, if not for what sensitivity ‘is’, at least for what it has been for the last 150 years. The paleos will never arrive at anything.

Conclusion:

Thanks Alberto for that summation.  It deserves wide appreciation