Message Font: Serif | Sans-Serif
No. of Recommendations: 7
DB2: I am saying that looking at paleo data does nothing to constrain climate sensitivity estimates. They vary widely and come with more unknowns and larger uncertainties... Many forcings and feedbacks, such as clouds, can be difficult or impossible to reconstruct. Initial climate states and available feedbacks and mechanisms will vary greatly. Multiple states of a climate system can exist for a single set of boundary conditions.

You would be correct about paleo data if this were indeed the case. But it is not.

Parrenin et al. (2013) present a truly remarkable result. The 12 kyr from 22 to 10 kyr before present, a period spanning the Last Glacial Maximum and continuing into the start of the Holocene, was a time of great climate changes. Now if the climate situation was really as complex as you claim, then one would think it hard to find *any* regularity in these data. But just look at the plot of temperature change dT vs CO2 for these ice core data (for which the ice age has been very carefully calibrated). I've plotted these data below in blue (Parrenin et al. never actually show this remarkable figure), along with modern (1880-present) data (in red):

The CO2-dT plot shows a nearly perfect linear relationship, with an (amazing) correlation coefficient of R=0.994! This means that 98.8% (=R^2) of the variance in temperature over this 12 kyr period is explicable in terms of a change of CO2. This means that, statistically, these two variables are dependent (and effectively the same variable).

This highly-correlated behavior between CO2 and dT continues for the entire Vostok ice core record going back nearly half a million years:

The regression line is nearly identical to that found from the Parrenin et al. data, but the correlation isn't quite as good (R=0.87, which is still good). But most of the more poorly correlated data comes from the deeper ice (older than 100 kyr, shown in grey in this figure), for which it is harder to accurately register the temperature and CO2 data. It is probable that the most for this error comes from inaccurate ages, and if a more accurate age determination could be carried out, the CO2-dT plot would resemble the Parrenin et al data (which samples nearly the entire range in dT and CO2 of the 400+ kyr Vostok record).

The tightly defined linear relation between temperature change and atmospheric CO2 is remarkable. The simplest explanation is that this describes the equilibrium surface temperature relation of the Earth as a function of atmospheric CO2 concentration -- and it is totally reproducible over time. The temperature goes up and down many times over the course of the Vostok record due to small changes in orbital forcing being amplified into large temperature changes by the large climate sensitivity of the ice albedo feedback. But despite the almost chaotic behavior of the temperature sequence, the climate system remains confined to the CO2-dT equilibrium line -- the climate system just slides up and down this line, over and over again. Far from being complex, the paleoclimate system appears to be controlled by a single parameter that varies in step with the overall forcings.

Keep in mind that the orbital forcings never actually repeat over time (they are not really periodic), and neither do the albedo (ice sheet area) forcings, which depend on prior history. So if the climate system really depended sensitively upon the precise nature of these forcings, and the initial climate states -- as you claim -- you would not expect to see this simple behavior over time. Instead the climate system behaves precisely as one would expect if the equilibrium response depended only on the total forcing (with the details of the individual forcings being unimportant). It also means that the changes in paleoclimate forcings always occurred more slowly than changes in the climate system response (about 500-1000 yrs, set by ice sheet growth and decay rates), since otherwise we should see departures from equilibrium when forcings changed too quickly for the climate to keep up.

This means if we know the position of the climate equilibrium curve, we can predict *future* climates as well, as long as we don't exceed the bounds of the available paleoclimate regimes. This is a significant issue because nearly all of the easily accessible, recent paleoclimate data sample *colder* climates, whereas we are heading the other way. Hence the great interest now in reconstructing the warmer climate of the Pliocene. There is a lot more scatter in the Pliocene data, but I believe this is just to the difficulty of recovering CO2 abundances prior to the ice core record.

It is obvious from the Parrenin figure that the current climate system is far from the equilibrium curve, and moving away from it at an ever increasing rate, as CO2 emissions incresae year by year. Why is this? The Kirchner paper reference by DB2 mentions two possibilities:

In particular, we need to understand why, despite greenhouse gas concentrations that are unprecedented in recent Earth history, global temperatures have not (yet [in 2002]) risen nearly as much as the correlations in the ice core records would indicate that they could... [1.] This might indicate that the glacial-interglacial temperature shifts were amplified by ice-albedo fedbacks that are now much less influential (since the ice caps are already nearly gone, relative to ice age conditions). [2.] Alternatively, it may indicate that we have not yet fully felt the effects of important positive feedbacks ... that will amplify the warming experienced to date.

It is now clear that point [1] is not the cause, because temperatures in the Eemian interglacial were nearly 2 C warmer than at present, with substantial melting (about 1/3) of the Greenland ice sheet. Yet, the Vostock record shows the climate system remained on the same equilibrium line throughout the warmth of the Eemian, wheres it is now far away from this line.

Unfortunately, the deviation from the equilibrium climate relation is very likely due to point [2]. We are now driving changes to the greenhouse forcings a hundred times or so times faster than that due to natural variations in the past. This is simply too fast for the climate system to keep up and relax back to equilibrium. Some day in the future, when we stop the rapid increase in atmospheric greenhouse gases, the climate system will relax back to equilibrium.

How does knowing the equilibrium CO2-dT relation help? If this equilibrium relation is indeed universal, and we know the timescales to relax back to equilibrium, then given the history, and projected future history of carbon input to the atmosphere, we can work out (quantitatively) how the climate system will evolve. This is a project I'm working on, and I'll provide updates as I progress.

Print the post  


When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.