Earth’s temperature over the past two million years


A new reconstruction of global average temperature over the past two million years has recently appeared in Nature (Snyder, 2016). That is quite a feat and a first for this duration. The figure below, made by Jos Hagelaars, shows Snyder’s temperature reconstruction, combined with the observed warming since 1880 and projected warming until the year 3000 for two IPCC scenarios, RCP6.0 and RCP8.5.


The RCP8.5 can be viewed as a “no mitigation” scenario, whereas RCP6.0 would be a “limited mitigation” scenario. It is clear that in both scenarios global warming over the next centuries will take us out of the temperature realm of the past two million years. A similar figure (which I tweeted yesterday) but then with temperature projections stopping in the year 2100 can be found here.

Even though lauded as a very valuable and novel contribution to the field, Snyder’s reconstruction has also been criticized because the temperature amplitude between glacial and interglacial states appears relatively large (~6 degrees) compared to other recent reconstructions, e.g. by Shakun et al (2012) (~4 degrees). Somewhat related, Snyder estimates the global average temperature during the previous interglacial (Eemian) to be warmer than now, whereas e.g. Hansen et al (2016, under review) argue that they are similarly warm. By the way, sea levels were 6 to 9 metres higher in the Eemian than now. Sea level responds very slowly to a change in temperature, yet another sign of the vast inertia in the climate system.


Somewhat overshadowing the actual temperature reconstruction that Snyder presented was her calculation of an earth system sensitivity (ESS) based on a correlation between temperature and CO2 over the past few glacial cycles. The earth system sensitivity denotes the long-term temperature response to a doubling in CO2 concentrations, including e.g. the response of ice sheets (which is typically excluded from the more often used equilibrium climate sensitivity, ECS). She then applied the ESS value of a whopping 9 degrees, obtained from this simple correlation, to the current warming, stating in the abstract:

This result suggests that stabilization at today’s greenhouse gas levels may already commit Earth to an eventual total warming of 5 degrees Celsius (range 3 to 7 degrees Celsius, 95 per cent credible interval) over the next few millennia as ice sheets, vegetation and atmospheric dust continue to respond to global warming.

Where “commit” means that this level of warming would be eventually expected based on current CO2 concentrations.

As Gavin Schmidt wrote, this is simply wrong.

The reason why I think it’s wrong is that in her calculation of ESS she takes the radiative forcing caused by albedo changes (resulting from the massive change in ice coverage between a glacial and interglacial state) and assumes it to be a feedback on the CO2 induced temperature-change.

There are two issues with this:

1) In reality both the changes in albedo (reflectivity) and CO2 concentration are feedbacks on the orbital forcing, and the relation in the one direction (a change in earth’s orbit causing a temperature change which in turn causes albedo and CO2 levels to change) is not necessarily the same as the relation in the reverse direction, as is currently happening with human-induced increases in CO2. Gavin Schmidt makes this point in two consecutive posts at RealClimate (here and here), though you might also want to read Hansen’s take, who has used a similar approach as Snyder did).

2) The ESS value obtained would (ignoring the more complex first point) perhaps be applicable to a glacial-interglacial transition, but decidedly not to an interglacial-‘hyperinterglacial’ transition, where the ice-albedo feedback would of course be much smaller because of the much smaller ice-covered surface area.

This second point was also made by James Annan in response to Hansen’s 2008 Target CO2 paper, where he essentially used the same method as Snyder is using (but came to a smaller ESS value of 6 degrees, because Snyder uses a greater temperature-amplitude between glacial-interglacial). Hansen noted in his paper though that “The 6°C sensitivity reduces to 3°C when the planet has become warm enough to lose its ice sheets.”

In other words, using Snyder’s very (and probably too) high ESS value to project future warming is unwarranted and wrong.


Tags: , , , , , , ,

4 Responses to “Earth’s temperature over the past two million years”

  1. thomaswfuller2 Says:

    Why do you treat the Representative Concentration Pathways as projections of potential realities? They are not.

    As I’m sure you are aware, they are ad hoc explanations of possible ways the planet could achieve pre-assigned levels of radiative forcing. Those levels were given as a starting point in the exercise to develop inputs to climate models.

    They are not projections. The researchers have specifically said they are not projections for over a decade. A casual look at the ad hoc explanations they came up with show that they are not at all serious, especially RCP 8.5.

    Worse, RCP 8.5 has already been invalidated by choices we have collectively made about green energy, emissions control and population.

    Only those wishing to foster a sense of panic would hold RCP 8.5 out as either a carrot or stick for influencing policy.

    You wouldn’t do that, would you Bart?

  2. Jos Hagelaars Says:

    “Why do you treat the Representative Concentration Pathways as projections of potential realities? They are not.”
    “RCP 8.5 has already been invalidated by choices we have collectively made about green energy, emissions control and population.”

    To me the RCP’s are clearly described by Bart as “scenario’s” and the RCP8.5 scenario is described as a “no mitigation” scenario.”. Indeed agreements are made about CO2-emission control so hopefully we will not follow that path. But you never know, there are still people in some governments who think climate change is a hoax. And we are still quite close to RCP8.5:

  3. wert Says:

    TWF: You wouldn’t do that, would you Bart?

    I agree with Thomas here, presenting RCP8.5 as no-mitigation future is an untrue statement. Please come back and tell you didn’t mean this.

  4. Bob Brand Says:

    As Thomas Fuller mentions:

    Worse, RCP 8.5 has already been invalidated by choices we have collectively made about green energy, emissions control and population.

    So it is actually mitigation (“emissions control”) which “invalidates” RCP8.5. By extension, without any mitigation RCP8.5 might be a valid projection — quite possibly even on the low side if one would consider only CO2 emissions.

    A point of criticism on RCP8.5 is that the net methane emissions are rather high, which impacts CO2-equivalent and it is the reason why the graph shows “>1000ppm CO2eq”.

    On the other hand, the RCP8.5 scenario already contains some plausible developments which would happen to limit CO2 emissions even without mitigation policies, such as an increasing role for natural gas and nuclear.

    See Riahi et al 2011:

    This paper summarizes the main characteristics of the RCP8.5 scenario. The RCP8.5 combines assumptions about high population and relatively slow income growth with modest rates of technological change and energy intensity improvements, leading in the long term to high energy demand and GHG emissions in absence of climate change policies.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: