Sequestration of ocean surface water by the Gulf Stream

There’s one story that the climate science community likes to scare the world’s population with even more than The Methane Time Bomb and that is saturation of the upper ocean layers with CO2 leading to ocean acidification and the extinction of all carbonate based ecosystems. The worry is that on our static planet removal of CO2 from the upper ocean layers by the deeper ocean layers takes place on “a very long time scale” of hundreds of years.

This concept is totally at odds with my own perceptions of wild oceans and ocean currents churning seawater about on a daily basis. From my geochemistry background I have recollection of the ocean mixing time being roughly 1000 years. And this got me thinking about the Gulf Stream, the gigantic ocean conveyer system that carries water from the Indian Ocean, round Cape of Good Hope, northwards along the full length of the Atlantic until it eventually sinks in the North Atlantic somewhere between northern Norway and Greenland.

The Gulf Stream transports nearly four billion cubic feet of water per second, an amount greater than that carried by all of the world’s rivers combined.

That is one BIG number. The calculation below the fold suggests that the Gulf Stream sequesters the equivalent of the surface waters (333m layer) of the whole Atlantic Ocean once every decade. Map image from Met Office.

4^9 ft3 per second
= 2.4^11 ft3 per minute
= 1.44^13 ft3 per hour
= 3.456^14 ft3 per day
= 1.262^17 ft3 per year
= 3.6^15 m3 per year
= 3.6^6 km3 per year

I’m going to make the assumption that the annual flow of the Gulf Stream at the surface is matched by a similar reverse flow in the depths and therefore the flow rate equals the sequestration rate. Hence the Gulf Stream sequesters 3.6 million cubic kilometres of surface water every year. Is that still a BIG number? The upper ocean layer is normally taken to be the top 300 m or so, for simplicities sake I’m going to assume the top 333.3 m. 3.6^6 km3 spread out in a layer 333.3 m deep would cover an area of 3*3.6^6 = 10.8^6 square kms. The area of the Atlantic Ocean is 106.4^6 square kilometres. Hence The Gulf Stream would take 106.4/10.8 = 9.85 years to sequester the surface layer (333.3 m deep) of the Atlantic Ocean.

The total volume of seawater on Earth is 1.37^9 km3. Divide this by the annual sequestration of The Gulf Stream (3.6^6 km3) and you get 380 years for The Gulf Stream to cycle a volume of water equivalent to all the world’s oceans. Of course some of the water in the Gulf Stream may simply be going round and round but there must surely also be significant mixing along the edges and on the surface and this zeroes in on the 1000 year ocean mixing time quoted at the start. And of course The Gulf Stream is only one of many systems churning seawater on a daily basis.

These numbers surprised me, check my sums and let me know if I made a mistake.

Image source.

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

79 Responses to Sequestration of ocean surface water by the Gulf Stream

  1. Joe Public says:

    An excellent insight into oceanic (thermo) dynamics. Thanks Euan.

    • dennis coyne says:

      The thermohaline in the Atlantic only transports about15.5 (14-17) million cubic meters/sec of water, so the 113 million cubic meter per second flow of the gulf stream is an overestimate by a factor of 7.3. Only a portion of the gulfstream flow sinks to become part of the AOMC (thermohaline circulation of altlantic). There is also a contribution from the Southern Ocean of 11(10 to 12) million cubic meters/sec. For the total world oceans we have a thermohaline circulation flow rate of 26.5 million cubic meters per second about 1/4 of Euan’s estimate.

      Also the thermohaline circulation goes to depths of at least 900 meters so using 333 meters as Euan has done would tend to give estimates about 3 times too low. When we combine these two problems of flow rate(factor of 4) and depth (factor of 3), Euan’s estimates for the time for the ocean to mix are low by a factor of 12.

      There may indeed be a lot of mixing of the surface layers of the ocean, the assumption that the surface flow rate of the Gulf stream is matched by flows at depths of 300 meters or more at the same rate may be incorrect, According to encyclopedia Brittanica the flow of the Gulf stream is about 30 million cubic meters per sec and in some places the depths are as much as 790 meters.

      • Euan Mearns says:

        Dennis I understand and accept your first paragraph. But you lose me on the second. I wasn’t trying to estimate ocean mixing time but the time it would take thermo haline circulation to cycle a volume of water equivalent to the world’s oceans. If I overestimated thermohaline circulation by a factor of 4 then surely that is it. Where does depth come into it? That is relevant to the time taken to circulate a surface area of ocean but I don’t see why it should affect the estimate for global recycling.

        I have as a consequence of this epic blogging reached the conclusion that ocean circulation cannot remove CO2 from surface to depths since the return flow to surface will likely have higher C than that sequestered. In fact I am having trouble with the whole concept of the solution pump.

        • dennis coyne says:

          Hi Euan,

          You talked about the top 333 meters of ocean, if the thermohaline is about 30 SV then the time to mix the entire ocean would be about 1300 years. You need to think in terms of dissolved inorganic carbon rather than co2.

          You are correct on the factor of 4, I was confused by the 333 meters, which really should be about 700 meters for mixing by wind and waves, I would also agree that the mixing of the top 700 meters is much quicker than 1300 years, but remember that not all ocean currents move as much volume as the gulf stream, we would need average flow rates throughout the surface of the ocean, I have no idea what that would be. Note that there are varying estimates of the gulf stream flow rate as well, the lower end estimates would have the top 700 meters of the Atlantic mixed in 60 years of so, I don’t know if this would apply to the oceans of the world.

          None of these processes work in isolation, you need the solution pump, biological pump and carbonate pump working together to make sense of the system.

          • Euan Mearns says:

            Agreed that in surface ocean water it is bicarbonate that is the species of interest at pH around 8. I have currently reached the view that the main process to remove CO2 from the atmosphere is photosynthesis. I don’t see how the so called solubility pump can possibly work. And I will stand by my view that the process that removes CO2 from the atmosphere needs to be modelled separately from the processes that eventually sequester the organic matter subject to rate limitations from “sink saturation levels”.

          • dennis coyne says:

            Hi Euan,

            I am trying to understand the carbon cycle of the Earth System. Restricting one’s attention to carbon dioxide only is quite limiting in my opinion. Read the Archer papers if you want to understand the Carbon cycle.

            A model of the fast processes without an understanding of the limited buffering capacity of the Ocean (which already contains 38,000 Gt of dissolved inorganic carbon[DIC]) and the effect of calcium carbonate formation and dissolution on the DIC levels will tell us little about how anthropogenic carbon emissions are eventually sequestered by the Earth system.

          • Euan Mearns says:

            A model of the fast processes without an understanding of the limited buffering capacity of the Ocean (which already contains 38,000 Gt

            Dennis you are already laying claim to understanding by saying it is limited.

          • dennis coyne says:

            Hi Euan,

            It is not very hard to understand that the buffering capacity might be limited. As the Volume of the Ocean is not unlimited and the mixing time is not instantaneous, the likelihood that the buffering capacity might be limited seems self-evident. Perhaps there is something that I am missing, I am assuming here that you are not claiming the buffering capacity of the Ocean is unlimited over millennial time scales.

          • Euan Mearns says:

            Of course there are limits, but you don’t know what they are. And by using the terminology “buffering capacity” you are still thinking chemical species and not billions of tonnes of rotting plankton where one of the rate limiting factors might be O2 availability. I saw a chart with an anoxic zone the other day. With anoxia, the limit to how much organic material the ocean can hold would be huge. And you need to note that the Pacific is more laden than the Atlantic and other oceans. And at some point you’re going to have to come to terms with what appears to be a fact that the ocean system is driven by biology and not solution chemistry.

            Would appreciate your thoughts on the long comment at end of the thread.

  2. A C Osborn says:

    Euan, do you have any data on what Oceanic Volcanoes do in terms of bringing what was cold water, but now heated, to the surface?
    Also what affect does the slosh of Tides have on the lower depths, perhaps Clive would know something on that?
    One thing we do know is that there are many different currents in the very deeps moving the water all arounf the world. The NuSchool Earth has a nice view of a lot of them.,1.49,323

    • Euan Mearns says:

      I suspect hot water flow from mid ocean ridges is too small to significantly impact the very large body of ocean water above. Over most of the oceans tidal effects are very small in terms of vertical and lateral movements. Its only where these get amplified along coastal areas that they appear large. But I believe there ay be a host of thermohaline processes working on all scales that may cause perturbations to “static ocean layers”.

      This is a great source..

      Its this chart I really wonder about combined with the ocean conveyor map in my post. Deep pacific water contains much higher concentrations of CO2 and has very low pH compared with surface waters. According to my calculations there must be about 3.6 million km3 of this water bubbling to surface in the N Pacific each year. Its much more extreme than anything we see on the surface. Something doesn’t add up.

  3. Pete says:

    “This situation may accelerate global warming by affecting the complex oceanic food chain. Phytoplankton are tiny plants which float in the water column. Zooplankton are tiny animals which graze on the phytoplankton. These two organisms are critical to the oceanic food supply. In the Antarctic spring, as the ice recedes, tremendous blooms of plankton feed off carbon dioxide absorbed from the air by the oceans. By this process, vast amounts of CO2, an effective “greenhouse” gas, are removed by planktonic plants, thus helping to keep CO2 levels down in the atmosphere.”

  4. itzman says:


    Wild non peer reviewed systems engineering think:

    All that is required to create a complex non linear dynamic system are negative feedback paths, with delay, and non linearity.

    The earth’s cliamte must be overall within the bunds it stays within, negative feedback or we would be at absolute zero or as hit as venus.

    Overall the ultimate dominant negative feedback is of course radiative balance. Apart from residual heat, tidal action, and internal nuclear decay the earth gets all of its surface heat from the sun and loses it all to space by radiation.

    HOWEVER what is in between is far more complex. Take water. in tropical maritime climates te planet never warms too much locally, because water vapour rises, condenses, forms clouds and increases the daytime albedo. And at night warm wet air is a greenhouse effect that stops direct radiation to space by reflecting or remitting infra red. In deserts, this does not happen, and the diurnal range is far greater. which supports the general idea that water is a far more powerful greenhouse and albedo increasing agent than anything CO2 comes remotely near.

    Also ice and snow are massively powerful albedo increasers so lots of ice and snow is actually a positive feedback system. If the plant became wholly icebound, it might stay that way for a very long time. There are possibilities that this has in fact happened.

    But the planet is inside still molten, and crusts and plates shift, and whilst the surface might acquire snow and ice, its likely that the deep oceans would not, so water would still be liquid underneath, with the ice and snow actually forming a greenhouse roof on the OCEAN geothermal heat.
    As long as there is ocean circulation, that would always mean that snowball earth could after crustal changes, restart the melt by moving deep warmer water around through plate tectonics. And we know that ice ages affect plate tectonics too over long time periods. So there is another long delay negative feedback path.

    However, considering today’s climate, broadly unchanged for 8000 years, what we have is the ability of ocean currents to move warm and cold water around. With pretty long lags of at least decadal levels, or longer. And several of them too. Do these represent negative feedback? they must, unless we are simply in a 100,000 year cycle because the shorter term fluctuations show no real trend.

    So what DO we know.?

    That water is complex, it can be positive or negative feedback and both in the same phase too – water vapour. Clouds shield us from the sun, and they cool the surface with rain and snow, and they transport heat to high above (most of)the carbon dioxide. But clouds keep us warmer at night, too. And snow and ice reflect sunlight.

    Complex, very complex, and very non linear. clouds are very strange things and happen for all sorts of reasons some of which are very non linear. Like the temperature going past a given point at which point water actually condenses into droplets. if condensation nuclei are present. Think Svensmark.

    Or because of terrain. Think rain shadow and deserts.

    Or because of ocean currents bringing warm water to polar regions, or being blocked by ice melt.

    It is not too far fetched to presuppose that our climate is dominated by water flow, which is affected by continental plate tectonics on geological timescales and by amplitude modulations in ocean currents in decadal and centennial ones.

    And has almost nothing to do with CO2 man made or otherwise.

    Are these long delay feedback paths and a non liner relationship between energy and temperature die to waters latent heat and so on, enough to account for all climatic variation – at least within te last few million years? without having to invent any externalities like solar variation and Milankovitch cycles?

    My educated guess is that they are MORE than enough to account for all the global warming and cooling.

    That there is in the end no such things as a ‘normal’ climate, that the climate system is a bounded chaotic system orbiting attractors of either ice age or interglacial, and in any given attractor the only thing you can say is that one year will never be the same as another for very long.

    We may expect little ice ages, mediaeval and holocene warm periods, to come and go as various feedback paths combine to cause hotter or colder conditions in one part of the planet or the other. Or indeed overall.

    Climate change is real, and happening, but its got nothing to do with carbon dioxide, and everything to do with water. And a lot to do with geography. And plate tectonics.

    The one thing you can predict about a complex non linear dynamic system is that even if its fully deterministic, it is totally unpredictable in any detail over longer than a few elements if time scale.

    By working out the time constant of ocean transport, even to a first order, you have added just the sort of feedback mechanisms with the right sort of time constant to give multi-decadal cliamte variations of the sort usually ascribed to CO2.

    One more nail in AGW’s coffin…

    • Yvan Dutil says:

      I am always impressed by the way people with a technical training are trying to find technical excuse to reject a fact that does not suit them,

      The problem of such argument is it is pure crap until you mange to put number and cross-validate it with various observation.

      Yes, this is complex, but climate scientist have working on that for decades.

      Dunning–Kruger effect obviously.

      • Euan Mearns says:

        You need to be careful Yvan. You don’t know who it is that you have decided to insult. The convention here is to try and win arguments by presenting data and reasoned argument.

        • Yvan Dutil says:

          I have change of research field many time. Unless you have taken the time to read 150 to 250 research papers, the value of your scientific opinion is generally close to zero. This is based on 20 years of experience as a scientist.

          As for the CO2, I suggest this video (I hate to use youtube as a reference, but this is the best concentrate of information I have found).

          • Euan Mearns says:

            I watched the vid but skipped through it. Maybe watch the whole thing one day. I thought large parts of it were utter rubbish – but I may be wrong. Take rock weathering for example where half the human population seem to have been brain washed into believing this somehow controls CO2 levels in the atmosphere and acts as some form of thermostat. Here’s the Grid Arendal C cycle.


            It shows that rocks are by far the largest carbon sink with 1,000,000,000 Gt of carbon. When you weather rocks they are going to release a vast amount of CO2, especially limestones, coals, calcareous mudstones. Most silicate rock weathers to mineral grains and only a small fraction is dissolved by chemical weathering. There’s no mention of this in that lecture. Rock weathering returns a relatively tiny amount of CO2 and bicarbonate to the oceans – this is trivial compared with the ocean atmosphere exchange. So I’m afraid I just don’t get the rock weathering story at all.

            His handling of the ice core record is also kind of pathetic. Fact is that CO2 peaks and then suddenly things get colder again. The bank comes along and writes off your debts.

            PS I have read a couple of hundred papers around climate topics. The saving grace is that many of them as individual pieces of the jig saw are reasonable enough research. Its the IPCC summations of the work that are the greatest problem for most sceptics. The IPCC position does not appear to have changed much over 15 years. All that work appears not to have added anything to understanding.

          • Dennis Coyne says:

            Hi Euan,

            It is not completely clear why a geochemist would not understand the rock weathering story.

            Not to insult your intelligence, but I am not either a geochemist or geophysicist so I found the following link helpful, perhaps you could explain what the author is missing?


          • Euan Mearns says:

            The weathering of rocks is estimated to involve the drawdown of about a gigaton of atmospheric carbon dioxide a year

            That is 0.27 Gt of Carbon where the ocean and plant fast sinks combined do over 200 Gt C. Its not a question of not understanding it, its a question of understanding that its totally irrelevant. The so called weathering sink is in fact ocean water. And there is no guarantee that the CO2 HCO3 in solution in river water ever makes it to the sea. What’s more this is a simple part of the global flux of CO2 and will not respond like the oceans and biosphere have done to absorb emissions because a) it is far too small and b) it operates far too slowly – a and b are in fact the same thing. Carbonic acid made from CO2 assists rock weathering and is converted back to CO2 where it ends up in the oceans via river water instead of taking the direct route known as the solution pump.

            And he uses the white cliffs of Dover as an example of sequestered CO2 when in fact he should be pointing out that the cliffs are there in the first place because vast amounts of that limestone have been eroded and weathered and probably added billions of tonnes of CO2 back to the atmosphere.

          • Euan Mearns says:

            I guess I need a correction here. Skeptical Science claims “drawdown” when in fact it is simply a part of the annual flux. The 200Gt I refer to is the annual flux.

          • dennis coyne says:

            Hi Euan,

            I agree on further reading that the rock weathering is not relevant on short time scales. The exchange of CO2 with the ocean and how it is sequestered seems that it should be pretty easy for a chemist to understand.


    • Euan Mearns says:

      Itzman is at least in part correct. I’ve shown this chart many times before from Bond et al 2001 (Science). What is shows are cycles in the detritus dumbed by drift ice in the N Atlantic. Bond et al hypothesise that the Labrador current periodically cuts across the N Atlantic truncating the Gulf Stream, chopping off that northern loop and this brings extreme cold winters to N Europe and drift ice gets much further S than at present.

      Climate science imagines that it has a dynamic coupled ocean atmosphere model while in fact to a geologist it is a static model. The patterns of ocean currents do change with time and this does affect climate or perhaps climate changes affect the currents since Bond’s data correlates with 10Be. Climate models can never be correct and have any predictive power until they encompass changes in ocean currents with time (acceleration / deceleration / geometry). I can hear a chorus saying I don’t know what I’m talking about, that chorus that confuses global average surface temperature with climate.

      But in any case, that chorus often just forgets that the principle mechanism for heat loss from the surface is convection and not radiation. Anything that changes the mean convection rate of Earth’s climatic systems will also change the radiative heat balance since deep convection transfers heat from surface to tropopause where it can more easily escape to space. The bit about convection is lifted from Haughton’s book. I also imagine that truncating the Gulf Stream will reduce the amount of water vapour at that high latitude, also impacting radiative warming

      Prof Dave Rutledge from CalTech has made the same point that you do about systems with only positive feed backs going to an extreme point and staying there. I suspect that d convection is one of the main negative feedbacks.

      • Yvan Dutil says:

        Well, climate model too model change in oceanic currents.

      • Yvan Dutil says:

        Did you ever notice that convection is include in the atmospheric heat transfer model?

        • Euan Mearns says:

          I’m quite sure that convection is included. But what I’m talking about are changes in mean convection rates with time.

          • Dennis Coyne says:

            Hi Euan,

            The physics of the models are based on conservation of energy and conservation of mass and so the changes in mean convection rates are built into the model.

          • dennis coyne says:

            On Ocean currents, physics is also used, though a problem here is that the temperature profile (in detail) of the ocean is largely unknown, so this is trickier.

            Note that if you follow mainstream climate scientists such as the people at Real Climate, they are not convinced of the methane bomb stuff, they would consider that alarmist. They are also very open about the areas of further research which are needed, and ocean temperatures are among several areas where better data is needed.

            Nobody in climate science thinks that there are not a lot of unknowns. Real Climate is a great resource.


            And Paul Pukite also does some interesting stuff (though he is not a climate scientist), nor am I.


      • dennis coyne says:

        Hi Euan,

        Many climate scientists are geophysicists, so I think they are somewhat familiar with geology.

  5. Lidia17 says:

    The zone where phytoplankton live is really only in the top 100m or so. Be aware that mixing is not going to happen consistently at each level, since at the boundary between two substances (air, water) the physics is going to be different. Also surface currents affect the top 100m more than the deeper currents do, it seems.

    • Euan Mearns says:

      Lidia, This is one of the exact points I made in my earlier post on the Bern Model. Obviously the air water surface process is physically different to the water – water diffusion and mechanical mixing processes and I question the wisdom of trying to combine these into a single equation.

      I assert that the fast air-water and air-plant processes remove 100% of CO2 emissions from air and that the slower water – water and bio-detritus burial processes (in deltas) remove 100% of emissions from the fast sinks allowing the fast sinks to go on doing their work.

      • dennis coyne says:

        Hi Euan,

        Here is the problem with separating the fast and slow processes.

        There is a limit to how much carbon the fast processes can sequester, once reach that limit there is an equilibrium between the rate that carbon is added to the atmosphere and removed from the atmosphere so that the net sequestration of carbon by fast processes would be zero.

        Let’s ignore the land or assume that at a given level atmospheric CO2 that the fast land sequestration processes are in equilibrium and focus on the ocean. Let’s also assume for the moment that only the top 333 meters of the ocean are important for sequestering carbon (this is not correct, but it simplifies the argument).

        If we are only going to worry about how fast we can mix the upper 300 meters of the ocean, then our concern would be how much net carbon can be sequestered by this slab of the ocean, we can easily imagine that at some level of atmospheric carbon dioxide that the net rate that carbon can be sequestered will reach an equilibrium where the net sequestration of carbon will be zero.

        When that point is reached any increase in the level of carbon dioxide in the atmosphere will only be sequestered in full if the increase can be fully sequestered by the fast process.

        I will make up a numerical example (but it will not be physically accurate, it is for illustration).

        Imagine that the world decided that potential negative effects of increased carbon dioxide was a problem and magically reduced carbon emissions to some level that allowed CO2 levels to stabilize at 400 ppm, lets say 20 years from now CO2 emissions at 15 Gt per year is the level that works to accomplish this (it is highly unlikely that this number is correct, we could call the level x, it is unknown). Note also that it may not be possible to stabilize the earth system in a simple way because the earth system is dynamic and may change over time both from natural processes and the higher atmospheric carbon levels which may induce positive feedbacks to the system, but we will ignore these for simplicity,

        At this new equilibrium all of the processes balance so that 100% of carbon emissions are sequestered and atmospheric carbon levels remain unchanged.

        What happens if the world decides, no problem, let’s increase carbon emissions, the earth can take care of it. Anthropogenic carbon dioxide emissions are increased by 1 Gt/year to see if the fast sequestering processes can handle it. I would argue that they cannot handle it in the short term. At equilibrium the fast sequestering processes are working as fast as they can (they are at their capacity), the same will be true of all of the links in the chain down to the slowest of the processes, which act on time scales of 30,000 years or so.

        Does this mean that the atmospheric CO2 will increase by 1 Gt and none of this “excess” carbon dioxide is sequestered? No.

        The increase in atmospheric CO2 levels will enable some of the CO2 to be sequestered, but not all of it. This is due to increased fertilization of plant growth and increased uptake of co2 by the ocean surface due to the higher partial pressure of co2 in the atmosphere, these will take up a portion of the extra CO2 over time, but about 20% will remain in the atmosphere for very long periods (30,000 years or more) of the other 0.8 Gt about 0.4 Gt will be sequestered over a relatively short time frame (10 years or so) and the rest over longer time frames up to about 300 years.

        • Euan Mearns says:

          but about 20% will remain in the atmosphere for very long periods (30,000 years or more)

          Dennis I agree that saturation of the fast sinks is an issue not to be ignored. Do you have any information on the CO2 HCO3 content of the upper ocean with time.

          You seem to be agreeing with me whilst simultaneously disagreeing with me on a number of points. For example you agree that ocean mixing time is about 1300 years and yet say that about 20% of CO2 will remain in atmosphere for 30,000 years.

          A major problem of course is that the deeper ocean has much higher CO2 than the surface and so I’m not sure how the exchange processes of the Grid Arendal C cycle are supposed to work at all? Any takers?

          The main point of this post is to highlight the approximate rate of ocean water recycling by the oceanic thermo haline circulation and to wonder out loud why I’ve never heard about this as a possible process for removal of CO2 in surface waters into the deeps.

          • Euan Mearns says:

            In fact, if you look at the ocean part of the C cycle we do see ocean mixing processes transferring CO2 from deeper to shallower levels! We know that 50% of emissions have been sequestered – where have they gone?

          • dennis coyne says:

            Where has the carbon been sequestered? Some in trees, some locked up in sea shells and carbonate ions.

            The movement of CO2 to the depths takes much longer than you estimated in your post, where your flow rate for the THC was too high by a factor of 4 and your depth of ocean considered was too low by a factor of 3 so that the rate of turn over is too low by an order of magnitude. There will be some mixing of CO2 into the top few hundred meters by wind and waves, but again there are limits to the solubility of CO2 in the surface waters and it takes time for the thermohaline to mix the surface with deeper waters, if we take the top 1000 meters it would be on the order of 350 years.

            There are two main ocean mechanisms, the solubility pump and the biological pump which sequester carbon over the longer term in the ocean.



            Note that there are many different uses of the terms fast and slow for carbon sequestering.

            Chapter 6 with a focus on the carbon cycle in that chapter (pp470-473) is a fairly up to date mainstream science view on this see link below:


            You should actually read more of the scientific literature so that you know what you are criticizing.

            The whole weathering thing is not really relevant. Most of the sequestered carbon ends up as dissolved inorganic carbon where the sink is 38,000 Gt of carbon.

            Over a 10,000 year period an initial pulse of carbon emitted into the atmosphere at say a 1500 Gt level (where a pulse might occur over 200 years from 1850 to 2050) would fall to about 20% at the end of the period (200 Gt) for total atmospheric carbon of 789 Gt which is about 372 ppm. Generally given the uncertainty in our understanding it is considered safer to limit anthropogenic emissions to 900 Gt (of which 500 Gt have already been emitted) to try to keep warming under 2 C.

            Just as an engineer would over build a bridge by a factor of 2 or three just to be safe, keeping carbon emissions on the lower side would be better in case the higher end (ECS=4.5 C) of climate sensitivity estimates turn out to be correct, rather than the lower estimates (ECS=1.5C) favored by some.

            If the low estimates are correct, we can always burn the fuel later, it is much more expensive fix the problem later, if higher climate sensitivity is correct.

          • dennis coyne says:

            Hi Euan,

            In fact, if you look at the ocean part of the C cycle we do see ocean mixing processes transferring CO2 from deeper to shallower levels!

            The thermohaline circulation (THC) will mix the surface with the lower levels (down to 1000 m) on the order of 300 years, so the outgassing of CO2 as the deeper waters move to the surface takes quite a bit of time. Most of the sequestered CO2 is dissolved inorganic carbon circulating in the intermediate and deep ocean.

  6. What do you mean by “sequestered”?

    The flow rate is easy enough. It is the volume of water flowing through a fixed cross section in a unit time. So far so good.

    But to “sequester” it means moving the water from the surface to the deep. That means moving any given ft^3 of water over 20,000 ft vertically. I think it is this lateral, or vertical, displacement that you are missing.

    Flow rate != sequestration rate.
    moving 20,000 ft^3 one foot deeper is not the same thing as moving 1 ft^3 20,000 ft deeper.

    • Euan Mearns says:

      Stephen, good luck with moving all that water one foot down every second, I think you’ll find that after 1 milli second it has nowhere to go. The Gulf Steam sinks to form North Atlantic Deep Water and is an integral part of the engine of global thermohaline circulation.

      I saw a documentary a few months back on the sinking of HMS Hood by the Bismark. Britain’s biggest against Germany’s biggest warships. The Hood was sunk in the Denmark Straight between Iceland and Greenland. The wreck has been found and surveyed by ROVs (Remotely operated vehicles – mini subs). The footage was astonishing, the ROV’s struggled to hold station the ocean floor currents were so strong. It was like watching the Niagara river in the Niagara gorge – which is more spectacular than the falls.

      • Dennis Coyne says:

        Hi Euan,

        You are confusing the gulf stream and ocean conveyor belt (thermohaline circulation). You used the NOAA website as your source for the gulf stream estimate, from the same website we have:

        “Winds drive ocean currents in the upper 100 meters of the ocean’s surface. However, ocean currents also flow thousands of meters below the surface. These deep-ocean currents are driven by differences in the water’s density, which is controlled by temperature (thermo) and salinity (haline). This process is known as thermohaline circulation.”

        These thermohaline currents go to about 800 meters in depth so using 333 meters is not a good way to estimate, probably figuring the top 1000 meters would be better.

        Note that the flow rate of the ocean conveyor belt is about 20 million (10^6) cubic meters per second (according to source below, which also suggests a mixing time of 600 years rather than my earlier 1300 year estimate (based on another source which suggested the total conveyer belt flow was about 30 million m^3/sec). There may be some wind driven mixing from the Gulf stream that figures into this, the problem is complex.

        Good paper on thermohaline below(but I am no expert and this is an old paper(2001) which may not be correct based on latest research, it suggests a 600 year ocean mixing time)

        • Euan Mearns says:

          You are confusing the gulf stream and ocean conveyor belt (thermohaline circulation). You used the NOAA website as your source for the gulf stream estimate, from the same website we have:<

          I don't think so Dennis. My understanding is that the Gulf Stream is that part of the global thermohaline circulation that stretches from the GOM to W Europe. It carries warm saline water northwards where on going evaporation causes the salinity and density to increase as the water cools. Eventually it becomes too dense and sinks.

          Good luck with evaporating currents hundreds of meters below the surface which I think is probably referring to the return flow.

          • dennis coyne says:

            Hi Euan,

            Your understanding is incorrect, what you are describing is thermohaline circulation.


            The Gulf stream is a wind driven current that is a piece of the Atlantic portion (AMOC), but the figures in your post are based on this surface current which is larger than the thermohaline circulation by a factor of 4.


            “The Gulf Stream proper is a western-intensified current, driven largely by wind stress.[7] The North Atlantic Drift, in contrast, is largely thermohaline circulation–driven.”

            There is a distinction between the Gulf stream and the North Atlantic Drift.

            The main point is that the turnover of the top 1000 meters of the ocean is about 325 years and the top 333 meters might be quicker, but the smaller the piece of Ocean we talk about (about 9% of total Ocean volume in the case of 333 meters) the more quickly we run into alkalinity limits that would limit the carbon that can be sequestered.

          • Euan Mearns says:

            Dennis, I have blogged with you for many years, and you were as far as I recall always one of the saner voices out there. But I’m afraid your contributions to this thread appear to have descended into gibberish. You challenge me, I respond and then you agree with what I say but then continue to disagree.

            I’m not going to accept wikipedia sources as a distinction between the Gulf Steam and global thermohaline circulation. But that is not to say that I don’t accept this may be correct.

            According to your sources only 25% of the “Gulfstream” descends to form North Atlantic Deep Water. And so where does the rest of this water go?

            You seem to have two ocean currents crossing the Atlantic on the surface, travelling together, european black eel youngsters hatched in the Sargasso 2 inches long swimming along with it, but with totally different origins. When we have winds from the East, the gulf stream still flows.

          • dennis coyne says:

            Hi Euan,

            I apologize for not being clear. I am learning here as I go. As I am not an expert, I am finding sources of information on the internet. Do you accept as a possibility that the Gulf Stream might not all become a part of the thermohaline circulation. I have found several sources which estimate this at about 15 to 18 SV(million cubic meters per second) in the North Atlantic and about 10 to 15 SV in the Southern Ocean.
            If we accept the NOAA estimate of the Gulf Stream flow at 110 SV (Encyclopedia Britannica says it is 30 SV), then some of the Gulf stream flow remains at the surface (either 95 SV or 15 SV depending on which estimate is correct).

            So far the only thing we roughly agree on is that the mixing time of the Ocean is about 1000 years (I would put it at more like 1250 years).

            You use 333 meters to determine mixing of the surface of the Atlantic, if the 110 SV Gulf Stream estimate is correct you get this roughly correct( I would use 95 SV because some sinks and flows along the bottom of the ocean at 3500 meters). The usual depth people use for the surface is the top 700 meters so that would double your mixing time for the surface of the Atlantic.

            Paper on thermohaline circulation:


            For an understanding of the carbon cycle:



            The first few pages of the last paper explain why a simple exponential decay is not appropriate to describe the response of an injection of 500 Gt of carbon dioxide into the atmosphere.

  7. dennis coyne says:

    Hi Euan,

    Let’s assume your geochemistry professors knew what they were talking about and that the mixing time of the oceans is about 1000 years. I checked this, and it seems a good estimate. The ocean volume is 1.335 billion (10^9) cubic kilometers(km3) and has an average depth of 3.7 km. A 1000 year mixing time implies that the ocean conveyer belt flows at about 1.335 million (10^6) km3 per year or about 40 million cubic meters(m3) per second or 1.4 billion cu ft per second. So your estimate of the thermohaline circulation is high by a factor of 3. Normally we think about the top 1000 meters of the ocean in terms of dissolved inorganic carbon so if we restrict our thinking to this portion of the ocean (and make the simplifying assumption that there is no mixing between the top 1000 meters and the 2700 meters (on average) below this we would have a mixing time of 250 years and if we restrict ourselves to the upper third of this and assume again no mixing with waters below 333 meters, then we have about 83 years.

    Note that the actual ocean mixing time is a little more than 1000 years, more like 1300 years, so we would get 350 years for the top 1000 m and 117 years for the top 333 meters.

    Also the Gulf stream is wind and weather driven, it is different from the ocean thermohaline circulation.

  8. Euan Mearns says:

    Dennis, I’m sorry for getting a bit short yesterday evening but it does seem you are choosing to disagree with what I’m saying and then present alternatives that are not substantially different. My post is not intended to be a precise rendition of what is actually going on but rather an illustration of the potential scale of the speed of certain processes.

    I didn’t pluck 300 m out of thin air. It came from this NASA source (there are a lot of useful links here):

    The way the Grid Arendal carbon cycle is presented is in fact deceptive. Exchange surface with deep water is presented as a slow process, over 100 years. In fact it proceeds at exactly the same pace as the so called fast atmosphere – ocean surface process. Both exchange roughly 90 Gt C per year!

    So where does this leave the fast sink getting saturated argument? Part of the problem is that the intermediate and deep processes are bringing more acid and CO2 rich water towards surface, they are not sequestering. Perhaps this is all understood by the C cycle experts but it sure as hell does not come over that way in the on-line debate. Deeper water cannot remove CO2 from surface water by either diffusion or mechanical mixing since the deep water has more CO2 in it. That leaves the biological pump and sequestration by the Gulf Stream (or variants thereof). One problem with the thermohaline circulation is that it seems likely to me that it will degas more CO2 in the Pacific than it can sequester in the N Atlantic since it will likely bring CO2 laden water from the deep to surface?

    I’m quite happy to discuss the flow rate of the global thermohaline circulation and you mentioned alkalinity without providing any data or links.

    This seems to be a good source for explaining the subtleties of different currents in the N Atlantic.

    The Atlantic Meridional Overturning Circulation (AMOC) is a major current in the Atlantic Ocean, characterized by a northward flow of warm, salty water in the upper layers of the Atlantic, and a southward flow of colder water in the deep Atlantic. The AMOC is an important component of the Earth’s climate system.

    The Gulf Stream is a component part of AMOC or vice versa.

      • dennis coyne says:

        Agreed, it is a very interesting paper. They point out quite a bit of uncertainty, unfortunately it would take quite a bit of study to sort out the disagreements. One nice thing about IPPC AR5 is they focus on those areas where there is agreement and where there is substantial disagreement, they simply say there are differences in opinion and thus greater uncertainty.

        I like all three of your sources linked above, thanks. That last one was very heavy reading.

        • Euan Mearns says:

          This is where our scientific approaches diverge. I see vast areas with poor to zero understanding and believe it is impossible to have a model that gets remotely close to reality for so long as those blanks exist. That is why the IPCC keeps getting it wrong. But they are exploring every variant of why they are right before they gat around to accepting they are simply wrong. And it therefore becomes a travesty when the climate science community begins to make claims about near unanimity for their findings. Of course the main material change in AR5 was to extend the range of climate sensitivity downwards. The range is so large is basically admits that the IPCC doesn’t have a clue what is going on. And everyone agrees with that – I’m not being flippant here, those are the exact circumstances.

          • dennis coyne says:

            Hi Euan,

            Disagree completely, even though the models are not perfect and there are large areas that need better data and models. I think you give climate scientists much too little credit.

    • Euan Mearns says:

      The biogenic rain to the deep sea has important mineral components: calcium carbonate, mostly from coccolithophorids (phytoplankton) and for- aminifera (zooplankton), and opal, mostly from diatoms (phytoplankton) and radiolaria (zoo- plankton). The calcium carbonate (CaCO3) com- ponent is important for atmospheric carbon dioxide in its own right (Figure 2). In contrast to organic carbon, the production, sinking, and burial of CaCO3 acts to raise atmospheric carbon dioxide concentrations. This is unintuitive, in that CaCO3, like organic carbon, represents a repository for inorganic carbon and a vector for the removal of this carbon from the surface ocean and atmos- phere. The difference involves ocean “alkalinity.” Alkalinity is the acid-titrating capacity of the ocean. As it increases, the pH of seawater rises (i.e., the concentration of Hþ, or protons, decreases), and an increasing amount of CO2 is stored in the ionic forms bicarbonate (HCO23 ) and carbonate (CO232). This “storage” is achieved by the loss of Hþ from carbonic acid (H2CO3), which is itself formed by the combination of dissolved CO2 with its host H2O. Thus, an increase in ocean alkalinity lowers the pCO2 of surface waters and thus the CO2 concentration of the overlying atmosphere. The carbonate ion holds two equiva- lents of alkalinity for every mole of carbon: CO232 must absorb two protons before it can leave the ocean as CO2 gas. In the precipitation of CaCO3, CO232 is removed from the ocean, lowering the alkalinity of the ocean water and thus raising its pCO2. We can describe this in terms of a chain of reactions: the CO232 that was lost to precipitation is replaced by the deprotonation of a HCO23 , generating a proton. This proton then combines with another HCO23 to produce H2CO3, which dissociates to form CO2 and H2O, yielding the summary reaction: Ca2þ þ 2HCO23 ! CaCO3 þ CO2 þ H2O. Thus, the biological precipitation of CaCO3 raises the pCO2 of the water in which it occurs.

      The biological formation of CaCO3 affects atmospheric CO2 from two perspectives. First, the precipitation of CaCO3 in surface waters and its sinking to the seafloor drives a surface-to-deep alkalinity gradient, which raises the pCO2 of surface waters in a way that is analogous to the pCO2 decrease due to the biological pump; this might be referred to as the “carbonate pump.” Just as with the CO2 produced in deep water by organic matter degradation, the chemical products of the deep redissolution of the CaCO3 rain are even- tually mixed up to the surface again, undoing the effect of their temporary (,1,000 yr) sequestra- tion in the abyss. Second, ,25% of the CaCO3 escapes dissolution and is buried, thus sequester- ing carbon and alkalinity in the geosphere, on a timescale of thousands to millions of years (Figure 2). An excess in the loss of alkalinity by calcium carbonate burial rate relative to the input of alkalinity by continental weathering will drive an increase in the pCO2 of the whole ocean on the timescale of thousands of years. In summary, CaCO3 precipitation can alter atmospheric pCO2 by generating a surface-to-deep gradient in seawater alkalinity (the carbonate pump) and by changing the total amount of alkalinity in the ocean.

      The CaCO3 cycle is a central part of the effect of biological productivity on atmospheric CO2. However, it is not within the strict definition of the biological pump, which deals specifically with organic carbon. Moreover, the effect of the CaCO3 rain is determined not only by the actual magnitude of the rain to the seafloor but also by its degree of preservation and burial at the seafloor, a relatively involved subject that is treated elsewhere in this volume. Thus, in our discussions below, we try as much as possible to limit ourselves to the geochemical effects of the biogenic rain of organic matter, bringing CaCO3 into the discussion only when absolutely necessary and then trying to focus on its biological production and not its seafloor preservation.

    • dennis coyne says:

      Hi Euan,

      I was using the AMOC incorrectly in my earlier comments. This is often confused with the thermohaline circulation(THC) but it is in fact different. See

      There are areas where I agree with you and others where I do not. I agree that on the scale of less than 10,000 years the rock weathering story is of little relevance.

      I disagree with your assessment that the fast processes that sequester carbon in the surface of the ocean can exchange the dissolved inorganic carbon (DIC) at the same rate with the deeper ocean, these processes occur on the order of 300 to 1000 years. There is a limit to these processes as well which depend on removal by calcium carbonate formation which is slower still ( the order of 10,000 to 50,000 years).

      This is the reason we cannot use a simple exponential decay for atmospheric carbon dioxide for a proper model.



      • Euan Mearns says:

        Dennis, I find this exchange very useful to iterate towards some form of understanding.

        I disagree with your assessment that the fast processes that sequester carbon in the surface of the ocean can exchange the dissolved inorganic carbon (DIC) at the same rate with the deeper ocean, these processes occur on the order of 300 to 1000 years.

        I’m writing another post that includes this, not because it is popular but because it is important to get to the bottom 😉

        This is the reason we cannot use a simple exponential decay for atmospheric carbon dioxide for a proper model.

        The 2 Tau model I favour is not exponential but is one where 20% of emissions linger.

        I will try to read your links tomorrow.


        • dennis coyne says:

          Hi Euan,

          I enjoy the exchange as well and I am learning a lot. Thanks.

          I apologize for the occasional cheap shots, I will try to do better in the future. I will also attempt to read your links as well.

        • dennis coyne says:

          Hi Euan,

          The two tau model is pretty close to an exponential, but the various processes we are discussing is the reason why we would need at least 3 taus, one for CaCO3, one for DIC, and one for fast processes. That is based on my current level of understanding, which will hopefully improve.

    • dennis coyne says:

      I agree that both links above are excellent (they agree with other sources I have read) Thanks.

  9. Euan Mearns says:

    Dennis, this is one section from larger post I’m working on for your consideration and comment. I really don’t think the solubility pump can work, but what does happen is that CO2 enters the ocean surface layer where it is consumed in vast quantities by phytoplankton. These are removed by gravity which is a ver fast process;

    The Representation of the Ocean Biological Pump

    The representation of the ocean biological pump [ref] shown in Figure x bears important information. It shows 3 Gt organic carbon in the surface ocean and this tallies with the number in the Grid Arendal cycle (Figure x). But it also shows net primary production of 50 Gt per year and net export of 10 Gt C per year. The Tau (half life) for surface ocean C is only 0.06 years – 22 days! So this is not the Amazon forest, growing slowly over centuries but an extremely rapid turn over of microscopic plankton.

    Figure x The ocean  biological pump from Sigman and Haug [ref]. According to this scheme the surface oceans never contain more than 3 Gt organic carbon and yet they export 10 GtC per year brought about by the extremely fast cycle of growth, death and removal of dead organisms mainly by gravity. This scheme also 

    There are two main types of plankton in the oceans that form the base of the food chain. Phytoplankton contain chlorophyl and fix CO2 from the upper ocean layer via photsynthesis and they are considered to be plants. Zoo plankton eat the phytoplankton. Ocean algal blooms are one manifestation of phytoplankton growth. The deeper ocean is subject to a continuous rain of this dead organic material that locks atmospheric CO2 into microscopic creatures that die and sink evidently transferring 10 GtC from upper to deeper ocean layers each year.

    There is in fact major uncertainty over the size of this flux. This UK government source [NERC ref x] saying:

    The biological carbon pump is a major term in the global carbon cycle, transferring approximately 5-15 GT C yr-1 from the surface ocean to the oceans interior (Henson et al., 2011). It is of comparable magnitude to the annual increase in CO2 in the atmosphere driven by anthropogenic remobilisation of fossil fuel reserves and without it we believe that atmospheric CO2 would be order 200ppm higher (Parekh et al., 2006). Small changes in its functioning and or strength could radically affect ocean atmosphere partitioning of CO2.

    Deep ocean carbon chemistry appears to be dominated by the biological pump with 38,000 GtC (that is 38 trillion tonnes) of rotting plankton in the ocean depths (Figure x). On average about 10 GtC being added each year. I find the representation of this process on the Grid Arendal Model (Figure x) to be woefully inadequate. First, representing marine organisms by 3 GtC misses completely the annual productivity of 50 GtC per year. Second, the thin green arrow removing 4 Gt per annum as a very slow process over 100 years seems to miss completely current understanding that 5 to 15 Gt of dead marine organic carbon sinks from shallow to deep levels every year [ref x]. Grid Arendal have a second thin green arrow removing 6 Gt per annum of dissolved organic carbon by very slow process, which I’m guessing makes up the 10 GtC which everyone seems to agree is a good median estimate. And so Grid Arendal do have a net 10 GtC per annum transfer of C from surface to deep layers but they categorise this annual flux as very slow. Explanation required please!

  10. dennis coyne says:

    Hi Euan,

    I am not familiar with the Grid Arendal Model, but a search on Grid Arendal gives a website with presentations for the lay person to understand environmental issues. These sorts of presentations are often inadequate and in many cases are prepared by people with an incomplete understanding of the science. Think of it as a secondary school teacher trying to present quantum mechanics to an average student, usually a little gets lost in the presentation.

    Often these fluxes are presented in net terms.

    A point of some confusion are the fast and slow processes and how they are divided. In climate science the slow processes are the rock weathering, etc:

    “A second, slow domain consists of the huge carbon stores in rocks and sediments which exchange carbon with the fast domain through volcanic emissions of CO2, chemical
    weathering (see Glossary), erosion and sediment formation on the sea floor (Sundquist, 1986). Turnover times of the (mainly geological) reservoirs of the slow domain are 10,000 years or longer.”

    From page 470 of the IPCC AR5 WG1 (Chapter 6).

    The “fast processes” are all the other processes that sequester carbon and have turnover times between a few years (atmosphere) to millennia (deep ocean).

    “Deep ocean carbon chemistry appears to be dominated by the biological pump with 38,000 Gt (that is 38 trillion tonnes) of rotting plankton in the ocean depths ”

    The part about rotting plankton is not quite right, about 700 Gt is dissolved organic carbon, most of the 38,000 Gt is dissolved inorganic carbon, if you think of dissolved carbonate as rotting plankton then it would be correct, I may just think in different terms.

    On the annual fluxes and slow change, I think the slow change refers to the slow change in the size of the sink. About 13 Gt is exported from dead marine biota to the intermediate and deep ocean annually (in both organic(2 Gt) and inorganic(11 Gt) forms. The size of the carbon sink in the intermediate and deep ocean is about 37,000 Gt and it imports 13 Gt and exports 11 Gt each year (to and from the surface ocean) and about 0.2 Gt goes to sediments. Since 1750 the intermediate and deep ocean carbon sink has increased by 155+/-30 Gt and for the past decade has been increasing at about 2 Gt per year, as 0.005% per year is a pretty mall number, some would consider this a slow change 🙂 See figure 6.1 on page 471 (Chapter 6 of IPCC AR5 WG1).
    I also think “slow process” may refer to the slow vertical movement of ocean water below 700 meters.

    I would recommend Chapter 6 of the IPPC AR5 WG1 as a better source than the Grid Arendal Model see

    • Euan Mearns says:

      Dennis, thanks for link to AR5, I found this. It actually addresses many of the concerns I have. However, please note they are exporting 2 GtC as DIC and 11 Gt directly as marine biota. Phytoplankton are plants and not CaCo3 shelled beasties (at least I think they are).

      • dennis coyne says:

        I think we are on the same page see my comment below. I think of it as marine biota to DOC (dissolved organic carbon) and 2 Gt of DOC to deep and intermediate sink and 11 Gt of DIC to deep and intermediate ocean sink, but that is just a terminology difference of little importance.
        In the end we probably will not agree fully, but we both may learn something in the process. I have learned much so far. Thank you.

        • dennis Coyne says:

          And looking closely at the chart once again, I am not really sure, I get the impression that most of the 37,000 Gt of carbon is in the form of dissolved inorganic carbon and a smaller pool is dissolved organic carbon based on what I have read, but I really am a novice on this stuff I have never studied oceanography.

        • Euan Mearns says:

          Hi Dennis, I too have learned a huge amount and so the appreciation is mutual! Its just that I’m coming from the angle “what have they done wrong?” – and there’s no harm in that! This carbon cycle model may appear simple at first glance but it is spectacular. One thing I’d note is that volcanism and rock weathering balance at 0.1 Gt / y – which is totally insignificant for the debate about where have 50% of emissions gone?

          I too came across the carbonate deposition resulting in CO2 increase a couple of days ago and decided to body swerve that part of the story. Carbonate is not stable in any of our deep oceans today. Where it is being formed on shelf seas may or may not be important – there is only a small fraction of geology gets preserved.

          One issue that probably separates us most right now is the longevity of dCO2 in the atmosphere. I just don’t know what the evidence is that this is going to linger for hundreds to a thousand years. I understand phil Chapman’s model. But as far as I can tell, the photosynthesis pumps will continue to pump until a new equilibrium is restored if emissions are switched off. The accelerated rate will decline with time as PCO2 decreases.

          • dennis Coyne says:

            Hi Euan,

            Yes that is a crucial difference. I agree on the Rock weathering, the key idea that you may be missing is that there is a limit to how much carbon can be removed from the atmosphere by biological activity. The evidence is the relatively stable equilibrium established from 10,000 years BP to 1750 AD. Why is it that the biological activity didn’t draw down CO2 in the atmosphere to very low levels before 1750? It is these same equilibrium processes that control how quickly carbon dioxide is removed from the atmosphere at present. It is very fairly clear to me what is going on based on the Archer papers (in 2005 and 2009, there is another from 2008 which I have not read yet), if you have not yet read those I would start with the 2005 paper and maybe try to find the 2008 paper with Archer and Brovkin and then the 2009 paper). As a geochemist I think you will find them straight forward (no doubt you have forgotten more chemistry than I have learned).

    • dennis coyne says:

      You are correct about the solubility pump, it does not seem to be of great importance. Looking at the AR5 Chapter 6 figure 1, the biological pump is responsible for most of the movement of carbon from the surface ocean to deeper waters, the marine carbonate pump acts in the opposite direction with regard to CO2 tending to increase dissolved CO2 (carbonic acid) as calcium carbonate forms in producing calcareous shells.

      “Carbon is transported within the ocean by three mechanisms (Figure
      6.1): (1) the ‘solubility pump’ (see Glossary), (2) the ‘biological pump’
      (see Glossary), and (3) the ‘marine carbonate pump’ that is generated
      by the formation of calcareous shells of certain oceanic microorganisms
      in the surface ocean, which, after sinking to depth, are re-mineralized
      back into DIC and calcium ions. The marine carbonate pump operates
      counter to the marine biological soft-tissue pump with respect to its
      effect on CO2: in the formation of calcareous shells, two bicarbonate
      ions are split into one carbonate and one dissolved CO2 molecules,
      which increases the partial CO2 pressure in surface waters (driving a
      release of CO2 to the atmosphere).” page 472 IPCC AR5 WG1 Chapter 6

      Box 6.1 on pp 472-3 is also of interest, about 20% of a 1000 Gt pulse (injected to the atmosphere over 300 years) will remain in the atmosphere after 1000 years (orange line in panel c of figure 1 in box 6.1). This would correspond with about 376 ppm of CO2 about 1000 years from now if the midpoint of the pulse is 2013 and we limit total carbon emissions (including land use change) to only 1000 Gt.

      At the link below is a chart with a 3 Tau model reflecting very slow to fast processes an emissions scenario with 1090 Gt of C from all sources including land use change (all in the form of CO2 other global warming gases are ignored. Emissions stop in 2112 and atmospheric CO2 over 2190 Gt is shown (left scale has 2190 subtracted from atmospheric CO2 levels). CO2 reaches a maximum in 2093 at 489 ppm and falls to 474 ppm in 2150.

      • dennis coyne says:

        I double checked and 1090 Gt of carbon emissions would correspond with 385 ppm CO2 1000 years out, so in 3013 (assuming 2013 is the center of the 1090 Gt pulse) we would expect 385 ppm. My model was extended out to 3013 and the result is 389 ppm for the three Tau model in 3013, so the model follows the Archer et al results fairly well.

        The carbon dioxide persists in the atmosphere for a long time, after 1000 years with 1090 Gt of carbon emissions we finally get back to 2006 CO2 levels if carbon emissions remain zero from 2112 to 3013. This ignores all other gases like methane, etc.

        It is not at all clear that emissions will be limited to 1100 Gt of carbon, 1200 to 1300 Gt is a more realistic estimate and as coal prices rise it could be 1500 Gt or more.

  11. dennis coyne says:

    A key quote from the 2005 Archer paper(p4 of 6):

    “At first glance, the high atmospheric fraction may seem
    puzzling. The ocean contains 50 times more dissolved
    inorganic carbon than does the atmosphere, so it might seem
    as though pretty much all of the anthropogenic carbon
    should dissolve in the ocean. Roger Revelle [Revelle and
    Suess, 1957] realized, however, that the pH equilibrium
    chemistry of seawater will limit the uptake of CO2, by a
    factor now known as the Revelle buffer factor. At the
    preanthropogenic pH of the ocean, the buffer factor ranged
    from 9 in the tropics to 15 in high-latitude surface waters.
    The effective ‘‘size’’ of the ocean CO2 buffering capacity
    can be estimated as 40,000 Gtons of ocean DIC divided by
    the buffer factor from cold surface waters (15) to yield
    about 2500 Gtons buffering capacity. We expect that added
    CO2 will partition itself between the atmosphere and the
    ocean in proportion to the sizes of the reservoirs, and in
    the ocean we expect that size to be the buffering capacity. The
    relative sizes of the preanthropogenic atmosphere and the
    atmosphere plus ocean buffer are proportioned 560:(560 +
    2500) equals 18%. This crudely predicted atmospheric
    fraction is comparable to the model atmospheric fraction
    after 1000 years, which ranges from 14 to 30%, depending on
    the size of the fossil fuel release. As the anthropogenic CO2
    acidifies the ocean, the Revelle buffer factor increases,
    decreasing the buffer capacity of the ocean.”

    • Euan Mearns says:

      Dennis, I have not yet read the Archer papers. One of the aspects of running a blog is that I research and write a couple of articles per week and answer maybe 50 or so comments. Reading this key quote from Archer 2005 (that I really appreciate getting) is that this describes Phil Chapman’s model – so maybe it is the Archer model. But as I understand things right now the deep oceans do not contain 40,000 Gt of DIC. DIC as I understand it is bicarbonate – HCO3-. What they contain mainly is 40,000 Gt of rotting phytoplankton that sank there quickly by gravity. I’m guessing this is cellulosic material.

      The capacity of the biosphere to expand and contract is an open question in my opinion.

      • dennis coyne says:

        “Atmospheric CO2 is exchanged with the surface ocean through gas
        exchange. This exchange flux is driven by the partial CO2 pressure difference
        between the air and the sea. In the ocean, carbon is available
        predominantly as Dissolved Inorganic Carbon (DIC, ~38,000 PgC;
        Figure 6.1), that is carbonic acid (dissolved CO2 in water), bicarbonate
        and carbonate ions, which are tightly coupled via ocean chemistry. In
        addition, the ocean contains a pool of Dissolved Organic Carbon (DOC,
        ~700 PgC), of which a substantial fraction has a turnover time of 1000
        years or longer (Hansell et al., 2009)”

        From Chapter 6 of AR5, WG1 (p472)

        The 40,000 Gt of DIC was chosen for a round number in the Archer paper, dissolved inorganic carbon is defined in the quote above.

        Note that Phil Chapman’s model assumes that 80% of the emissions will be acted on quickly by very fast (photosynthetic) processes. Let’s assume that rather than assuming 20% of CO2 remains forever we use the 2009 Archer estimate that the e-folding time is about 13,000 years for 21.7% of emissions, this gives us about 20% remaining at 1000 years after emissions (this is the CaCO3 part of the cycle), then we have the biological pump, which we will assume acts on the order of 300 years(Your earlier link suggested 1000 years, I am being conservative). We also assume 25.9% of emissions are acted on by this Tau 300 piece, the rest (52.4%) is fast processes with a Tau of 0.43. This last value for the shortest Tau was chosen to match the model to 1910 to 2013 data for atmospheric CO2.

        You have asserted that if emissions of anthropogenic carbon drop to zero, that atmospheric CO2 will drop rapidly. What is it that drives a change in net uptake of carbon by the atmosphere? In other words, why does the biosphere (land and ocean net flux) currently absorb 3.9 Gt/a and about 1 Gt/a before 1750? I would argue that (ignoring nutrient limitations to simplify) it is primarily the change in atmospheric CO2 (dCO2) that has driven this increase in carbon uptake (U) by the biosphere. As dCO2 goes to zero, U will become negative and the excess carbon that has been stored in the Ocean will gradually be released back into the atmosphere. See chart below for a 4000 Gt CO2 (1090 Gt C) pulse over the 1850 to 2112 period with zero anthropogenic carbon emissions from 2112 to 3000 AD.

        • Euan Mearns says:

          The biological pump, represented by the right downward arrow, greatly enhances the surface–deep CO2 gradient, through the rain of organic matter (“Corg”) out of the surface ocean. The “carbonate pump,” represented by the left downward arrow, is the downward rain of calcium carbonate microfossils out of the surface ocean. Its effect is to actually raise the pCO2 of the surface ocean; this involves the alkalinity of seawater and is discussed in Chapters 6.04, 6.10, 6.19, and 6.20. Almost all of the organic matter raining out of the surface ocean is degraded back to CO2 and inorganic nutrients as it rains to the seafloor or once it is incorporated into the shallow sediments; only less than 1% (,0.05 Pg out of ,10 Pg) is removed from the ocean/ atmosphere system by burial (the downward blue arrow). This is in contrast to the calcium carbonate rain out of the surface (the downward purple arrow), ,25% of which is buried.

          I see from this they think the rain is CaCO3 and this goes back to CO2 at depth – CO2 would be favoured since it is less alkaline down there. But the removal process is very fast and is gravitational.

          If emissions switch off, the sequestration rate will initially be unchanged, but I agree that the rate of removal will decline with time as PCO2 drops with atmospheric CO2. I’ve not thought through the consequence of that yet. Would we not follow some sort of curve that mirrored the history of emissions?

          • Dennis Coyne says:

            Hi Euan,

            Why would the sequestration rate be unchanged?
            There would be a change in the rate that carbon dioxide is added to the atmosphere, this would effect the partial pressure of CO2 (tending to reduce it) and as a consequence the fertilization by excess CO2 would be reduced thereby reducing net primary production.

            Why is it that the net flux of carbon to the land and ocean combined has increased since pre-industrial times? Or do you believe that is not the case?

            We would not follow a curve that matches emissions. The fact that more emissions are not sequestered indicates that there is a limit to this process, the limitation is likely to be in part how quickly carbon is removed from the surface of the ocean and how quickly the CaCO3 formation proceeds, the speed of these processes has little to do with the path of emissions.

            The partial pressure of CO2 drives the rate of reaction of the biological processes. About 10 Gt of emissions are added each year (including land use change) about 55% is sequestered so let’s call it a net increase to the atmosphere of 4.5 Gt which results in a 2 ppm increase of pCO2, the 5.5 Gt of carbon taken up is the response to what otherwise would have been about a 5 ppm increase in CO2 if none of the atmospheric C had been sequestered. If we take away the emissions of 10 Gt , then you take away the change in partial pressure driving the reaction that takes up the excess C and in addition the reduced partial pressure results in a higher flow of CO2 from the ocean to the atmosphere.

        • Euan Mearns says:

          Dennis, I read the Archer article and my opinion at present is that it is fanciful rubbish.

          The Myth of The Rock Weathering Sink

          It seems to have become engrained in the folk lore of climate science that weathering rocks somehow creates a sink for CO2. The first time I read about this some years ago in a paper by J. Hansen I didn’t understand it and I do not understand it today. The storyline goes CO2 combines with water to make carbonic acid that weathers rocks and is subsequently fixed into river water as calcium bicarbonate ions and shortly thereafter gets dumped in the sea. The flux is small, around 0.4 GtC according to IPCC AR5. This gets blown into a huge amount by integrating the effect over geological time. I will argue that this is not valid. The correct approach is to compare the size of the annual flux with other processes. Rock weathering merely moves a tiny quantity of CO2 from atmosphere to ocean via a different route. Furthermore, since speciation is pH dependent, and waters associated with silicate bedrocks are often acid, it is not clear to me that the bicarbonate ion would be favoured.

          The AR5 carbon cycle does in fact show 1GtC per annum being released back to atmosphere from the riverine flux (freshwater outgassing). The riverine system works as follows:

          Rock weathering: 0.4 GTc
          Export from soils to rivers: 1.7 GtC
          Total input to rivers: 2.1 GtC

          Freshwater outgassing: 1 GtC
          Burial in estuaries and deltas: 0.2 GtC
          Export from rivers to oceans: 0.9 GtC
          Total export from rivers: 2.1 GtC

          Half of the rock weathering component seems to go straight back into the atmosphere and the approximate 0.2 GtC that makes it to the sea is insignificant compared with the ±80 GtC ocean-atmosphere and ±120 GtC biosphere-atmosphere exchanges.

          What rock weathering does do is to liberate cations into solution that may eventually become the salt in sea water or the Ca in limestone. But there are a host of weathering processes that can do this, for example the hydrolysis of plagioclase feldspar to the clay mineral kaolinite.

          Totally absent from the rock weathering debate is the potential liberation of CO2 from the vast C reservoir that rocks contain. Grid Arendal sees between 66,000,000 and 100,000,000 GtC in those sinks, the biggest number by far on the chart. Bacterial processes for example consume C and produce CO2.

          • Dennis Coyne says:


            Rock weathering is a pretty small part of the Archer story, it is about the biological and carbonate pump.

            So the whole “rock weathering debate” is really not at issue.

            Did we read the same paper? Which paper did you read?

            The rock weathering in the 2005 paper, though tested for completeness, had very little effect on the results.

            I guess when people have different points of view, they understand written English very differently.

            I thought the paper was excellent.

            Wow, unexpected.

          • Dennis Coyne says:

            Hi Euan,

            You are arguing that rock weathering is either not consequential or it is a potential source of CO2, I agree on the first point on the second point, in preindustrial times the land and ocean had a net flux to the atmosphere, volcanoes added to this, the rock weathering story is the explanation offered for how a long term equilibrium was established before anthropogenic emissions became very large (after 1750). For this reason I would say that rock weathering is likely to absorb carbon rather than release it, but I agree that the effect will be very slow.

            That is why Archer models it as an effect over 400,000 years, perhaps that is not slow enough for you 🙂

            I ignore the 5000 Gt case because I think that is too large (though there are people who believe such estimates, I would agree that they are not realistic.)

            So I focus my attention on the 1000 Gt and 2000 Gt cases because the eventual emissions are likely to fall between these two cases (my guess is that Archer would agree with this, no doubt the 300 Gt and 5000 Gt cases were chosen as limiting cases.)

            If you focus on Figure 1 and Table 1 you will see that there is very little difference between the CaCO3 case and the silicate weathering case.

            You may not like the choice of a 3 C sensitivity for deep ocean temperature, If you think there will be no temperature increase in the deep ocean, then one would simply choose the no temperature feedback case, if you think 1.5 C is the correct temperature sensitivity then you would choose the average of the no feedback case and the case with feedback and ignore the silicate weathering case if you wish.

            For 1000 Gt we would have 17% of the emissions remaining after 1000 years with no T feedback and no silicate weathering, and 20% with T feedback (at 3 C sensitivity), If the sensitivity were 2 C, I would estimate 19%. Note that if one were to include silicate weathering it would tend to reduce the fraction of CO2 remaining to about 18% for the 2 C sensitivity case.

            Now I expect that 1000 Gt is too low for anthropogenic emissions and expect around 1300 Gt at least. So using a linear approximation, we get about 23% remaining for the 2000 Gt no silicate weathering and 2 C T sensitivity (and unchanged when silicate weathering is included). For 1300 Gt I would estimate 19.5% of the emissions would remain after 1000 years if sensitivity were 2 C instead of 3C (and about 21% at 3C sensitivity).

            Perhaps you could point out the “fanciful rubbish” parts, with some specific quotations?

          • Euan Mearns says:

            Dennis, we will resume this conversation on Friday I presume. I’ve reached the conclusion that it is the mass of plants that determines CO2. If you switched emissions off, CO2 would fall with an exponential decay. How would it know when to stop? I think as partial P drops the rate of plant growth slows until a new equilibrium is reached at about 280 ppm.

            And a question. What physical evidence is there that CO2 has a long residence time in the atmosphere? Archer was quoting 300 years.

          • Euan Mearns says:

            Perhaps you could point out the “fanciful rubbish” parts, with some specific quotations?

            The idea that anthropogenic CO2 release may affect the climate of the earth for hundreds of thousands of years has not reached general public awareness.

            The maximum amount of fossil fuel carbon that could ultimately be released would seem to be about 5000 Gton C, on a timescale of several centuries.

            Weathering of the CaO component of igneous rocks acts to drag carbon from the atmosphere/ocean and deposit it as CaCO3 on the seafloor. The silicate weathering thermostat hypothesis is that the rate of igneous rock weathering increases with increasing atmospheric CO2, primarily by acceleration of the hydrologic cycle.

            The 400 kyr timescale of silicate weathering thermo- stat dominates the mean lifetime of a CO2 perturbation. Without and with the T feedback, the values are about 30 and 34 kyr, respectively. These lifetimes are nearly inde- pendent of release magnitude, with a slight increase as the release approaches 5000 Gtons, exhausting the CaCO3 dissolving capacity of the ocean.

  12. dennis coyne says:

    Hi Euan,

    The evidence is the record of CO2 levels for the last 12,000 years. Archer said if insist on using exponential decay (to describe processes which have much more in common with diffusive processes, see Fick’s Law), then we should assume about 20 to 25% of CO2 remains forever, the rest will decline with an e-folding time of roughly 300 years. This is intended as a rough approximation to give to the layman

    Why is it that equilibrium is reached at 280 ppm? What caused the net uptake of carbon to increase. It is not the residence time of individual carbon dioxide molecules that is important, it is the overall picture of how carbon cycles through the system that I am trying to understand.

    It strikes me as strange that you can see how at 280 ppm that the rate that carbon is removed from the atmosphere is balanced by other processes that emit carbon into the atmosphere so that over the past 10,000 years there was relatively little change in pCO2 up to 1750(about 20 ppm possibly due to land use change from 7500 BP to 1000 BP). All of these processes continue to operate at present. You cannot model the system very well if you ignore the processes that are emitting carbon into the atmosphere and only model the removal processes.

    From a system perspective, let’s ask the question how long does 2020 Gt (260 ppm) of carbon dioxide remain in the atmosphere? If we had posed the question 12,000 years ago, the answer should have been, “At least 12,000 years.” If you are concerned about individual carbon dioxide molecules and how long they remain in the atmosphere, I would say you are asking the wrong question, what matters is the number of carbon dioxide molecules in the atmosphere, residence time is of little relevance. It may be coincidence, but 13,000 for Tau works well for 21% of emissions with a Tau of 300 for 24% of emissions and the rest (55%) with a fast Tau of 3.

    If there is a rapid exponential decay of carbon dioxide, what is the mechanism that causes it to stop abruptly at 280 ppm?

    • Euan Mearns says:

      Exponential decay does not stop abruptly, it stops tangentially. And as I suggested that is where PCO2 may be in equilibrium with the biosphere – drop below that and the biosphere stops removing net CO2. It doesn’t die its just in balance.

      I’ve more or less convinced myself that the main mechanism for removal of CO2 is photosynthesis. There may be a small element of diffusion at the ocean surface but that gets driven by removal by phytoplankton. You can model the atmosphere using a single exponential decay – that is where all this started with Roger. So there is no need to invoke the long lived and infinite Taus.

      • dennis coyne says:

        You need to account for the removal of carbon from the surface of the ocean, there is a cycle involved, you have to account for the increased carbon in the ocean, which the focus on photosynthesis and its net uptake of 13 Gt of carbon per year (in the ocean) is not the whole story, the terrestrial sink is roughly in balance so the focus should be on the ocean. I think if you look at a mass balance over time you will find your story does not work. The ocean does not have unlimited capacity to absorb carbon and it takes time for carbon to cycle through the system.

      • dennis coyne says:

        Euan, the diffusion happens below the level where plankton are active, which is the top few hundred meters of the Ocean.

  13. Dennis Coyne says:

    Hi euan,

    As I said before the 5000 gt case can be ignored and the silicate weathering is also not that important and reduces the con fraction.
    This is a serious paper calling it fanciful rubbish is not particularly convincing.

    We will just have to diagree. A curve fit to data does not make a model. You need some chemistry and physics to back it up. Perhaps you could find peer reviewed literature which refutes archer’s 2005 and 2009 papers.

  14. Pingback: The Carbon Cycle: a geologist’s view | Energy Matters

Comments are closed.