The CO2 control knob fallacy?

Judith Curry has a recent post called CO2 ‘control knob’ fallacy (?). Admittedly there is a bracketed question mark at the end, but I get the impression that Judith thinks that using the term ‘control knob’ when referring to CO2 is probably a fallacy.

The term CO2 control knob is, admittedly, a bit of an analogy and analogies are never perfect. This one is, however, I think quite a good one. It really just refers to the idea that climate change (often measured as temperature changes) normally has some kind of driver (the control knob), the effects of which are amplified by other forcings and feedbacks. A classic is the 100000 year long Milankovitch cycles. Milankovitch cycles are variations in the Earth’s orbit around the Sun that cause variations in Solar forcing.

At the start of a Milankovitch cycle, the solar forcing increases, causing surface temperatures to rise. This then increases the atmospheric concentrations of CO2 and other greenhouse gases; increases atmospheric water vapour concentractions; reduces ice and snow cover and changes vegetation leading to changes in albedo. All this leads to a rise in surface temperature, globally, of about 5oC (although NH ice core samples show variations of around 10oC). Towards the end of the cycle, solar forcing drops; atmospheric greenhouse and water vapour concentrations decrease; ice and snow cover increase, and the system cools back down to the temperature at the start of the cycle. The main driver of this is the variation in solar forcing (driven by orbital variations) and, hence, in this case this is the control knob.

Today, the general view is that the rise in atmospheric CO2 concentrations is the main driver of climate change/global warming. Hence, the general view is the CO2 is the control knob and, hence, if we wish to avoid (or mitigate) against harmful climate change we should act to reduce the release of CO2 into the atmosphere. So, what is Judith’s issue? It essentially seems to be related to the roughly decade long pause (or hiatus) in surface warming. Judith says

Well, the ‘pause’ should give pause to anyone who thinks that CO2 controls temperature/climate/weather on the time scale of a decade. What about multi-decadal to century timescales? On these time scales, the big issue is the natural (unforced) internal variability……

If the climate is dominated by internal variability on these timescales, then external forcing causes only a small deviation in the climate, and the climate on these timescales is on a very long ‘leash’ with regards to external forcing.

So, Judith seems to be implying that the pause in surface warming since around 1998 is a consequence of internal variability and could exist for multi-decadal to century timescales. I don’t think many dispute that the hiatus is a consequence of internal/natural variability, but many would dispute that it could last for many decades or centuries. Also, it’s clear that internal variability is not really masking/cancelling anthropogenic warming. It is simply modulating the long-term warming trend, producing periods when surface warming is slower than expected and then others when it will be faster. Ten of the hottest years on record have been since 1998 and a recent article I read was arguing that, in the absence of anthropogenic forcing, we would have expected 0.3 to 0.4 degrees of cooling.

A more fundamental reason why internal variability is not that significant, from a global warming perspective, is that we appear to still be accruing energy at a rate of 1022J per year. Internal variability appears to be simply influencing how this energy is distributed through the system, rather than reducing overall warming (although there may be some influence from aerosols). So, is there a chance that internal variability could continue to maintain a slowdown for many decades or maybe a century?

At the current rate, it will take 50 years to increase atmospheric CO2 concentrations to 500ppm. This means that there will be an increase of radiative forcing, due to CO2 alone, of 5.35 ln(500/400) = 1.2 Wm-2. Currently we have a top-of-the-atmosphere (TOA) energy imbalance of about 0.5 Wm-2. There will likely also be changes in other forcings or feedbacks, but there is no real evidence to suggest that these should act to reduce the overall forcing. Hence, if surface temperatures were to remain flat, the TOA energy imbalance would increase to at least 1.7 Wm-2 and energy would be accruing at least 3 times faster than it is today.

There is a fairly basic issue with the possibility then that surface temperatures could remain flat for 50 years or more. The mass of the land and atmosphere is about 1019 kg with a specific heat capacity of 1000 J kg-1 K-1. Hence it would take 1022J to increase the average temperature by 1oC. If, in 50 years time, we’re accruing energy 3 times faster than today (3 x 1022J per year), then no more than 0.2% of this energy excess can heat the land and atmosphere if surface temperatures are to rise no faster than 0.05oC per decade. As little as 1% could cause surface temperatures to rise at 0.3oC per decade. This is also ignoring that other forcings and feedbacks would likely increase the energy imbalance compared to what it would be through CO2 alone. Essentially, we would need the fraction heating the land and atmosphere to decrease with time if we wish the surface warming hiatus to continue.

So, I find Judith Curry’s post a little strange in that it appears as though Judith understands the basics of radiative forcings and the influence of CO2, and yet seems to think that internal variability could produce a century long hiatus even as the radiation forcing due to CO2 continues to increase. It would require an ever increasing fraction of the energy excess going into the oceans and an ever decreasing fraction heating the land and atmosphere. I guess anything is possible, but some things seem to be appealing more to magic than to fundamental science.

This entry was posted in Climate change, Global warming and tagged , , , , , . Bookmark the permalink.

12 Responses to The CO2 control knob fallacy?

  1. BBD says:

    All this leads to a rise in surface temperature of about 10oC.

    There is debate about the difference in GAT between the LGM and the Holocene, but I think ~5C would be a more defensible estimate.

  2. BBD says:

    Whoops, hit “submit” too soon. As for the rest of your calculation vs JC’s – dare I say it – somewhat nonsensical notions of a multi-decadal or longer hiatus, you have made a far stronger case than she does, or could, given the physical constraints. I frequently wonder quite what JC is thinking of.

  3. I was referring to the changes during the Milankovitch cycles which (I thought) were about 10oC.

  4. BBD says:

    Well, the LGM is the coldest part of the last glacial, so we appear to be talking about the same thing… Unless you have in mind *NH* temperature change (centred on the N Atlantic) during the Younger Dryas event?

  5. Possibly I am. I was really just basing this on the classic Vostok Milankovitch figures which seem to show about a 10oC variation in temperature. So you’re saying that globally the temperature variation was close to 5oC than 10oC?

  6. BBD says:

    Yes. The Vostock and Greenland cores show the effects of polar amplification. Global average temperature change between glacial and interglacial states is generally estimated to be about 5C. This is the basis for the empirical estimate of S in Hansen & Sato (2012) which contains a referenced discussion of LGM/Holocene temperature estimates.

  7. Thanks, I shall have to correct that.

  8. Skeptikal says:

    I don’t think many dispute that the hiatus is a consequence of internal/natural variability, but many would dispute that it could last for many decades or centuries.

    Considering that natural variability is poorly understood, on what basis would you (or anyone else) dispute that it could last for many decades or centuries?

  9. Marco says:


  10. As Marco has already said, history, but that was also the point of the last part of the post. If Judith does not dispute the basics of radiative forcings and does not dispute that we’re accruing energy at about 1022J per year, then sustaining the hiatus for many decades or even a century would require an ever decreasing fraction of the excess heating the land and atmosphere and an ever increasing fraction going into the oceans and other heat sinks. Anything’s possible but some things are just very unlikely.

  11. James Westwood says:

    I’d dispute it because natural variability is not poorly understood at all. It’s difficult to model but not poorly understood. So while there are uncertainties with the predictions we can still with a very high level of certainty rule out the Curry’s hypothesis.

  12. Yes, I think that’s a good point. People seem to confuse the inability of an ensemble of models to all reproduce the same internal variability as indicating that we don’t understand it at all well. As far as I understand it, we might not be able to precisely predict the evolution of internal variability well, we still understand the actual processes involved quite well.

Comments are closed.