Guest Commentary by Chris Colose, SUNY Albany
It has long been known that characteristics of the Earth’s orbit (its eccentricity, the degree to which it is tilted, and its “wobble”) are slightly altered on timescales of tens to hundreds of thousands of years. Such variations, collectively known as Milankovitch cycles, conspire to pace the timing of glacial-to-interglacial variations.
Despite the immense explanatory power that this hypothesis has provided, some big questions still remain. For one, the relative roles of eccentricity, obliquity, and precession in controlling glacial onsets/terminations are still debated. While the local, seasonal climate forcing by the Milankovitch cycles is large (of the order 30 W/m2), the net forcing provided by Milankovitch is close to zero in the global mean, requiring other radiative terms (like albedo or greenhouse gas anomalies) to force global-mean temperature change.
The last deglaciation occurred as a long process between peak glacial conditions (from ~26-20,000 years ago) to the Holocene (~10,000 years ago). Explaining this evolution is not trivial. Variations in the orbit cause opposite changes in the intensity of solar radiation during the summer between the Northern and Southern hemisphere, yet ice age terminations seem synchronous between hemispheres. This could be explained by the role of the greenhouse gas CO2, which varies in abundance in the atmosphere in sync with the glacial cycles and thus acts as a “globaliser” of glacial cycles, as it is well-mixed throughout the atmosphere. However, if CO2 plays this role it is surprising that climatic proxies indicate that Antarctica seems to have warmed prior to the Northern Hemisphere, yet glacial cycles follow in phase with Northern insolation (“INcoming SOLar radiATION”) patterns, raising questions as to what communication mechanism links the hemispheres.
There have been multiple hypotheses to explain this apparent paradox. One is that the length of the austral summer co-varies with boreal summer intensity, such that local insolation forcings could result in synchronous deglaciations in each hemisphere (Huybers and Denton, 2008). A related idea is that austral spring insolation co-varies with summer duration, and could have forced sea ice retreat in the Southern Ocean and greenhouse gas feedbacks (e.g., Stott et al., 2007).
Based on transient climate model simulations of glacial-interglacial transitions (rather than “snapshots” of different modeled climate states), Ganopolski and Roche (2009) proposed that in addition to CO2, changes in ocean heat transport provide a critical link between northern and southern hemispheres, able to explain the apparent lag of CO2 behind Antarctic temperature. Recently, an elaborate data analysis published in Nature by Shakun et al., 2012 (pdf) has provided strong support for these model predictions. Shakun et al. attempt to interrogate the spatial and temporal patterns associated with the last deglaciation; in doing so, they analyze global-scale patterns (not just records from Antarctica). This is a formidable task, given the need to synchronize many marine, terrestrial, and ice core records.
The evolution of deglaciation
By analyzing 80 proxy records from around the globe (generally with resolutions better than 500 years) the authors are able to evaluate the changes occurring during different time periods in order to characterize the spatial and temporal structure of the deglacial evolution.
Shakun et al. confirm Ganopolski’s and Roche’s proposition that warming of the Southern Hemisphere during the last deglaciation is, in part, attributable to a bipolar-seesaw response to variations in the Atlantic Meridional Overturning Circulation (AMOC). This is hypothesized to result from fresh water input into the Northern Hemisphere (although it is worth noting that the transient simulations of this sort fix the magnitude of the freshwater perturbation, so this doesn’t necessarily mean that the model has the correct sensitivity to freshwater input).
The bi-polar seesaw is usually associated with the higher-frequency abrupt climate changes (e.g., Dansgaard-Oeschger and Heinrich events) that are embedded within the longer, orbital timescale variations. However, numerous studies have indicated that it also sets the stage for initiating the full deglaciation process. In this scenario, the increase in boreal summer insolation melts enough NH ice to trigger a strong AMOC reduction, which cools the North at the expense of warming the South. The changes in Antarctica are lagged somewhat due to the thermal inertia of the Southern Ocean, but eventually the result is degassing of CO2 from the Southern Ocean and global warming. In particular, CO2 levels started to rise from full glacial levels of about 180 parts per million (ppm), reaching 265 ppm 10,000 years ago (or ~2.1 W/m2 radiative forcing), and with another slow ~15 ppm rise during the Holocene.

Figure 1: Simplified schematic of the deglacial evolution according to Shakun et al (2012). kya = kiloyears ago; NH = Northern Hemisphere
The evolution of temperature as a function of latitude and the timing of CO2 rise are shown below (at two different time periods in part a, see the caption). There is considerable spatial and temporal structure in how the changes occur during deglaciation. There is also long-term warming trend superimposed on higher-frequency “abrupt climate changes” associated with AMOC-induced heat redistributions.

Figure 2: Temperature change before increase in CO2 concentration. a, Linear temperature trends in the proxy records from 21.5–19 kyr ago (red) and 19–17.5 kyr ago (blue) averaged in 10° latitude bins with1σ uncertainties. b, Proxy temperature stacks for 30° latitude bands with 1σ uncertainties. The stacks have been normalized by the glacial–interglacial (G–IG) range in each time series to facilitate comparison. From Shakun et al (2012)
What causes the CO2 rise?
The ultimate trigger of the CO2 increase is still a topic of interesting research. Some popular discussions like to invoke simple explanations, such as the fact that warmer water will expel CO2, but this is probably a minor effect (Sigman and Boyle, 2000). More than likely, the isotopic signal (the distribution of 13C-depleted carbon that invaded the atmosphere) indicates that carbon should have been “mined’ from the Southern ocean as a result of the displacement of southern winds, sea ice, and perturbations to the ocean’s biological pump (e.g., Anderson et al., 2009).
This view has been supported by another recent paper (Schmitt et al., 2012) that represents a key scientific advance in dissecting this problem. Until recently, analytical issues in the ice core measurements provided a limitation on assessing the deglacial isotopic evolution of 13C. Because carbon cycle processes such as photosynthesis fractionate the heavy isotope 13C from the lighter 12C, isotopic analysis can usually be used to “trace” sources and sinks of carbon. A rapid depletion in 13C between about 17,500 and 14,000 years ago, simultaneous with a time when the CO2 concentration rose substantially, is consistent with release of CO2 from an isolated deep-ocean source that accumulated carbon due to the sinking of organic material from the surface.

Figure 3: Ice core reconstructions of atmospheric δ13C and CO2 concentration covering the last 24 kyr, see Schmitt et al (2012)
Skeptics, CO2 lags, and all that…
Not surprisingly, several people don’t like this paper because it reaffirms that CO2 is important for climate. The criticisms have ranged from the absurd (water vapor is still not 95% of the greenhouse effect, particularly in a glacial world where one expects a drier atmosphere) to somewhat more technical sounding (like criticizing the way they did the weighting of their proxy records, though the results aren’t too sensitive to their averaging method). There’s also been confusion in how the results of Shakun et al. fit in with previous results that identified a lag between CO2 and Antarctic temperatures (e.g., Caillon et al., 2003).
Unlike the claims of some that these authors are trying to get rid of the “lag,” Shakun et al. fully support the notion that Antarctic temperature change did in fact precede the CO2 increase. This is not surprising since we fully expect the carbon cycle to respond to radical alterations to the climate. Moreover, there is no mechanism that would force CO2 to change on its own (in preferred cycles) without any previous alterations to the climate. Instead, Shakun et al. show that while CO2 lagged Antarctic temperatures, they led the major changes in the global average temperature (including many regions in the Northern Hemisphere and tropics).
It is important to realize that the nature of CO2’s lead/lag relationship with Antarctica is insightful for our understanding of carbon cycle dynamics and the sequence of events that occur during a deglaciation, but it yields very little information about climate sensitivity. If the CO2 rise is a carbon cycle feedback, this is still perfectly compatible with its role as a radiative agent and can thus “trigger” the traditional feedbacks that determine sensitivity (like water vapor, lapse rate, etc). Ganopolski and Roche (2009), for example, made it clear that one should be careful in using simple lead and lags to infer the nature of causality. If one takes the simple view that deglaciation is forced by only global ice volume change and greenhouse feedbacks, then one would be forced to conclude that Antarctic temperature change led all of its forcings! The communication between the NH and Antarctica via ocean circulation is one way to resolve this, and is also supported by the modeling efforts of Ganopolski and Roche. This also helps clear up some confusion about whether the south provides the leading role for the onset or demise of glacial cycles (it apparently doesn’t).
A number of legitimate issues still remain in exploring the physics of deglaciation. For instance, the commentary piece by Eric Wolff references earlier deglaciations and points out that solar insolation may have increased in the boreal summer during the most recent event, but was still not as high as during previous deglacial intervals. It will be interesting to see how these issues play out over the next few years.
References
- P. Huybers, and G. Denton, "Antarctic temperature at orbital timescales controlled by local summer duration", Nature Geoscience, vol. 1, pp. 787-792, 2008. http://dx.doi.org/10.1038/ngeo311
- L. Stott, A. Timmermann, and R. Thunell, "Southern Hemisphere and Deep-Sea Warming Led Deglacial Atmospheric CO 2 Rise and Tropical Warming", Science, vol. 318, pp. 435-438, 2007. http://dx.doi.org/10.1126/science.1143791
- A. Ganopolski, and D.M. Roche, "On the nature of lead–lag relationships during glacial–interglacial climate transitions", Quaternary Science Reviews, vol. 28, pp. 3361-3378, 2009. http://dx.doi.org/10.1016/j.quascirev.2009.09.019
- J.D. Shakun, P.U. Clark, F. He, S.A. Marcott, A.C. Mix, Z. Liu, B. Otto-Bliesner, A. Schmittner, and E. Bard, "Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation", Nature, vol. 484, pp. 49-54, 2012. http://dx.doi.org/10.1038/nature10915
- D.M. Sigman, and E.A. Boyle, "Glacial/interglacial variations in atmospheric carbon dioxide", Nature, vol. 407, pp. 859-869, 2000. http://dx.doi.org/10.1038/35038000
- R.F. Anderson, S. Ali, L.I. Bradtmiller, S.H.H. Nielsen, M.Q. Fleisher, B.E. Anderson, and L.H. Burckle, "Wind-Driven Upwelling in the Southern Ocean and the Deglacial Rise in Atmospheric CO 2", Science, vol. 323, pp. 1443-1448, 2009. http://dx.doi.org/10.1126/science.1167441
- J. Schmitt, R. Schneider, J. Elsig, D. Leuenberger, A. Lourantou, J. Chappellaz, P. Köhler, F. Joos, T.F. Stocker, M. Leuenberger, and H. Fischer, "Carbon Isotope Constraints on the Deglacial CO 2 Rise from Ice Cores", Science, vol. 336, pp. 711-714, 2012. http://dx.doi.org/10.1126/science.1217161
- N. Caillon, J.P. Severinghaus, J. Jouzel, J. Barnola, J. Kang, and V.Y. Lipenkov, "Timing of Atmospheric CO 2 and Antarctic Temperature Changes Across Termination III", Science, vol. 299, pp. 1728-1731, 2003. http://dx.doi.org/10.1126/science.1078758
#43, #46
OT. I’ll just point out that the spirit is willing but the flesh is weak — and there are inducements aplenty for apathy and procrastination.
At some point you may have to ask whether it’s time (for some of us at least) to abandon politely whispering in peoples’ ears, drag out the big brass gongs and kick out the jams…
On statewide New Hampshire polls, we’ve been tracking beliefs about climate change four times per year. The latest (April 2012) results make an interesting picture:
http://img.photobucket.com/albums/v224/Chiloe/Climate/timeseries_warmop_2.png
The following is part of an entry sent to the bore hole. SecularAnimish criticized me, using information that I think is incorrect, and I explain why. By leaving SecularAnimist’s comment unanswsered, it gives the impression that it is correct.
SecularAnimist (#43)
About public opinion polls. You cite polls that find that people say they are concerned with climate warming. That’s simply not true (I don’t expect climate scientists to understand polls). People tend to say “yes” to whether they are concerned about something, like reducing climate warming, that seems good. But that is meaningless without knowing how conerned. In fact few people care much about climate change.
The real facts are shown in a recent Harvard survey of young Americans, 18-29 years old. The researchers asked people to rank domestic policy concerns, 12 of them, and “combating the impacts of climate change” came in last. See:
http://www.iop.harvard.edu/var/ezp_site/storage/fckeditor/file/spring_poll_12_topline.pdf
http://www.nytimes.com/2012/04/28/opinion/blow-young-peoples-priorities.html?_r=1
T Marvell wrote: “The question is why they selected these lags.”
Well, since one of the authors of Rahmstorf and Foster (2011) is a regular contributor here at RC, why don’t you ask him? You could also head over to Open Mind and ask the other author. And finally, you could ask John Nielsen-Gammon why here.
They’ve shown their work, now how about you show yours.
T Marvell wrote: “The question is why they selected these lags.”
The reason is that those were the lags which gave the best fit to the data. It’s rather a standard approach. We were hardly the first to determine that there’s a lag in the impact of el Nino (and volcanic eruptions) on temperature. It was even demonstrated in a paper by prominent global-warming denialists (who correctly estimated the lag, but the rest of their thesis was rebutted here).
If your “regression analysis shows that El-Nino has a huge impact on temperature, but only for the same month (witout lag)” then your regression analysis is wrong.
P.S. The claim that “when one says that El_Nino affects temperature, one is largely saying that temperature affects temperature” is incorrect, because it’s mistaken to flatly assert that “temperature is an element of El_Nino measures.” It is for some (such as the Nino3.4 index), but isn’t used at all for measures such as the Southern Oscillation Index.
I’d say that the one who is “fishing for the best model to get the results you want” is you.
T Marvell
Oh, my. You seem to be claiming that:
1. a pool of 18-29 year olds serves as a proxy for people of all ages regardless of the results of another, alll-ages poll and
2. the fact that some things concern them more than global warming proves that they’re not concerned about global warming.
Sad.
“(I don’t expect climate scientists to understand polls)”
Sure they do. I talk to climate scientists about polls all the time.
L. Hamilton @51:
Yes indeed, it’s fascinating that independents are making a big change.
Being sent to the bore hole is not something to be proud of. Why advertise one has maundered on or otherwise transgressed so badly that one has descended to being a frightful bore? Why bother coming here, where you have access to top people ready and willing to address your questions, only to promote yourself? Are you so sure of yourself that you have nothing to learn? I doubt it.
#53 Jim Eager:
I asked for the same, but it seems t marvell has no time to present his work or link to his publications. He is now fully occupied with educating the scientists here with how shortsighted and smallminded they are. Compared to Galileo, Einstein, Feynman, did I miss out somebody.
Marcus
Eager (53) and Tamino (54, 55)
About Foster and Rahmstorf
http://iopscience.iop.org/1748-9326/6/4/044022/fulltext/#erl408263bib19
They use a single lag (e.g., lagged two months) of the Multivariable ENSO Index (MEI) to adjust the temperature data. There are several problems with their procedures. First they use levels data, as opposed to differenced (change) variable, which one should never do without testing whether the data series is stationary. Climate data are generally non-stationary, requiring that the variables be differenced (in fact, climate scientists assume that temperature is non-stationary). Second, they give very little information about how they selected the lag. They say they get the “best fit”, but using what procedure? Third, and most important, they select only one lag of MEI, without looking into whether using more than one lag would provide a better fit, as I’m sure it would. Other odd choices are that they do not use all the years available for analysis, and they had different lags for differerent temperature measures.
I used a Granger regression to study the relationship betwen MEI and global temperatures, from 1960 to present. The dependent variable is global temperature, NCDC data for monthly anamalies. Both temperature and MEI are differenced (current year less prior year). I entered 12 lags of MEI (along with 12 lags of temperature, making it a Granger regression). All 12 lags of MEI have similar coefficients, and most are statistically significant. That is, there is no reason to select any one lag as being more important than others, and if one selects only a single lag, one is leaving out many additional important lags.
It is interesting that the MEI produces many significant lags, while ENSO3.4, which I have used in the past, is only significant to one lag.
Foster and Rahnstorf also factor in volcanos, using an aerosol index. Volcanos affect temperature almost exclusively in the summer, but their analysis assumes volcanos have the same impact year around.
As for my contention that the El_Nino measures contain ocean temperature as an element: First, it is important that when using lagged, differenced variables and a Granger regression, this is not important because there is no current-year relationship in the regression, directly or indirectly. Even so, it is worth noting, in case anyone wants to compare current-period temperature and El_Nino, the measures do indeed contain elements of temperature.
The Southern Oscillation Index (SOI) is indirectly based on temperature. It is “based on the (atmospheric) pressure difference between Tahiti and Darwin, Australia. It is highly correlated with tropical sea surface temperature anomaly indices recorded in Niño3.”
http://www.cpc.ncep.noaa.gov/products/outreach/glossary.shtml
Low atmospheric pressure tends to occur over warm water (and high over cold water), partly because of deep convection over the warm water. So the SOI is indirectly based on temperature because it is largely caused by it.
As for the MEI, Of the 6 elements, one is ocean temperature and another is pressure.
http://www.esrl.noaa.gov/psd/enso/mei/
T. Marvell,
It would appear that you are claiming that ENSO is equally likely to manifest at any of the following 12 months–but we know that is not so.
Likewise, we know that volcanic eruptions can affect temperatures year-round
Finally, “correlated with temperature” is not the same as temperature based.
Susan @58,
Yes, the tracking poll’s uptick among Independents is visually striking. It’s too soon to call it significant; we’ll see what happens in June.
Meanwhile there’s a paper coming out later this month with findings about how ideology interacts with science literacy to affect climate or science-related views.
t. marvell said:
There’s no reason you can’t analyze non-stationary data, but more to the point, what’s important is whether the *noise* is stationary. The residuals from the fit in Foster & Rahmstorf are indistinguishable from a stationary time series. I doubt you bothered to find out. First-differencing, on the other hand, introduces very strong lag-1 autocorrelation, which I seriously doubt you accounted for.
Lags were selected as those which gave the best fit from a sizeable range of possibilities which are made explicit. “Best fit” means least sum of squared residuals. A real scientist would make that assumption (since the fit it by least-squares), which is why the reviewers expressed no confusion and raised no objection. With this point you make it abundantly clear that the only thing you’re doing is trying to find fault. But you’re doing a miserable job of it.
Nobody claimed that including more lags won’t give a better fit. In fact it *will*. But that in no way invalidates the results of F&R, and it certaintly doesn’t alter the fact that when you remove a realistic estimate of the influence of exogenous factors, the remaining temperature evolution contradicts all claims that global warming has “stopped” or even “slowed down.”
The time span was chosen as that for which all five data sets have coverage. Having different lags for different temperature measures was indicated by the *data itself*. The lags are only estimates, you wouldn’t expect all the data sets necessarily to give the same result, and the difference between surface and lower-troposphere temperature lags is almost certainly physically meaningful.
Now you’re just making stuff up. This is really pathetic.
The Southern Oscillation Index is the pressure difference between Tahiti and Darwin. Your claim that because it correlates with temperature it is somehow “based on” temperature, is absolutely nonsense.
Your screed is blatantly, obviously, and pathetically nothing more than an attempt to save your own face by throwing mud on F&R. It demonstrates that your bias makes you unqualified to discuss the subject intelligently. You do not discuss, you babble.
Feel free to have the last word. I’m done with you.
L,
We will have to see if the trend continues. It may well be a consequence of the warm winter, as oppose to any real movement. The uptick appears too suddenly, such that it may not be sustainable.
“It may well be a consequence of the warm winter, as oppose to any real movement.”
Yes, but if so — interesting that it’s just with Independents. But that’s why I said that we’ll see what happens in June. One cool feature of our time series is that it has temporal/spatial resolution fine enough to see weather effects, should they occur.
t. marvell:
I’ve decided to try to help you understand where you’ve gone wrong. I’m skeptical that you’ll listen with an open mind, but I’ll try anyway.
By first-differencing the temperature data, you’ve eliminated the trend. But it is precisely in the trend that the causal relationship with CO2 lies. Steadily rising CO2 levels are the cause of the overall rising trend in temperature for at least the last 40 years and most of it for the last century or so. If you leave the trend in place, the long-term correlation of temperature and CO2 (or more generally, climate forcing) is obvious.
Of course correlation is not causation. The causation is in the physics, which frankly, cannot be denied sensibly.
When you eliminate the trend, only the short-term fluctuations remain. One cannot hope to detect the causal chain CO2 ==> temperature in them, because temperature fluctuations are dominated by other factors (including ENSO, volcanoes, solar variations, and “natural variation”) which overwhelm the impact of CO2 change on short time scales. That’s one of the points of Foster & Rahmstorf 2011.
In addition, there are short-term factors which affect both temperature and CO2, which could mislead you into believing the opposite causal direction. One such factor is el Nino. It changes surface temperature, and it also changes patterns of rainfall over large areas of the globe. These changes alter the biospheric carbon cycle, and can significantly affect how much carbon is cycled through plant matter, in turn causing changes in atmospheric CO2. This has been well-known in geophysics for quite some time. Hence the causal chain el Nino ==> *both* temperature and the carbon cycle, could easily mislead you if you eliminate the trend in temperature data (even more so if you also eliminate the trend in CO2).
I know that first-differencing to remove trend is common in time series analysis, and especially prevalent for those who approach it from an economics background. But in this case, it eliminates the very phenomenon which one is trying to determine — the effect of CO2 on long-term temperature trends. Seeking that very real relationship using only the short-term fluctuations is destined to fail.
There is a lively debate about Foster & Rahmstrof, between me on one side and Ladbury and tamino on the other. I think this is very important, for both substantive and methodological reasosn.
First, to answer Ladbury (#63),
1 – He said, “It would appear that you are claiming that ENSO is equally likely to manifest at any of the following 12 months–but we know that is not so.” I didn’t say that. I said there are “similar coefficients,” some significant and some not. The point here is that Foster and Rahmstorf selected only one lagged value of ENSO, when they should have added all lags that made a difference in the analysis. It looks like they are cherry picking.
2 – He questions my contention that the impact of volcanos is largely limited to the summer. Let me quote from WJ Burroughs, Climate Change, 2ed ed, p. 161, “. . . major erruptions produce a substantial drop in summer temperatures for two or three yeara. If anything, the effect on winter temperatures is a warming, . . .” My own regression research corroborates this. If Foster and
3 – He said, about ENSO measures “‘correlated with temperature’ is not the same as temperature based” He misquote me. I said temperature “is and element of” ENSO measures (such that same-year relationships between temperature and ENSO are largely an identity.
To answer tamino (65).
About stationarity – in time series analysis one simply must difference a variable if it is not stationary (unless it is cointegrated with another variable, which is not the case here). There is probably nothing more basic in time series analysis. Otherwise one gets what is technically called a “nonsense regression.” There are many tests for stationarity; the most basic is the Dickey-Fuller test, and it shows that temperature is not stationary [it is I(1)]. Neither tamino nor Foster & Rahmstrof test for stationarity. Tamino also claims that first differincing introduces autocorrelation. Without differencing temperature data, one gets huge (positive) autocorrelation (which is typical of a non-stationary variable in levels). Differencing typically introduces negative autocorrelation, not as bad as the positive correlation in levels.
Again tamino says that the lags were selected by finding the best fit in least squares. What test was used, if any? More important, tamino admits, which is obvious, that Foster and Rahmstorf did not actually look for the best fit because they did not look to see if they could get a better fit after entering more than one lag.
Tamino (and Foster & Rahmstrof) say the time span was chosen to be the same for all five data sets. Why is that important? In general, in time series analysis one should use all the data available.
About volcanos, see above. Foster & Rahmstorf should not have used the volcano variable.
tomero continured (#68).
This makes no sense. First differencing is done to prevent a “nonsense regression.” It is not done to remove a linear trend, which can be easily done in levels analysis. The fact that two variables generally move in the same direction does not imply causation. Even if there is causation, it says nothing about which variable causes which. This whole post is about the two-way causal relationship between CO2 and temperature.
This brings up a problem with Foster & Rahmstrof that I did not see before. Regression analysis assumes that the researcher is sure that the dependent variable (left side) does not cause the independent variable (right side). As this post makes clear, higher temperatures lead to more CO2 because warmer water outgasses more. The Foster & Rahmstrof analysis suffers from “simultaneity bias”, which renders the results are meaningless.
To summarize, the Foster & Rahmstrof paper sufferes from 3 unquestionably deadly mistakes:
1) doing the regression in levels
2) not properly looking for the best fit, because they include only a single lag
3) simultaneity bias.
[Response: You are wrong on multiple counts here. The evidence would be whether your proposed methodology would succeed in determining the correct causality in a perfect model situation that contains the relevant physics – it won’t. Volcanoes do have an impact on the annual temperature – see Thompson et al (2009), lean and rind etc. And Hansen et al 2007, shindell et al 2004 for more detailed model data comparisons. As taming notes, first differencing removes the signal one is trying to explain, and if this is a problem for your methods, so much the worse for your methods. The physics says that a linear trend in GHGs will cause (all other things being equal) a trend in temperature. Declaring that this is non-stationary and must be removed is perverse and aphysical. However, please remember that no one is claiming that CO2 is causal purely because it is going up- that would be prone to spurious regression issues. – gavin]
[PS: This general subject has been discussed ad nauseum in previous years – see here for instance, and has appeared in the literature as well (for instance, Kaufmann and Stern (2002)). You are following a well-worn path. – gavin]
Tamino #66
To dismiss one common method of dealing with serial correlation in climate data by saying, “the causation is in the physics” won’t wash. Most thinking sceptics accept the “the physics”; what they argue about is the balance between the natural and anthropogenic component of temperature change. If there was perfect correlation between the dependent and independent variables then the first order differences would also be correlated. The use of regression to identify the relationship between the differences would only break down if the underlying trend was linear.
Foster and Rahmstorf deal with autocorrelation in an appendix to their paper by considering the reduction in the number of effective of degrees of freedom.
The question of autocorrelation is an important one and it is probable that no ideal method dealing with it exists. It should however be possible to demonstrate that the underlying relationship remains valid despite the presence of autocorrelation. Any paper which ignores it would, rightly, be panned by both thinking and unthinking sceptics.
“About stationarity – in time series analysis one simply must difference a variable if it is not stationary (unless it is cointegrated with another variable, which is not the case here). There is probably nothing more basic in time series analysis.”
Cointegration requires finding a linear combination of the variables that is stationary. In F&R and some other cases, the models serve this purpose. If the residual series generated by a model show no significant autocorrelation, as tested by Ljung-Box portmanteau statistics, then they are not distinguishable from white noise — a covariance stationary process.
“About stationarity – in time series analysis one simply must difference a variable if it is not stationary (unless it is cointegrated with another variable, which is not the case here). There is probably nothing more basic in time series analysis.”
Cointegration requires finding a linear combination of the variables that is stationary. In F&R and some other cases, the models serve this purpose. If the residual series generated by a model show no significant autocorrelation, as tested by Ljung-Box portmanteau statistics, then they are not distinguishable from white noise – a covariance stationary process.
“There is probably nothing more basic in time series analysis.”
T Marvell, you have absolutely no idea who you are jousting with, do you?
T. Marvell,
OK, please tell me that you aren’t really saying that an RSS fit is “cherrypicking”.
And please tell me that you are not really saying that if A correlates with B then A must be based on B. That is not just wrong, it is a logical fallacy.
It makes no sense to first difference a variable if you are interested in the trend of that variable. That is simply nonsense. You aren’t even making sense now.
Ron Manley: “Most thinking sceptics accept the “the physics”; ”
They merely reject or ignore its consequences.
Re: #70 (Ron Manley)
I didn’t bring up the physics because of serial correlation. It was to address the issue of “correlation is not causation.” I think you’re confused about the meaning of this.
Re: #69 (t. marvell)
No, there isn’t a “lively debate.” Your response is just absurd.
Perhaps much of your nonsense is rooted in the fact that you have confused “stationary” with having a “unit root” (which is what the Dickey-Fuller test is for, not for stationarity). And by the way, when it comes to unit root in temperature data, no, there is not.
I was foolish to think you might listen with an open your mind, but it was worth it for the sake of other readers. Enjoy your extended stay in Dunning-Kruger land…
A rapid depletion in 13C between about 17,500 and 14,000 years ago, simultaneous with a time when the CO2 concentration rose substantially, is consistent with release of CO2 from an isolated deep-ocean source that accumulated carbon due to the sinking of organic material from the surface.
Two things.
First, the chart underneath it seems to show the concentration falling – the vertical axis in the upper figure is reversed, perhaps the one in the lower figure should be?
Second, any idea what the isolated deep-ocean source might be?
Jim Eager:
Ssssh it will be more fun if he doesn’t figure it out.
If I understand the issue correctly we can determine that the Mak cycles produce relatively large regional variation in energy hitting the surface but globally the differences are very small. The question then becomes why do we get ice ages?
When I look at the globe I see mainly land mass in the Northern Hempisphere (especially near the pole) and mainly ocean in the South (again especially near the pole). A 30 W/m change in energy hitting the surface of the earth is enough to cause snow to accumulate – but this can only happen on land. Hence when the Mak cycle leads to reduce energy in the North ice sheets form, this increases albedo which reduces overall energy being absorbed by the plant and we have an ices. Conversely when the cold is in the South the ice sheets cannot form on open ocean?
How much does the above (which clearly happens) account for the observed historical record and therefore how much is left to explain but other factors such as CO2?
If my maths are correct – ice sheets down to the 45deg North will have a signifacnt albedo effect?
Regards
Ladbury (74) You say “please tell me that you are not really saying that if A correlates with B then A must be based on B. That is not just wrong, it is a logical fallacy”
Yor’re not immune from the fallacy. In comment 38 you said “if you look at the warming over the past 35 years, the correlation with delta ln(CO2) is quite significant.” From the context it is clear that you are arguing that more CO2 leads to higher temperatures (you answered my earlier comment that I could find no evidence that CO2 affected temperaturee in near term; I said nothing about a lack of correlation). Of course, the correlation you mention could also be due to higher temperatures increasing CO2 due to outgassing.
Comments 62 on, about Foster & Rahmstrof –
tamino – your’re right, I checked and the monthly temperature data that Foster & Rahmstrof use is trend stationary, so regression in levels with a trend is OK.
It seems that you have no answer to my original objection to Foster & Rahmstrof, that their “best fit” procedure is bad, mainly in that they would have had a better fit by using more lags. More lags would have given them a more accurate adjustment of the temperature series.
Ladbury (74) About regressing in differences with trending variables – differencing does not lose the trend; it goes into the intercept coefficient. Again, if the trending variable isn’t stationary, one ordinarily has to difference.
– About cherry picking. Climate scientists on RC often accuse others of cherry picking. Only others do it?
The marvell wrote “Of course, the correlation you mention could also be due to higher temperatures increasing CO2 due to outgassing.”
No, it could not, because sea water pH is decreasing, which means the ocean is absorbing more CO2 than it is outgassing, unless you know of some other new source of CO2.
T Marvell, OK, so now you are saying that it is a logical fallacy to cite the combination of a good correlation between ln[CO2] and temperature rise + the well known physics of greenhouse warming as evidence in favor of the consensus view? Really? Isn’t that pretty much the basis of the scientific method?
Let’s be clear here so I can figure out just how dumb you are: Are you really questioning the existence of the greenhouse effect, which has been known since the 1850s? Are you questioning whether CO2 is a greenhouse gas? Or do you think that somehow the greenhouse properties of CO2 magically stop once it reaches 280 ppmv?
As to you contentions of cherrypicking, no one claimed immunity from that failing–I merely said that taking the best fit according to least-squares is NOT cherrypicking, and to contend otherwise is ludicrous.
Finally, your contention that F&R should have used multiple lags is problematic. It is certainly true that allowing multiple lags might have given a better fit–after all it has to be at least as good since the >=1 lag models contain the single-lag model as a subset.
However, the goal in scientific modeling is not mere explanation but rather predictive power, and generally the simpler model that explains the data adequately has greater predictive power. The model in F&R is elegantly simple and does a good job of showing that a linear trend due to CO2 + a few forcings that we know to be operant and important are sufficient to explain most of the variability in all of the temperature datasets. That is an important and robust result–but then to understand that, you’d have to be able to find your but with both hands and a GPS when it comes to scientific modeling.
T Marvell
Two problems:
Where’s all the CO2 going that humans pour into the atmosphere?
Why does the observational evidence support the fact that the oceans currently are a sink, not a source?
You’re really going into la-la land now.
T Marvell,
It’s a bit hard to find words to describe the range of techniques used by artificial skeptics to discredit progress in science. “Cherry picking” is just one shortcut to describe that, for example, isolating counter-trendaceous segments of the temperature record doesn’t work, or that each of the proxy records has shortcomings but taken in sum they add to our knowledge. An open-minded reading of Mike Mann’s latest gives an overview of much of the hard work over time to try to find ways to extend and verify the record, and others have tried as well.
http://www.skepticalscience.com/graphics.php?g=47
Your tone is arrogant, but your conclusions make you sort of undressed in public to real scientists who do this work every day.
It’s not that real questions don’t need to be answered: they do. It’s that these kinds of questions come up over and over and over … hundreds, thousands of times, and the responses are not heeded.
They are a kind of dragon’s teeth and the more they are discredited the more they appear.
It is not surprising that scientists, who are not known to suffer fools gladly, to say the same thing in a different way, like most intelligent people are impatient with nonsense, are more than tired of posturing, whether naive or dishonest.
Unfortunately, this is all playing out in a real world that is heading, like the Titanic, for steadily worsening conditions as the real causes and problems are ignored and politically motivated naysayers decide they know everything bout reality.
@ t marvell
Lively? Maybe. Debate? No, not so much. You express no awareness of whom the two individuals are that you are engaging in “debate” with nor the credentials they bring to the table. And you consistently ignore the advice & feedback given you by Gavin (do you even understand who he is?) plus the mountainous body of evidence in flagrant contradiction to your position.
You espouse nonsense.
If the Drake passage became ice covered a the Last Glacial Maximum, that would have had a large influence on scale and patterns of Thermohaline Circulation. Note the question marks around the southern tip of South America, and the fact that the Falkland Islands kelp genetics indicate ice scour at the LGM – Figure 3 http://www.pnas.org/content/106/9/3249.full. Coupling from the Antarctic Circumpolar Vortex to the Antarctic Circumpolar Current, flow through the Drake Passage, and Eckmann transport driving the Atlantic Meridional Overturning Current would all be changed – see http://climate.gmu.edu/research/drake.php, especially Figure 4. Perhaps the upper panel in figure 4 is similar to the conditions of an ice covered Drake Passage, with the cutoff current loop to the South of 50 degrees representing a carbon pool of unventilated water. As Milankovic warming in the norther hemisphere is carried into the shallow return current, the Antarctic Ice margin retreats; a tipping point is reached where the Drake Passage flow restarts or increases above a threshold. This starts bringing old water from deeper in the ocean, releasing more CO2 into the atmosphere, and driving the end to the glacial age. (Yes I know it’s a lot of speculative qualitative handwaving &;>)
Ladbury (83) Anderson (85). I am not a skepting about human-caused global warming. I am very skeptical about the ability of climate scientists to convey that idea to the public. Their defensiveness is enough to turn people off. Any criticism gets emotional reactions. Anderson: In my view, the climate scientists are standing behind the helm of the Titanic, looking at the iceburg, but doing almost nothing.
I’ve been trying to research the skeptics’ claims. Some don’t pan out at all, like the notion that sun spots or cosimic rays are behind the temperature growth. On the other hand, I cannot rule out other claims, particually that there is no statistical relationship indicating that CO2 causes temperature growth. A big issue here is that any correlation between the two can be due to CO2 outgassing. Gavin – please tell these jerks that this is likely. Just because skeptics make this claim does not mean it is false. This whole post is about the two-way relationship between temperature and CO2.
Gavan linked me to an article by Kaufmann and Stern, saying that greenhouse gasses are cointegrated with (i.e., closely tied to) greenhouse gases. Cointegration says nothing about causal direction, yet they claim that the cointegration is due to greenhouse gases causing temperature. Any knowledgable skeptic would laugh at that paper. From the point of view of an outsider, Foster & Rahmstrof has the “look and feel” of confirmation bias.
As part of the general defensiveness, I’m sure that you consider me a “closet skeptic”. I’m not. But I am very angry that the climate science community is “circuling the wagons” in defense, rather than trying to do an effective job of persuading people to act to combat global warming.
I’m trying to figure out why, statistically, there is no causal link from CO2 to temperature. A theory I am working on is that a strong direct link only occurs in cold temperatures, where the greenhouse effect is greatest. Links to temperature in other areas are diffuse. So far this holds up statistically, at least with respect to land temperatures in the north. I’ve got a lot more work to do.
Can anybody point me to sources that explain why the greenhouse effect is greatest at the poles, and how much greater it is? I’ve just seen general statements to that effect.
One last, but interesting, point. All this has gotten me to look more at the issues of stationarity. I was surprised to find out that land tempature since 1900 is stationary (NCDC data, monthly anamolies, ADF with 12 lags). Ocean temperature is far from stationary, but it is trend stationary. There appears to be some self-regulating factor on land, while ocean temperatures plow upwards.
“particually that there is no statistical relationship indicating that CO2 causes temperature growth”
I think that skeptics have rediscovered that (statistical) correlation does not prove causation.
On the other hand, if polyatomic gas molecules in the atmosphere absorb Lambertian infrared radiation coming from the ground and re-emit it isotropically, the effective emissivity decreases, and the surface will warm until balance between insolation and radiation is achieved. More CO2 from fossil fuels will increase the intercepted radiation, and further raise the temperature. More water vapor from rising temperatures will further increase the temperature rise. Rising temperatures will cause outgassing of CO2 from the oceans, but its C12/C13 ratio will be that of the atmosphere when sinking thermohaline circulation took the CO2 from the atmosphere ~1600 years ago, which is different from fossil fuels. This will cause a higher climate sensitivity, but with a long lag ’til equilibration(800 years, perhaps; sound familiar?). Rising temperatures will melt snow and ice, mostly near the poles, and cause larger temperature rises at the poles, firstly in the Northern Hemisphere.
On the third hand, Causation forces correlation; if these causal chains have occurred, there will be correlation. Is this model wrong? Nope. If you describe the model with equations, instead of narrative, and plug in the numbers, the models are inaccurate; partly because the initial conditions aren’t known precisely(measurement noise), partly because the system is chaotic(turbulent, locally divergent – THIS rising bubble of warm moist air got a head start because a farmer plowed his field yesterday, decreasing the albedo, starting the thunderstorm that sucked up lot of nearby warm moist air, and spawned a tornado that flattened Picher OK), and partly because those two things interact(the butterfly effect).
Have the skeptics created an internally consistent narrative model which describes a different outcome? Let us take for example something you touched on, “…A big issue here is that any correlation between the two can be due to CO2 outgassing…” plus the argument that “the Medieval Warm Period was warmer than today” or “CO2 increases lag temperature increases by 800 years”. The ice cores don’t show 390ppmv+ CO2 that would have outgassed if these arguments were true. See Figure 3 of “Atmospheric CO2 fluctuations during the last millennium reconstructed by stomatal frequency analysis of Tsuga heterophylla needles” Kouwenberg et al http://fm1.fieldmuseum.org/aa/Files/lkouwenberg/Kouwenberg_et_al_2005_Geology.pdf Yet another hockey stick, and declining CO2 ~950-1200 AD – the peak of the MWP according to http://www.co2science.org/data/timemap/mwpmap.html
As for paleoclimate CO2/temperature changes around the Younger Dryas, “The reconstructed atmospheric CO2 reduction of ≈25 ppmv indicates a temporarily enhanced North Atlantic sink for CO2 at the time of the 8.2-ka-B.P. cooling event.” http://www.pnas.org/content/99/19/12011.long; somewhat less than the ~110ppmv modern increase in CO2.
Which leads me to another question – the melting glacial/Greenland/Antarctic ice water is depleted in CO2(check out the bubbles in your ice cubes) – how much additional CO2 is being sequestered by this runoff into the oceans, and what happens to CO2 increase when we run out of glaciers? My S.W.A.G. is that there’ll be so much runoff from Greenland/Antarctica by then that it won’t make much difference.
Dodge 89 – I have no trouble with this.
You say “if polyatomic gas molecules in the atmosphere absorb Lambertian infrared radiation coming from the ground and re-emit it isotropically, the effective emissivity decreases, and the surface will warm until balance between insolation and radiation is achieved.” What I need to know is whether the absorbsion and emmissions depends on temperature – is it greater for CO2 when temperatures are colder? Do you know of anything I can read about this? As a non-climate scientist, I have difficulty navagating the literature.
69 t marvel says, ” There is probably nothing more basic in time series analysis.”
I’m assuming you’re not a climate scientist. Please correct that assumption if appropriate.
When I read something by not just an expert, but a top-notch one, and my initial thought is that it violates something that is “most basic”, yet other experts in the field don’t seem to “get” the “obvious” flaw, then I try very hard to figure out what it is I’M missing. The PIOMAS thread is a perfect example. I STILL don’t understand the reactions – on one hand the models are “predicting” triple the current ice (since ice has a short memory, initial conditions should only have a weak influence), yet on the other hand people seem to be predicting another record or near record low while at the same time not objecting about the “super-flawed” models. It makes no sense to me. It is “most basic” that either the ice should double or triple quite quickly or the models are seriously flawed. My conclusion is that I don’t quite get it yet, not that the experts are failing at undergraduate-level concepts. Perhaps you’re in a similar situation here as I’m in there.
And if anyone would kindly go to the PIOMAS thread and help me with my conundrum, I’d appreciate it. Thanks.
T Marvell:
Why would gavin tell us jerks that this is likely given that the observational evidence is that the oceans are a net *sink* not *source* in today’s world?
You have absolutely no clue as to what you’re talking about.
Here’s just one paper addressing the issue.
A more extensive search will lead you to a long, long list of references making this clear.
Your claim is not “likely”. Not even close.
T. Marvell:
Well, Gavin is too polite to call me a jerk, but I’m happy to be one if that helps. However, I’m not one of the the many knowledgeable people here who are worth the attention if you can get those plugs out of yours ears/brain.
The arrogance is even more apparent when the namecalling comes forward.
“circling the wagons” is not what is happening here. People are trying hard to help you out, and you are not paying attention and doing your homework, which irritates people who have better things to do than be ignored.
It is profoundly boring to have someone self-righteously trying to reinvent the wheel and demonstrating a lack of observation and critical thinking in the process.
T Marvell,
If you have this notion that ocean outgassing is the cause of the current CO2 increase, I recommend you do some simple calculations for yourself (see below). The calculations will require you to look up some things, but they’re all freely available on the internet. No scientific literature required.
The calculations are as follows:
1. Calculate the average annual increase in total CO2 in the atmosphere in the last decade
2. Calculate the average annual net emissions of CO2 by humans over the last decade
3. If you believe oceans are responsible for the increase, calculate how big its average net annual emissions must be to overwhelm those of humans; you calculated the latter value in point 2. Just assume that the net annual emission by the ocean must be more than 50% of the net emissions by humans
4. Now calculate how much CO2 must be taken up, annually, by some third ‘source’ (a net sink) in order to allow point 1 to still be valid.
You will have to make some assumptions here and there.
Now, identify that “sink”, or ask those that claim the increase is due to ocean outgassing to identify that sink for you. Assess its likelihood of being a sink of such a large amount of CO2. Remember, this calculation you just did only applies to the last decade.
Good luck!
t. marvell: “On the other hand, I cannot rule out other claims, particually that there is no statistical relationship indicating that CO2 causes temperature growth.”
OK. I’m gonna try to be nice here. Where in the hell did you get this little nugget? The phrasing doesn’t even make sense! What is more, it is simply too vague even to tell you where you are going wrong. Ferchrissake, we have a 35-year warming trend; we have diurnal, seasonal and latitudinal dependence that is exactly what would be predicted from greenhouse warming; we have a cooling stratosphere even as the temperature warms; we have the glacial/interglacial cycle…do I need to go on? Where in the hell are you getting this crap?
WRT the source of the CO2, Brian has already pointed out the isotopic signature of the CO2, but over and above this, we’ve produced about 2x as much CO2 as has gone into the atmosphere! The oceans are becoming more acidic. You really haven’t looked at this much, have you?
WRT your “theory”, sorry, it doesn’t hold up. Temperature effects on the absorption of IR by greenhouse gasses won’t help you, as they are pretty tiny over the range of terrestrial temperatures. The distribution of energy states will follow the Maxwell Boltzmann equation. The reason you are having trouble finding out why you are wrong in the climate literature is because you are wrong on a much more fundamental level. This is basic atomic physics.
Even more fundamentally, you are trying to solve a problem that doesn’t exist by proposing a theory based on physics you don’t understand. Please, please, please, quit going to the skeptic sites until you have enough of an understanding of the actual science to avoid being fooled.
The reason why climate science has “circled the wagons” against the denialists is the same reason why biologists have circled the wagons against creationists or that medical scientists have circled the wagons against the anti-vaxxers. Science–science itself, not just climate science–is under attack. The same people are behind this as were behind the pro-tobacco crap of the 60s-90s–the exact same people.
t marvell, to say you are ignorant of climate science is not an insult. It is a fact. The way to rectify that ignorance is to learn the science, and you need a strong grounding in the mainstream science before you can evaluate its strengths and weaknesses. I have a PhD in physics. It took me about 2 years to learn enough of the basics that I felt grounded enough to understand what was going on. You seem like a smart guy, you could probably do it in a comparable time. There are plenty of others who have gone through the same process and would be happy to help you out as well. Do the math.
t marvell
Ah yes “circling the wagons”, you do realize that that’s a golden oldie from the threadbare grab bag of lame denialist epithets?
So the frustrating thing seems to be here that in the “marketplace of ideas” AGW isn’t selling properly, so it must either be a bad product or badly presented. Perhaps we should just pause for a moment, take a deep breath and bow our heads to quietly reflect on this odd notion and the worldly marketplace upon which it is modeled.
You’d think that after the last few years, people would be a little more skeptical of marketplaces in general, seeing them less as shining citadels on the mountain top and more as tawdry bedlams of hawkers and posers, grifters and marks. On the other hand, I guess I can see how those wishing to elude culpability or deny their gullibility might wish to double down on burnishing the old illusions.
All I can say is wakey, wakey Gold Bricks! It’s time to get up and greet the new day. There’s work to be done. Perhaps we should begin by taking out all that libertarian garbage thats been piling up over the last few decades.
t. marvel,
While claims of outgassing are valid during previous eras, there is one component present today that was not around back then. That is the burning of carbon-based fuels. In order for the CO2 increase to mostly a result of increasing temperatures, we would need to prove that the amount of CO2 being emitted is minor in comparison. Since most calculations contend that the amount of CO2 emitted shoudl has resulted in twice the increase measured, I would agree that the CO2 rise is not a result of outgassing.
I say it’s word salad from a Turing Test.
Dhogaza 92, Anderson 94, Marco 94, Ladgurn 96, Dan H. 97.
The physics couldn’t be clearer – as water warms it can hold less gas, including CO2. If you are contesting outgassing, you are physics deniers.
I have discussed this in earlier comments – in the post on Lindzen in March – see 460, 482, 517. Gavin accepted the reality of outgassing. Again, my research finds that higher ocean (but not land) temperatures lead to more Co2 a few months later. The results are highly significant
statistically, although the magnitude is small compared to human-produced Co2.
[Response: Many things are correlated with the seasonal cycle. Concluding that seasonal correlations must relate to ocean outgassing is not however a sensible conclusion (indeed, this is a very small component). – gavin]
You can do a quick test of this relationship. Ladbury said above (#37) “if you look at the warming over the past 35 years, the correlation with delta ln(CO2) is quite significant.” Is that correlation due to CO2 increasing temperature, or temperature affecting CO2? To begin with, the former seems unlikely because the global warming effects of CO2 are unlikely to become manifest right away (the correlation between differenced variables, which Ladbury uses, only pertains to same-year relationships). I propose the following test:
Using annual CO2 and temperature levels (not anomalies), take logs and differences as Lanbury does (thus getting, essentially, yearly percent changes). Correlate last year’s CO2 with this year’s temperature, and then the other way around. It’s an easy analysis, and certainly Ladbury can do it. The results: last year’s CO2 has a negative correlation with current year temperature (almost certainly because volcanos give off CO2).
[Response: Nonsense. Volcanic signals are orders of magnitude too small to emerge in such an analysis. Most of this is related to deforestation variability and ENSO (and impacts of ENSO on the terrestrial carbon cycle). Nothing to do with outgassing. – gavin]
The correlation between lagged temperature and current CO2 is positive, large, and highly significant, much more so than the correlation between current year CO2 and temperature changes. The relationship occurs only for ocean temperature. That is strong evidence that higher ocean temperature increases CO2 in the short term.
Carbon sinking obviously is a countervailing factor. Whether it overshadows the outgassing effect depends on timing and location. Sinking takes time, since ocean mixing is slow below 200m. Outgassing is greater in select areas, such as when warm water mixes with cold water, causing the cold water to lose CO2. I don’t know if anyone has modeled the tradeoff between outgassing and carbon sinking over the short term (a year or two) that is relevant to Ladbury’s correlation and my regressions. Empirically, however, it is clear that outgassing dominates in the short term. I admit that the long term is more important, but Ladbury’s results are just for the short term. This positive feed-back between temperature and CO2 can accellerate AGW, and I hope it is incorporated into the climate models.
T Marvell,
Maybe I misunderstood what you meant, but to me it sounded as if you claimed ocean outgassing is responsible for the current increase in atmospheric CO2. PLease correct me if I am wrong.
The calculation I proposed you’d perform would show you there’s a big problem in that claim: we’d need a sink that has taken up well over 100 Gigatons of carbon in the last decade alone.
Regarding outgassing, please note this paper:
http://www.climate.columbia.edu/sitefiles/file/2002Takahashi.pdf
which is freely available, so no excuse for not reading it.