A new hypothesis for deglacial CO2 rise?

It is well recognized that warming from a cold glacial period to a warmer interglacial is accompanied by increases in Carbon Dioxide concentrations, which at least amplified Milankovitch-induced forcing over millennial timescales. Atmospheric CO2 concentrations were around 100 ppmv lower than pre-industrial values during deep glacial climates. A key question in paleoclimate studies that is still not sufficiently addressed is how changes in solar insulation could have forced the feedbacks in atmospheric chemistry observed in proxy records.

There are many parts to the puzzle of how atmospheric CO2 changed between glacial and interglacial variations. A simple start is in ocean temperature, where we know from basic chemistry principles that a colder ocean would be better at “holding” more CO2 (since gas is more soluble in cooler water), thus lowering atmospheric concentrations. However, because so much fresh water was locked up in ice, the glacial oceans were saltier than today by a few percent, which reduces the solubility of CO2 gas in water. Changes in the biosphere contributed to some variations in atmospheric chemistry. Increased stratification of surface waters or increases in sea-ice cover that might have limited the release of CO2 to the atmosphere may be very important. There are many competing factors (explored for instance in Sigman and Boyle 2000 ). Also invoked as explanations are the finer points of the ocean carbon cycle such as the productivity of “the biological pump” (i.e., a process where CO2 fixed in photosynthesis is transferred to the deep ocean resulting in sequestration (storage) of carbon) as well as nutrient utilization of high-latitude oceans. Of particular importance is the large Southern Ocean which is key in the atmosphere/ocean CO2 equilibrium. There is evidence of injection of C14-depleted carbon into warmer times (meaning injection from a reservoir long isolated from the atmosphere, such as the deep sea). Most workers point to the Southern Ocean as a locus of deglacial CO2 release, evidence for this also coming from the correlation between atmospheric CO2 and Antarctic temperature records.

In a recent issue of Science, Anderson et al. 2009 provide insight into what may have caused a change in Southern Ocean circulation that would have altered the ocean-air CO2 exchange so substantially. Their answer involves the wind.

The authors use biological proxies to examine changes in upwelling during the last deglaciation. Maximum production of biogenic silica (opal) corresponds with maximum supply of dissolved nutrients to surface waters by the upwelling of nutrient-rich deep waters. Enhanced upwelling in the Southern Ocean during deglaciation is robust. Increased upwelling in the Southern Ocean coincided with the rise in atmospheric CO2 concentrations and Antarctic temperatures during the last deglaciation.

Supporting examples can be found in Abrupt Climate change events. Decreases in upwelling occur during the Antarctic Cold Reversal (14.5 to 12.5 thousand years ago), a time of anomalous cooling in the Southern Hemisphere just before the Younger Dryas in the North. Upwelling resumed after this event and was accompanied by a rise in CO2. Similar changes in CO2 are seen with the rapid warming of Antarctica at Heinrich event 1, interruption of CO2 rise at the Bølling-Allerød, and then CO2 rise warming occurring with Antarctic warming during the Younger Dryas (I didn’t get this backwards for those familiar with the YD; keep in mind the asymmetry between hemispheres as a result of the “see-saw” effect as a result of re-organizing ocean circulation).

Alley, 2007

I now do a brief detour to do a brief description of the Intertropical Convergence Zone (ITCZ) and the “westerlies” in either hemisphere.

In the Northern Hemisphere, viewed from above, surface winds blow clockwise and outward, away from subtropical highs.


As you can see, the westerlies are the winds north of the Horse latitudes (between about 30 and 35 degrees N and S under subtropical highs) and “underneath” the Polar Regions. In the North, the surface winds blowing from the northeast out of the southern flanks of the anticyclones are the trade winds. The reason air doesn’t flow directly from the equator to the poles, but rather is deflected, is because of the Coriolis effect (i.e., the planet is rotating). And contrary to what your high school science teacher probably said, that has nothing to do with how water spirals down your toilet or bathtub.

The trade winds of the two hemispheres converge in a broad east-west equatorial belt of calm winds, known as doldrums. The most active weather occurs at the ITCZ, a belt with thunderstorms that parallel the equator. The world-wide average location of the ITCZ (the “heat equator”) is located around 10 degrees N, due to the fact the North has more land and is a bit warmer.

So what’s that have to do with anything? The process that exposes the deep ocean water to the atmosphere today is the upwelling around Antarctica that is forced by the westerly winds over the Antarctic Circumpolar Current (ACC). Anderson et al. 2009 propose that during these times of ocean re-organization when the North was cool, the ITCZ shifted closer to the equator and the southern westerlies shifted further toward Antarctica. This image is taken from the “Perspective article” in Science

Fluid Dynamics Lab in Princeton

When the strongest westerlies are relatively close to Antarctica and are aligned with the ACC, they are associated with more upwelling. If the glacial westerlies were further to the equator than now, then there would be a buildup of respired CO2 in the deeper ocean as a result of reduced wind stress near antarctica. In general, glacial times follow this pattern. The interesting thought to take from this paper is that this has been a leading mechanism to cause CO2 changes over glacial-interglacial timeframes.

I personally have a quibble with the focus of the paper on NH circulation anomalies that are associated with the Abrupt climate shifts (e.g. Heinrich events and D-O events) for the simple reason that the full range of CO2 variability between glacial and interglacial cycles is evident over the last 800,000 years, not just the last (de)glacial. And in fact, the full range (~180-280 ppm) is amazingly consistent over each peak and trough, whereas it is not self-evident that abrupt climate anomalies followed similar patterns. Remember that temperature wise, these are Hemispheric-scale and relatively short events super-imposed on a much longer Milankovitch-paced timeframe. The argument needs to be extended over a glacial timeframe, not a YD timeframe. I’m convinced by the concept, but I’m not convinced that they’ve demonstrated a significant mechanism for CO2 fluctuations over the last million years, particularly since the spatial and temporal extent of the temperature anomalies are different between glacial-interglacial and stadial-interstadial variations. This criticism may be ill-founded though.

Conceptually, this follows the format of a positive CO2 feedback (also proposed for instance in Toggweiler et al 2006) since warmer temperatures and higher CO2 levels are associated with a poleward displacement of the westerly winds, which according to this hypothesis, would result in further “mining” of CO2 from the ocean to put into the atmosphere.


20 responses to “A new hypothesis for deglacial CO2 rise?

  1. Do you or these “scientists” know that CO2 LAGS temperature. It has nothing to do with causing temperature rise.

    Response– Oh Please. I would enjoy having at least one intelligent conversation without feeling the need to include these contrarian talking points.– chris

  2. Are the authors essentially arguing for single CO2 source which then becomes well-mixed? Essentailly the Southern Ocean being the “controller” of glacial-interglacial CO2 variations…

  3. Chris – Thank you for commenting on our recent paper in Science. I would like to respond to the points raised near the end of your comment. Specifically, I will try to put relatively short events like the Younger Dryas into the context of the global transition out of the last ice age.

    To begin, as noted in the second paragraph of your comment: “There are many parts to the puzzle of how atmospheric CO2 changed between glacial and interglacial variations.” Our paper addresses only one part of the puzzle, but we believe that having this part in place will accelerate progress on filling in the rest of the picture.

    In the penultimate paragraph you state: “Remember that temperature wise, these are Hemispheric-scale and relatively short events super-imposed on a much longer Milankovitch-paced timeframe. The argument needs to be extended over a glacial timeframe, not a YD timeframe. ”

    You are correct. The Younger Dryas is just one of several periods of severely cold climate conditions in the northern hemisphere during the last ice age. The others include Heinrich Events 1 through 6 as well as cold intervals preceding the numbered Heinrich Events. As pointed out in Figure 6 of our paper, ice core records reveal that Antarctica warmed and atmospheric CO2 rose during time periods surrounding each of these northern cold events. Our contribution toward filling in the puzzle was to add evidence that upwelling in the Southern Ocean also increased during these events. Combining our results with evidence from other studies, we concluded that the zonal wind systems shifted southwards during each northern cold period, creating conditions that favored upwelling in the Southern Ocean and release to the atmosphere of CO2 that had accumulated in the deep ocean.

    But, this clearly is not the whole story for the 80-100 pm rise in atmospheric CO2 that occurred at the end of each of the past several ice ages. This is evident because the ice core records indicate that following most of the northern cold intervals, during which atmospheric CO2 rose by 10 to 20 ppm, Antarctica then cooled again and CO2 fell back to roughly the level that existed prior to the start of the northern cold interval. The ice age did NOT end.

    So, you are correct to ask: What is special about the terminations of the ice ages? Why did atmospheric CO2 rise by 80-100 ppm instead of 10-20 ppm? What kept CO2 levels high for thousands of years? This is where the complex details of the climate – CO2 connection still need to be worked out. But we have some valuable clues.

    For one thing, as noted in your posting on “Follow up to lecture by Chris Walcek at HVCC ” (which is excellent, by the way, and I encourage you to add it to your listing of “Recent Posts”), another important factor affecting the transition from ice ages to warm interglacial periods is “orbital forcing”; i.e., the spatial and seasonal variability of solar insolation that reaches the surface of the earth.

    Several papers have pointed out that ice ages end when earth’s orbit is configured to provide maximum summer solar insolation at high northern latitudes. These conditions favor melting of the northern ice sheets and lead to the end of ice ages. This is the original Milankovitch hypothesis, and it has been explored in detail by several recent papers (e.g., see Figure 4 in Kawamura et al., Nature, Vol 448, 23 August 2007, doi:10.1038/nature06015; also see papers by Maureen Raymo and by Peter Huybers).

    Our hypothesis is that the end of an ice age and an 80-100 ppm rise in CO2 require the combination of orbital conditions that favor melting of the northern ice sheets (Milankovitch) together with a cold event in the northern hemisphere that creates upwelling favorable conditions in the Southern Ocean to vent CO2 from the deep ocean. Whether you consider this a feedback or a synergistic effect, the combination of high summer insolation in the north *together with* greenhouse warming from CO2 vented from the Southern Ocean is required to end an ice age.

    We believe that the hypothesis described in the paragraph above applies to ice ages in general, and not only to the Younger Dryas and the end of the last ice age. If you look at the CO2 record from Antarctic ice cores you see that each glacial period is characterized by short intervals where CO2 rose by 10-20 ppm. We suspect that those intervals are analogous to similar features of the last ice age, as shown in Figure 6 of our paper. Unfortunately, the Greenland ice core record does not extend back in time far enough to tell us if the earlier intervals where CO2 rose by 10-20 ppm coincided with periods of intense cold in the northern hemisphere. Nevertheless, the regular relationship between rising atmospheric CO2 and cold northern conditions during the last ice age makes it reasonable to suggest that short intervals of rising CO2 associated with intense cold in the northern hemisphere were characteristic of previous ice ages as well. These events occurred every several thousand years, but conditions were not right to end the ice age. Therefore, it is reasonable to suggest that the end of an ice age generally involved a combination of summer insolation in the north that was favorable for melting ice sheets together with winds in the south that were favorable for upwelling and release of CO2 to the atmosphere.

    Other factors are involved as well in creating the deglacial rise of 80-100 ppm CO2. Warming ocean temperatures during the transition out of each ice age reduced the solubility of gases in seawater, thereby providing a positive feedback for rising atmospheric CO2. The lowering of ocean salinity by melting ice sheets increased CO2 solubility, partially offsetting the positive feedback due to rising temperatures. The release of CO2 to the atmosphere caused by increased upwelling in the Southern Ocean caused the ratio of alkalinity to total dissolved inorganic carbon in seawater to increase. That change in ocean carbonate chemistry increased the preservation and burial of calcium carbonate in deep sea sediments, which added another positive feedback by driving the partial pressure of CO2 in seawater upward. This process is referred to as calcium carbonate compensation and it operates in the reverse sense as well (i.e., in lowering atmospheric CO2 during the initial phases of the ice ages). Once the ice sheets melted and sea level rose enough to cover the continental shelves, coral reefs and carbonate platforms started to expand. This removed additional calcium carbonate from the ocean, further reducing the ratio of alkalinity to dissolved inorganic carbon in seawater. Following the same principles invoked above for the increased burial of calcium carbonate in deep sea sediments, the expansion of coral reefs and calcium carbonate deposition on continental shelves provided another positive feedback that added to rising atmospheric CO2. In contrast to processes in the ocean, expansion of the biosphere on land following the melting of ice sheets and warming of earth’s climate represented a negative feedback, drawing CO2 out of the atmosphere.

    All of these factors combined to create the observed 80-100 ppm difference in CO2 between glacial and interglacial periods. Our work puts into place one key piece of the puzzle, but more work is needed to quantify the contributions of each of the other processes identified above.

    In addition to responding to your posting, I would like to add short post scripts to the comments by Tony and by John.

    For Tony: The fact that atmospheric CO2 lagged behind the initial change in climate at the end of past ice ages is a red herring frequently tossed around by global warming skeptics. No scientist that I know believes that changes in CO2 provided the initial cause of glacial-interglacial climate change. Rather, CO2 represents a feedback that amplified climate change that was triggered by other processes. It was recognized decades ago that insolation alone is likely insufficient to drive the changes in earth’s climate between glacial and interglacial periods. Scientists were looking for a feedback mechanism to amplify the effects of insolation before the first ice core records revealed that atmospheric CO2 was lower during glacial periods. It seems that CO2 provides that amplifier.

    For John – In order to put as much CO2 into the atmosphere at the end of each ice age as is observed in the ice core records, the deep ocean must have been involved. No other reservoir of carbon could have put that much carbon into the atmosphere that fast. Old deep water masses that are enriched in respiratory CO2 come into contact with the atmosphere only in the Southern Ocean. So, the short answer to your question is “yes”. Although there were other processes involved (see my comments above), especially during the latter stages of the transition out of the last ice age, initially the Southern Ocean was the source of carbon that spread throughout the rest of the system.

    Response– Dr. Anderson, Many thanks for your comments and detailed explanations, which are no doubt educational to myself and other readers. You’ve elaborated on quite a bit. The public could use more of such dialogue outside of peer-reviewed literature. Feel free to drop in anytime.

    There is also similar discussion related to CO2’s role as a forcing and feedback going on in my recent thread on the Eocene-Oligocene transition.– chris

  4. Chris:I finally understand what you were saying in your paper The Greenhouse effect Pt2..That the energy of the IR photon is converted to the vibrational state of of the GHG molecule.At different vibrational modes they therefore absorb at more than one wavelength.OK, Then you start talking about diatomic molecules. The nuclei is stretched because they are tightly bound to the elcromagnetic wavelength.,BUT? they do not absorb heat and contribute to the greenhouse effect.How do they contribute to the greenhouse effect?I like what you said, that the motion of molecules themselves alter the spectrum, and their collisions cause their absorbtion.Can you tell me why methane would not have more vibrational states than co2.Thanks,Kipp

    Response– The other molecules do matter in the sense that CO2 can radiate because it receives energy from surrounding air molecules – usually N2, O2, argon– it’s just that they are not themselves greenhouse gases. CO2 molecules also excite surrounding air molecules and heat the atmospheric level at which they reside. The Martian atmosphere has very little mass and so cannot sustain a large GHG effect. Unfortunately, for a more comprehensive picture of the underlying quantum mechanics (also relating to your question on methane) you’re going to need to consult a textbook or expert in the subject. Introduction to Atmospheric Radiation by Grant Petty may be a good start. The HITRAN archive also had data on the allowable vibrational modes for various molecules– chris

  5. Just a question, than I will study this paper.Water vapour could increase by solar Irradiance alone.Accept that CO2, allows for the atmosphere where solar irradiance can reach the warmth necessary to make water vapour a feedback factor.K

    Response– I have no idea what you’re trying to say here. The H2O saturation vapor pressure goes up and down for any forcing (CO2 or solar irradiance) and the water vapor enhances the Earth’s sensitivity to forcings of all sort.– chris

  6. Chris:You answered both my questions,even the incoherent one.Thanks. I have been working at this for more than a month.In the second one I was just trying to say that CO2 causes more of a feedback than solar irradiance. Sorry to bug you.

    Response– Maybe in the 20th century it does, but the statement “CO2 causes more of a feedback than solar irradiance” doesn’t necessarily have to be true. The feedback (water vapor, ice-albedo, etc) more or less arises from the warming, not what caused the warming. So, whatever caused the majority of warming during a specific time period (whether it be CO2, methane, solar irradiance, martian beams, etc) will “cause more of the feedback.” I’m not sure anything meaningful arises from such wording though.– chris

  7. Dr. Anderson,

    Thanks for your lengthy comment.

  8. Chris:Yes the warming or temperature determine how much water vapour there will be. You also said that absorption by a particular gas occurs when the frequency of the electromagnetic radiation is that of the of the molecular vibrational frequency of the gas in question.From this I can assume that any greenhouse gas would have to meet this criteria.Also, that is why other gases pass right through transparently.I have taken all your advice including buying several books on Amazon.I really don’t like to ask you anything unless I am completely stuck and don’t have the books or understanding to move forward.I am 56 years old and have dedicated the rest of my life to spreading out the truth about a warmer world, what could happen,and it’s necessary mitigation. I have even stopped bashing deniers and prefer to use any psychology to win them over from the dark side. Humanity depends on some of us to grow up and face the truth, however difficult it may be. I hope I am not preaching to the choir. Thanks, KIpp

  9. Chris: I appreciated your input, and never asked Roy Ladbury for his input. He read a conversation I had and offered the answers just by reading through the posts. When I learn new science than I own it and can honestly understand what I am saying to deniers and delayers, to convince them of the urgency and change that must happen now. I am one of a handful of people at my site who does this. My credibility demands that I am much more than a puppet. I don’t think at this early stage anyone deserves an award until we have turned the corner to sustainable energy, and a world consensus exists.

  10. Chris,
    I’m interested in a closer look at the graph from Alley 2007 you’ve provided. Can you tell me where to find Alley 2007?


    Response– The full reference is “Wally Was Right: Predictive Ability of the North Atlantic “Conveyor Belt” Hypothesis for Abrupt Climate Change” Alley, Richard B., Annual Review of Earth and Planetary Sciences, May 2007, Vol. 35, Pages 241-272

    If you want me to e-mail you a PDF let me know– chris

  11. Thanks, Chris, and yes, I would be grateful if you’d send me a copy at {removed}.

    This post and its comments are a timely gem. I’m working on a presentation to my astronomy club on orbital

    forcing. My presention is my version of what Kipp Albert is doing, creating my own set of explanations so I can defend the science.

    An example of what I’m doing is here: (requires Flash plugin for your browser)

    In this program I’m trying to illustrate the orbital changes over time. This version works like a clock (turning the

    hands a 26k, 41k, and 100k years) but for a future version I hope to track the real values to show visually what the

    orbit was doing and how the forcing affected climate.

    I enjoy your primer approach to climate science. If I’m not commenting, please assume I’m reading and studying your post along with the references you provide.

    Also, I’m redrawing your global circulation diagram for my presentation. A question: does the Hadley cell move up

    and down in lattitude according to the season? essentially tracking the overhead sun?


    John Garrett

    Response– Sent. And yes, the Hadley cell ITCZ does change position as a function of season.– chris

  12. JohnG:It is rather counterintuitive to learn form the articulate Chris and go over to AccuWeather, and teach while having rocks thrown at me.Three out of Twenty are bad odds, but better than zero. It is worth it, I think, and I am learning at the same time.Peace.

  13. Kipp: Same here. I’m learning a lot and it’s worth it. And like you, I feel a need to correct obvious errors and distortions of climate science where I see them, in my case, in local media and trade journals.

    My underlying message is NOT that people must understand and accept what I think of the science of climate change; but rather, that as non-scientists (me), we must learn to recognize what is and isn’t competent peer reviewed literature. In other words, my goal is to promote scientific literacy. That’s why I’m drawn to this site; I can always find good reference material behind Chris’s explanations. (Also, this topic got me looking at back issues of Nature as far as 2001 when I first started subscribing. I missed so much when I first started studying climate change).


  14. Chris,

    First para: “solar insulation”. You meant solar insolation, right?

  15. Chris, when discussing this general subject you should not neglect to mention that the southern westerlies are moving poleward again, and with great rapidity as these things go, this time as a result of a poleward shift of the entire atmospheric circulation sytem due to expansion of the tropics driven by radiative forcing. Wally’s beast is on the move.

  16. Let me get this straight, the 200 year melt in the Antarctic of part of the ice cap, that caused a 20 meter sea level increase that ended the last glacial period 1000 years before CO2 started rising was due to CO2?

  17. Hi Chris, I apologize for the extremely late comment. In his description of other factors involved in atmospheric CO2 concentration, do you think Anderson might have missed a possible source of CO2 exchange in terrestrial vegetation and soil carbon? He mentioned the expansion of terrestrial biosphere only as a negative feedback, “drawing CO2 out of the atmosphere.” However, during glacial advance, vegetation and soil carbon that accumulated during warm periods, were covered by advancing glaciers, and carbon from that organic material was essentially stored beneath the ice during glaciations. During deglaciation, much of this carbon stored beneath the ice (and possibly much of it in permafrost) was released increasing CO2 atmospheric concentrations. This scenario was described by Zeng, 2003 (ADVANCES IN ATMOSPHERIC SCIENCES, VOL. 20, NO. 5, 2003, PP. 677–693). Zeng suggested this, along with other ocean-based mechanisms such as the change in ocean temperature, could account for the entire changes in CO2 concentration during glacial/deglacial cycles. I wouldn’t go that far, but I think that it may have indeed been a contributory factor, along with mechanisms described by Anderson et al., 2009 and the other factors he mentions above.

    Response– I don’t think they missed these factors, it just wasn’t within the scope of their study. They never claim changes in ocean upwelling are the only thing going on– chris

  18. I suppose I conceived that part of my comment poorly. It might have been more reasonable to say, here is an aspect of the deglacial rise in the concentration of atmospheric CO2 not covered by Anderson et al. 2009, in either their article or in Anderson’s comments. I guess what I was reacting to was Anderson’s comment “In contrast to processes in the ocean, expansion of the biosphere on land following the melting of ice sheets and warming of earth’s climate represented a negative feedback, drawing CO2 out of the atmosphere.” There he seems to be saying that the only impact of the terrestrial component of CO2 concentration change was to decrease atmospheric CO2. I fully agree with that statement, as the re-establishment of a vegetative cover over previously ice covered land did contribute to an increase in the uptake of CO2. But I also believe the terrestrial contribution to the rise in CO2 concentration during deglacial times has possibly been underestimated by many researchers. The terrestrial contribution might have lagged much of the marine contribution as its impact could only have occurred as ice sheets melted, much of which occurred well past the temporal extent of the pre-Boreal warmup.

  19. really good post and an intriguing hypothesis. Also excellent to hear Dr Anderson’s responses to your questions – very informative indeed. I’ll certainly be keeping an eye out to see if this idea gains further traction in coming years.

    As an aside I’d like to thank you for your excellent posts, some of which I’ve recommended to others in learning more about CO2 and the greenhouse effect.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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