Guest essay by David Archibald
President Obama didn’t start the war on coal. That war had its origins back in the 1970s. The nuclear industry joined the fray in 1982 with the establishment of the Carbon Dioxide Information Analysis Center (CDIAC) at Oak Ridge, part of the U.S. Department of Energy. The CDIAC collects data on carbon dioxide concentrations around the planet and conducts experiments with pre-ordained outcomes. By that I mean growing plants in elevated carbon dioxide concentrations to study the effects of that on growth rates but at the same time adding ozone so that the growth would be stunted. Not everything the CDIAC is completely useless though.
The pause in global temperature rise might cause a loss of faith in the global warming faithfully so the priests of the movement are required to provide an explanation. The explanation they have come up with is that the missing heat is hiding in the depth of the Altantic Ocean and will one day leap out at us when we are least expecting it. This is an illustration of the heat gone AWOL:
The illustration shows heat plunging into the depths as far as 1,500 metres. The oceans don’t work like that. Most of the heat energy of sunlight is absorbed in the first few centimetres of the ocean’s surface. Waves mix the water near the surface layer such that the temperature may be relatively uniform in the top 100 metres. Below that there is almost no mixing and no vertical movement of water.
This is where the CDIAC comes in handy. Following is a map of CDIAC voyages in the Atlantic Ocean:
And this is the temperature profile of A16 from almost 60°S to near Iceland, a distance of over 13,000 km.:
It shows how the Antarctic is a giant refrigerator for the planet. The dark blue in the bottom left is cold water below 1°C plunges near Antarctica and ponds in the deep ocean right up to the equator. The CDIAC voyages also record carbon dioxide data of course. This is the carbon dioxide and total alkalinity profile for A20, to the west of the A16 voyage:
Once again, most variation is near surface while the bulk of the ocean is effectively homogenous.
We didn’t need the CDIAC data to debunk claims of missing heat in the ocean depths but it is good to have empirical data. The CDIAC is well past its use-by date though. Apart from the unnecessary cost, it was conceived for a dark purpose under President Carter. The United States will need all the energy it can get soon enough.
David Archibald, a Visiting Fellow at the Institute of World Politics in Washington, D.C., is the author of Twilight of Abundance: Why Life in the 21st Century Will Be Nasty, Brutish, and Short (Regnery, 2014).
Reference:
Is Atlantic holding Earth’s missing heat?
Eli Kintisch
Armchair detectives might call it the case of Earth’s missing heat: Why have average global surface air temperatures remained essentially steady since 2000, even as greenhouse gases have continued to accumulate in the atmosphere? The suspects include changes in atmospheric water vapor, a strong greenhouse gas, or the noxious sunshade of haze emanating from factories. Others believe the culprit is the mighty Pacific Ocean, which has been sending vast slugs of cold bottom water to the surface. But two fresh investigations finger a new suspect: the Atlantic Ocean. One study, in this issue of Science, presents sea temperature data implying that most of the missing heat has been stored deep in the Atlantic. The other, published online in Nature Climate Change, suggests a warming Atlantic is abetting the Pacific by driving wind patterns that help that ocean cool the atmosphere. But some climate specialists remain skeptical. In a third recent paper, also published online in Nature Climate Change, other researchers argue that the Pacific remains the kingpin. One reason some scientists remain convinced the Pacific is behind the hiatus is a measured speedup in trade winds that drive a massive upwelling of cold water in the eastern Pacific. But there, too, the Atlantic may be responsible, modeling experiments suggest. A consensus about what has put global warming on pause may be years away, but one scientist says the recent papers confirm that Earth’s warming has continued during the hiatus, at least in the ocean depths, if not in the air.
Bart August 29, 2014 at 9:53 am
No, Edward, that is not at all what is happening. The linear trend component remains, but becomes a constant offset. The rate of change of atmospheric CO2 is proportional to temperature relative to a particular baseline temperature. We find that baseline temperature by comparing with the offset in the rate of change of CO2.
The model is
dCO2/dt = k*(T – To)
where To is the offset temperature, and k is a coupling factor. This is a first order expansion of the true relationship over a limited timeline. We do not know how long the timeline lasts over which the equation should be accurate, we only know that it is remarkably accurate over the past 56 years, since reliable measurements of atmospheric CO2 became available, and over which time the greater part of the modern rise is observed.
To fit anthropogenic additions in, we would have to decrease the k factor, but there is little room to do that and maintain a match with the higher order variability in the data. Hence, anthropogenic emissions cannot be a significant driver of atmospheric levels.
Incorrect as always Bart.
As your own graphs show you can only fit the temperature effect to the CO2 if you remove a large constant term from the slope.
A more correct model would be:
dCO2/dt = k1(dCO2ff/dt) +k2*(T – To)
Where k1(dCO2ff/dt) is the larger term, this is what the real world data shows us for annual data.
No, Phil. This is only the effect of the baseline temperature anomaly. It is true that this is an arbitrary parameter to fit, but there is also a slope in temperature, which is not arbitrary and cannot be removed.
That term matches the slope in dCO2/dt with the same scale factor with which the variations are matched. And, what that means is that there is no room for a significant effect from your dCO2ff/dt, because it also has a slope, and the slope is already accounted for by the temperature relationship.
[Snip. “beckleybud” sockpuppet. Banned. ~mod.]
pyromancer:
Can you provide specifics on how Phil has obliterated Bart’s post on this particular topic?
pyromancer,
We are always amused by assertions from the peanut gallery.
But if you want to be taken seriously, show where Bart is wrong. If you can.
Snip. Sockpuppet. ~ mod.
Snip. Sockpuppet. ~mod.
pyromancer76@gmail.com
No, that is not the argument. Please do not interfere and pollute the discussion until you do understand it. This is not high school. We do not need cheerleaders.
[Note: This sockpuppet is “H Grouse” and “beckleybud”. He has been banned multiple times. ~mod.]
Phil:
Let me attempt once more to explain this simply. Your having written this equation may help.
dCO2/dt = k1(dCO2ff/dt) +k2*(T – To)
There is a value for k2 which, with k1 set to zero, very capably fits the data. That value of k2 reproduces both the slope and the variability which match very closely.
If we boost the value of k1 to anything significant, we will have to decrease the value of k2 so that the slope will still match. If we do that, the variability will no longer match.
Note that this argument does not depend on the baseline anomaly offset. We would expect an offset, since the baseline chosen by the record keepers is arbitrary, and the odds are well against it being the precise value needed. But, that offset is, indeed, tunable. If that were all I had to go on, your argument would be apposite.
However, that is not all there is. We have to match both the slope, and the variability. The simplest way to do that is to choose k2 appropriately, and set k1 to zero. Occam’s razor alone therefore argues that this is very likely the answer. Beyond that, it is very typical behavior for a feedback loop, to suppress disturbances from the enforcement of the equilibrium condition. It is so typical that, this is almost surely what is happening.
Once again, everything is trending my way, and you are on the defensive. Emissions are accelerating. Atmospheric concentration is not. And, the lull in atmospheric rate of change matches the lull in temperatures. You would be well advised to rethink your position, before it comes tumbling down on top of you. Because, it is abundantly clear at this time that it will.
With only two adjusted parameters in this model, the projected temperature from Bart’s model matches the observed temperature with little error. Good job Bart!
The next step toward the development of a scientific model would be to determine whether the predicted relative frequencies of the outcomes of the events underlying this model match or fail to match the observed relative frequencies. Here we are hampered by the fact that these events do not exist.
gary gulrud August 28, 2014 at 2:16 pm
Discounting as gibberish that which you do not understand is the end of AGW.
Really, well that isn’t the case here because I understand the subject well enough to have taught it at the graduate level. However in this case there was no use discussing it with you since you couldn’t even get the definition of Kirchoff’s law right!
It should be plain that a quantum mechanical interaction does not change its spots because one looks at it alone. Your discrete spectral emissivity has no scientific foundation whatever.
More nonsense I’m afraid. The spectral emissivity is defined as the ratio of the spectral radiance emitted at a wavelength λ, to the spectral radiance at that wavelength emitted by a black body.
Bart September 2, 2014 at 7:43 pm
Phil:
Let me attempt once more to explain this simply. Your having written this equation may help.
dCO2/dt = k1(dCO2ff/dt) +k2*(T – To)
There is a value for k2 which, with k1 set to zero, very capably fits the data. That value of k2 reproduces both the slope and the variability which match very closely.
If we boost the value of k1 to anything significant, we will have to decrease the value of k2 so that the slope will still match. If we do that, the variability will no longer match.
That’s a problem of your model, you still ignore the Physics of the problem which is the known sensitivity of CO2 to T, to explain the annual growth in CO2 (~2ppm) by temperature increase alone would require a year on year growth of about 0.25ºC. that is an increase of about 4ºC over the last 16 years! I think we all know that that hasn’t happened.
However, that is not all there is. We have to match both the slope, and the variability. The simplest way to do that is to choose k2 appropriately, and set k1 to zero. Occam’s razor alone therefore argues that this is very likely the answer.
No, Occam’s razor says: “Entities should not be multiplied unnecessarily”, it does not say one should remove known parameters.
We know that there is an excellent linear correlation between pCO2 and cumulative FF emissions, you are not justified in arbitrarily ignoring it.
http://www.ferdinand-engelbeen.be/klimaat/klim_img/acc_co2_1900_2004.jpg
Since there has been a year on year growth in CO2 since 1960 your fit implies that To is rather low, what value do you use?
“…you still ignore the Physics…”
No, you are ignoring the data in favor of your pet hypothesis. Data are supreme. You have to fit your theory to the data, not the data to the theory. The data tell us the relationship in the modern era is well modeled by
dCO2/dt = k*(T – To)
Now, take that as your starting point, and then form your hypothesis for why.
“No, Occam’s razor says: “Entities should not be multiplied unnecessarily”, it does not say one should remove known parameters. “
It is not a “known parameter”. You are begging the question.
“We know that there is an excellent linear correlation between pCO2 and cumulative FF emissions, you are not justified in arbitrarily ignoring it.”
It’s a lousy fit, as is seen in the rate of change domain, where the variations do not match, and the long term evolution is diverging. Emissions are accelerating. Concentration isn’t.
“Since there has been a year on year growth in CO2 since 1960 your fit implies that To is rather low, what value do you use?”
You can see it in the plots. E.g., for the UAH data set, I’ve set
dCO2/dt = 0.22*T + 0.14 = 0.22*(T – (-0.63))
To = -0.63
That’s about how much the temperatures have changed since the previous temperature stall starting in about 1945, so not all that much.
I’ve been trying to follow this discussion which appears to attempt to design a model that relates CO2 with temperature. I may have missed key elements of the discussion, but I didn’t see any reference to the feedbacks present in the ‘real’ climate system, like water vapor feedback. In any case, the software depicted in the above video link might be of interest.
Hanzo
No, you are ignoring the data in favor of your pet hypothesis. Data are supreme. You have to fit your theory to the data, not the data to the theory.
Exactly, which is why your pet hypothesis of complete dependence on temperature has to be rejected.
The data shows that during the 90s the ocean absorbed a net ~2.0 PgC/yr, not the temperature dependent outgassing you propose. As Ferdinand has repeatedly told you the data shows that pCO2 increases by ~8ppm for a 1ºC temperature change, your fit requires a much larger coefficient than is observed. So as I pointed out and you ignored: “you still ignore the Physics of the problem which is the known sensitivity of CO2 to T, to explain the annual growth in CO2 (~2ppm) by temperature increase alone would require a year on year growth of about 0.25ºC, that is an increase of about 4ºC over the last 16 years! I think we all know that that hasn’t happened.”
Finally the data shows that there is “an excellent linear correlation between pCO2 and cumulative FF emissions”
In a vain attempt to counter this you do a bait and switch and plot the time dependence of monthly dpCO2/dt and annual emissions due to fossil fuels (not the same thing). The annual data is available why not use it? As explained to you multiple times the growth in pCO2 is not solely due to a single parameter as you appear to think, the major growth is due to FF emissions with modulation by Temperature variation so a matching between T and fluctuations in CO2 is to be expected but the total growth can not be just to T, the data doesn’t allow it.
To make matters worse your fit is not what you claim it to be:
You can see it in the plots. E.g., for the UAH data set, I’ve set
dCO2/dt = 0.22*T + 0.14 = 0.22*(T – (-0.63))
Where you use the monthly T anomaly rather than the actual temperature, why would you expect that to relevant? To get a decent fit between monthly data you’d have to include the seasonal change too (~2ºC).
“The data shows that during the 90s the ocean absorbed a net ~2.0 PgC/yr, not the temperature dependent outgassing you propose.”
The data show no such thing. Again, you are begging the question.
“Ferdinand has repeatedly told you the data shows that pCO2 increases by ~8ppm for a 1ºC temperature change…”
Ferdinand is wrong. The sensitivity is in ppmv/unit-of-time/ºC. That is an empirical fact. Why you (and Ferdinand) cannot wrap your head(s) around this simple observable fact, I have no idea.
“Finally the data shows that there is “an excellent linear correlation between pCO2 and cumulative FF emissions”
It’s a lousy fit, as I stated. It’s basically an observation that two series which are dominated by linear trends over the time interval examined are approximately affinely similar. That isn’t even noteworthy. Indeed, it is a tautology. To get the evidence to convict, you have to examine the fingerprints. And, the fingerprints for the anthropogenic culprit do not match. Those for temperature do.
Indeed, even the superficial match of low order polynomial behavior between virtual accumulation of emissions and measured concentration is currently diverging from affine similarity. Which is why, I suspect, that you cut off the data at 2004, fully a decade ago.
“…you do a bait and switch and plot the time dependence of monthly dpCO2/dt and annual emissions due to fossil fuels (not the same thing).”
This is getting bizarre. Annual emissions are a rate of change in units of mass per year. If I take yearly dCO2/dt, it shows the current lull even more distinctly.
“As explained to you multiple times…”
As you asserted multiple times. Your assertions are wrong. The data prove it.
“To get a decent fit between monthly data you’d have to include the seasonal change too (~2ºC).”
Not when the CO2 data are smoothed over twelve months.
This isn’t even remotely a close call, Phil. In years ahead, you will be embarrassed you took such an adamantly wrongheaded position. On the bright side, you’ll have plenty of company in the dunce corner. Since you have nothing new to add, show no indication of having given thought to the argument, and insist on repeating assertions without foundation, I see no further benefit to continuing this conversation.
Bart September 4, 2014 at 12:10 pm
“The data shows that during the 90s the ocean absorbed a net ~2.0 PgC/yr, not the temperature dependent outgassing you propose.”
The data show no such thing. Again, you are begging the question.
Indeed they do, for example:
M. Battle, M.L. Bender, P.P. Tans, J.W.C. White, J.T. Ellis, T. Conway, R.J. Francey
Global carbon sinks and their variability inferred from atmospheric O2 and δ13C
Science, 287 (2000), pp. 2467–2470
R. Keeling, S.C. Piper, M. Heinmann
Global and hemispheric CO2 sinks deduced from changes in atmospheric O2 concentration
Nature, 381 (1996), pp. 218–221
Takahashi et al.
Deep Sea Research Part II: Topical Studies in Oceanography
Volume 49, Issues 9–10, 2002, Pages 1601–1622
“Ferdinand has repeatedly told you the data shows that pCO2 increases by ~8ppm for a 1ºC temperature change…”
Ferdinand is wrong. The sensitivity is in ppmv/unit-of-time/ºC. That is an empirical fact. Why you (and Ferdinand) cannot wrap your head(s) around this simple observable fact, I have no idea.
Because it’s not true, read up on Henry’s law some time.
Indeed, even the superficial match of low order polynomial behavior between virtual accumulation of emissions and measured concentration is currently diverging from affine similarity. Which is why, I suspect, that you cut off the data at 2004, fully a decade ago.
No that’s the sort of trick you resort to, I just linked to Ferdinand’s plot. Here’s a plot of data up to 2013:
http://www.moyhu.org.s3.amazonaws.com/misc/ghg/m3.png
even better fit if land use change is incorporated too:
http://www.moyhu.org.s3.amazonaws.com/misc/ghg/m2.png
“As explained to you multiple times…”
As you asserted multiple times. Your assertions are wrong. The data prove it.
Actually the data is on my side as shown above.
“To get a decent fit between monthly data you’d have to include the seasonal change too (~2ºC).”
Not when the CO2 data are smoothed over twelve months.
So why do you use the unsmoothed monthly T anomaly in your equation for the smoothed CO2 data?
You take the monthly ML CO2 data, do a 12 month smooth of it and then take the derivative and plot it against the monthly UAH T anomaly (i.e. a sin term corresponding to the seasonal cycle has been subtracted).
You are not treating the fluctuations in the two quantities in the same way, why not detrend the CO2 and then subtract the seasonal cycle from it?
Since you have nothing new to add, show no indication of having given thought to the argument, and insist on repeating assertions without foundation, I see no further benefit to continuing this conversation.
Declaring victory and running away, a sure sign of a lost argument. Come back when you’ve learned something about the data and physics of the problem.
“Indeed they do…”
To the degree they are inferred from proxy measurements, they represent dozens of assumptions piled on top of one another. To the degree they rely on the predetermination that humans are driving atmospheric CO2, they are begging the question.
The direct measurements of CO2 since 1958 and temperatures in that interval contradict it.
“Because it’s not true, read up on Henry’s law some time.”
Wrong. You are imposing predetermined dynamics on the system. That is begging the question. You think that the dynamics ought to evolve according to Henry’s Law, but that is something that has to be confirmed. You cannot just assume it, and then proclaim it is so. This is a complex system. It does not have to behave as you think it should.
You must use the data to confirm your model, and in this instance, they disconfirm it. Your model is too simple, and does not match the behavior of this complex system.
“Here’s a plot of data up to 2013:”
And, as you can see, it is diverging from affine similarity near the end. If you plotted the data in the rate of change domain, you would see that the divergence is quite significant – emissions are accelerating, while concentration is not. Moreover, there is a better model which does not diverge at the end, and that is dCO2/dt = k*(T – To).
“…even better fit if land use change is incorporated too:”
If you torture the data, you can make it confess to any crime. The most direct, most modern, most reliable measurements confirm the dCO2/dt = k*(T – To) model.
“So why do you use the unsmoothed monthly T anomaly in your equation for the smoothed CO2 data?”
Because it is an anomaly, i.e., it has already had the periodic behavior smoothed out.
You’re clueless, Phil. Or, in denial. Or both. Haven’t we said all there is to say? You are arguing in circles, and refusing to deal with reality. What is the point of continuing?
This is no coincidence. You cannot refute it, so you turn a blind eye to it. Good luck with that. You are in for a rude awakening.
OMT: “…why not detrend the CO2 and then subtract the seasonal cycle from it?”
I do not want detrended CO2, I want the derivative. Subtracting the seasonal cycle out is a filtering operation. Averaging over 12 months is a filtering operation. Both filters squash the annual cycle. It really does not matter which you use. The WFT site allows easy yearly averaging, so I use it.
This is grasping at straws on your part. There is no legitimate different way of processing the data which will not lead to the same affine match between the rate of change of CO2 and temperature anomaly.
Bart. could you re-draw your graph using absolute T instead of an anomaly? It might make the relationship clearer.
It is the deviation from normal which drives the phenomenon.
You full well know that the “normal” used in the calculation of the anomaly is bogus.
I would need a clearer statement before agreeing or disagreeing. Let me replace “normal” with “pre-existing conditions”, and hopefully that will addresses your concern. Otherwise, you are going off on a tangent in which I am not interested.