Monday
Oct212013
by 
Bishop Hill
Bishop Hill Diary date: Murry Salby
Oct 21, 2013 Diary dates ;)
Murry Salby, who studies the carbon dioxide budget from his base in Australia, is visiting the UK at the start of November and will give a number of talks - two in London and one in Cambridgeshire.
Details are available at the link below.
Reader Comments (102)
Sorry, to keep repeating this, but does anybody know what happened to Salby's PhD student Evgenia Titova who has apparently vanished without trace
Seems to me her fate could reveal a lot about the reasons Salby was dismissed, especially if she was also dismissed
"...about the reasons Salby was dismissed..."
Does it matter? Some things are evident:
- Dr Salby is a scientist of stature [see his publication record and examine his textbook if in any doubt]
- His pointing out realities such as the following did not go down well with the CAGW priesthood at the university involved.
Anything more about Dr Salby's departure from the university is just fine detail and does not change the big picture.
It's hard from what we know of the university involved, to imagine that Ms Titova would simply be transferred to another supervisor. On the other hand, a research student is not an employee of the university and cannot normally simply be "dismissed". I'd suggest not speculating but simply waiting until the truth is eventually revealed.
http://newsle.com/person/evgeniatitova/8479547
Works for Morgan Hunt
That's all I can get, there's a twitter or facebook link on that page, but all the posts are in Russian.
Assuming it's her....
I hope he's got a return ticket! :-)
Has Salby published his theory yet?
You wouldn't have thought these two quotes were controversial. The main problem seems to be that the IPCC have ignored Henry's law in order to call the ocean a carbon sink - because they can't think of anywhere else (yes that is believe it or not, the official reason). However there have been massive carbon sinks discovered and published about in recent years (Northern forest tree widths, Amazon, deserts) because some researchers bothered to look in the real world rather than the modelised alternative. When these sinks are combined they don't need the ocean to be arbitrarily declared a sink at all and hence they don't have to propose an unphysical mechanism. Arhennius assumed Henry's law would be a positive feedback to any CO2 induced warming.
The thing is, if you assume CO2 comes from the sea then the question is, how much is left for man's contribution. If you say 50% then you have halved the problem. If you say all of it then you are labelled as anti-science - despite this assertion being the only one with scientific underpinnings in a warming world. They still keep Henry's law active of course; only to scare us that sometime in the uncertain future all this CO2 will burst forth at once and cause thermageddon. This contradicts their earlier assertions but that's the way climate science seems to work - a series of contradicting opinions declared as incontravertible "evidence" by the holy writ of consensus - simultaneously telling us we need to act but that it's already too late to act. Sure we'll make no difference whatsoever but we need to show leadership to the 3rd world by bankrupting ourselves.
Hope some of you can make it:
Climate Change: What We Know and What We Don‘t
with Dr Murry Salby at the House of Commons 1.30pm -3.30pm 6 Nov
Book now!
Eventbrite tickets here
http://www.eventbrite.co.uk/event/8872811819/efbnen
@rightwinggit
Wrong Titova, Morgan Hunt is a requirement* agency
[*recruitment agency. TM ]
I've booked. Anybody meeting in a pub for a snack/drink beforehand?
JamesG,
Forgive me, but I find your post a little bizarre. Textbook physical chemistry suggests that, over recent decades, the effect of the increase in partial pressure of atmospheric CO2 (increasing solubility) should be far larger than the effect of temperature change (decreasing solubility). This means that overall the oceans should be acting as a sink for CO2, unless there exists a subtle argument (which I have not seen) to do with equilibration of carbonate and bicarbonate ion concentrations to explain why this is not so. Anyone arguing that the ocean has actually been a net source of CO2 therefore carries the onus of explaining why our understanding of physical chemistry is flawed or too simplistic.
Good news. I would like to learn more about what he has to say. Hoping to see reports from all of these meetings here...
Paul_K,
I agree. While I have not heard Salby's arguments, they do seem to go against well-established phenomena. While there are aspects of climate science orthodoxy with which I take issue, I have been satisfied with the evidence for an anthropogenic increase in pCO2.
Prof Salby's talks in Sydney (2011, 2012) and Hamburg (2013) may be viewed here:
Sydney, 2nd Aug 2011
http://www.youtube.com/watch?v=YrI03ts--9I&feature=youtu.be
Sydney, 9 Sep 2012
http://www.youtube.com/watch?v=ZVCps_SwD5w&list=PLILd8YzszWVTp8s1bx2KTNHXCzp8YQR1z&index=3
Hamburg 18 April 2013
http://www.youtube.com/watch?v=2ROw_cDKwc0
It would be well worthwhile for anyone considering attending his UK lectuers to bone up beforehand to assist comprehension and prepare questions, if any.
Oct 21, 2013 at 2:23 PM | Paul_K
"Textbook physical chemistry suggests that, over recent decades, the effect of the increase in partial pressure of atmospheric CO2 (increasing solubility) should be far larger than the effect of temperature change (decreasing solubility)."
You have made the assumption that ocean content would otherwise be static. There is no requirement that it should be so. If upwelling waters are richer in CO2 than currently downwelling waters, Henry's law says that there should be a temperature modulated pumping action of CO2 into the atmosphere. This is, in fact, precisely what the data show.
Have a ticket for Nov. 6. If anyone wants to meet up before or after, let me know.
Having had a lot of discussions with Bart, who shares the same theory with Salby, here my thoughts:
- Salby and Bart have a theory that a small sustained change in ocean temperature over a baseline gives a constant stream of CO2 into the atmosphere. That is based on the very nice correlation between the temperature anomaly and the rate of change of the CO2 increase in the atmosphere.
- But while the nice fit is completely right for the short term variability, it is largely wrong for the slope of dCO2. The fit in slope is simply because both temperature over the past decades and the dCO2 trend are more or less linearly increasing, thus with the right factor and offset, one can always fit them both. But that is a fit of variables of a different order. In fact, by fitting the temperature anomaly to dCO2, one attributes the full slope of dCO2 to temperature alone, while the slope of dT is zero. The effect of temperature itself on the slope of dCO2 is zero. Thus another process is involved.
- So what makes the slope of dCO2? A good candidate are human emissions: these increased slightly quadratic over the past 50 years:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/temp_co2_acc_1960_cur.jpg
which gives about twice the slope of the observed dCO2:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/dco2_em2.jpg
- Salby and Bart's theory is that a temperature modulated extra upwelling from the oceans is the cause, but that fails a lot of observations: the natural circulation must mimic the increase in human emissions at exactly the same ratio in exactly the same time frame. But that is not observed in the residence time, the 13C/12C ratio, the 14C/12C ratio,...
- Last but not least, the oceans are a net sink for CO2 as observed in a few million measurements over decades. What Salby and Bart suggest violates Henry's law, as with a global ocean temperature increase of 1 K the increase of pCO2 in the oceans is not more than 16 ppmv. Thus maximum 8 ppmv increase is from the 0.5K temperature increase since 1960, the rest of the 70+ ppmv since 1960 is from the human emissions.
- What Salby and Bart forgot is that the ocean-atmosphere exchanges of CO2 is an equilibrium system where both releases and uptakes are directly proportional to the pCO2 difference between the two, according to Henry's law. For Salby and Bart there seems to be no reaction from the increased CO2 pressure in the atmosphere on the in- and outfluxes of CO2 to/from the atmosphere. But that is physically impossible:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/upwelling_bart.jpg
Thus all by all a nice theory only based on curve fitting, without any base in reality...
Oct 22, 2013 at 2:46 AM | Unregistered CommenterBart
Hi Bart,
If you are still following this thread...
I don't think that your plot proves what you think it proves. It is quite possible that it confirms (only) that the interannual changes in atmospheric CO2 are partly controlled by temperature variation - but we knew that anyway from Henry's law.
I think that you believe that it "proves" that temperature controls the rate of change of CO2? The problem is that there exists an almost perfect affine map between dCO2/dT and CO2 over the temperature range considered - with less than a 0.2% error. Hence, your plot may be showing something new but, more likely, it is just confirming the known relationship from Henry's Law. How would you tell the difference?
The correlation between the temperature anomaly and the rate of change of the CO2 increase in the atmosphere didn't come through in my previous message. I have added the trends, because Bart's T anomaly slope is not exactly the same as the slope of dCO2. The integral would give a too high increase of CO2 in the atmosphere.
By changing the factor and offset, one can have an exact match between T anomaly and dCO2, but that gives a smaller amplitude in the variability than what is observed. That is one of the many problems with this theory, as the variability and the slope are supposed to be from the same temperature dependent process. If the variability is mainly caused by the temperature variability and the trend is mainly caused by human emissions, there is no problem at all to agree with every observation...
Oct 23, 2013 at 7:28 AM | Ferdinand Engelbeen
Ferdinand is rationalizing. He does not understand the math. I have been around and around on this issue with him, and do not wish to go through it again with him right now.
Oct 23, 2013 at 8:44 AM | Paul_K
"The problem is that there exists an almost perfect affine map between dCO2/dT and CO2 over the temperature range considered - with less than a 0.2% error."
Using the total CO2 as your measurement obscures detail, because of the low pass filtering characteristic of integration. All that is left is essentially a low order polynomial, really just a trend plus a slight curvature. Human emissions have also been essentially a linear trend with some slight curvature. Do you know how difficult it is to find a superficial affine resemblance between two series dominated by trends with slightly same-sign curvature?
It is not difficult at all. You simply perform a linear regression of the one against the other to find the best fit affine parameters. If their slight curvatures are both the same sign, you will get a fairly good fit. So, basically, the putative correlation between human inputs and atmospheric concentration is based on 50/50 odds. They both have slightly positive curvature, so they appear to be correlated. But, the correlation is not at all extraordinary, and forms no conclusory evidence of genuine cause and effect.
This is the basis of the mistaken correlation asserted between human inputs and atmospheric CO2. They are affinely similar, but so what? It is based on low information signals, and the observed correlation is superficial and spurious. When you take the derivative, however, you reveal the fine detail, the bumps and wiggles which should match if the variables were truly correlated. What we see is that the bumps and wiggles never matched, and the slowly varying apparent affine similarity started diverging in about 1990, and the divergence is accelerating as temperatures begin their natural, cyclical downturn.
Contrast this to the plot I linked above, where every bump and wiggle matches, and continues to match post 1990, and up to the present time. This is the fingerprint which tells us suspect A, human inputs, is innocent, and suspect B, the temperature dependence, is the true culprit.
Oct 23, 2013 at 9:26 AM | Ferdinand Engelbeen
Ferdinand notes here a slight difference in the slopes from linear regression. This is clutching at straws. The trend lines depend very much on the start and stop dates, and the errors or noise in the data. The differences are not statistically significant.
The thing that nails the temperature dependent pump into the atmosphere as being the cause of the observed rise in CO2 is the trend line in the derivative. That is what accounts for the curvature of the total integrated CO2. If is not a free parameter in the affine fit, because the affine parameter which impacts it must be fitted to the variations, the bumps and wiggles.
Ferdinand notes that the trend lines for the CO2 derivative and the affinely mapped temperature do not match precisely over this finite set of data. I maintain that it is simply statistical noise, and the slopes match well enough. But, even if Ferdinand were right, he would have a big problem. The slope of the affinely mapped temperature trend is higher than the slope of the CO2 derivative. So, the human impact would have to be to draw CO2 out of the air, not put more into it.
The small discrepancy is, as I say, more likely due to errors and noise, i.e., measurement errors and outside transient influences which are not captured in the bulk global temperature anomaly measure.
As I said before, one can fit any linear trend line in the original values with any linear trend line in the derivative, but that doesn't prove causation.
Here a simulated comparison between T anomaly and dCO2 (based on the same CO2 increase in the atmosphere and the same temperature trend + sinusoidal variability), fitting all trends, no matter if the origin of the trend is 95% human and 5% temperature or 50-50 or 10-90, only by choosing the right factor and offset:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/sim_dco2_dT_Tanom_95.jpg
http://www.ferdinand-engelbeen.be/klimaat/klim_img/sim_dco2_dT_Tanom_50.jpg
http://www.ferdinand-engelbeen.be/klimaat/klim_img/sim_dco2_dT_Tanom_10.jpg
In all three cases it is possible to match T anomaly with the trend of dCO2. The amplitude of the variability gets smaller, the more that the temperature gets involved, but that may be compensated with some extra reaction of the unknow CO2 releasing process on temperature variations.
Everybody agrees that the short term variability in the CO2 rate of change is caused by temperature variability. But that doesn't say anything about the cause of the long term trend. The short term variability has a periodicity between seasonal and 2-3 years, the trend - if temperature caused - has a period of at least 600 years. Nobody can claim that the short term variability has the same origin as the long term trend: different processes are at work, where the long term trend is near fully independent of the short term one and most probably only slightly temperature dependent, per Henry's law.
And please, if you show a graph of the CO2 emissions and of the increase in the atmosphere, use the same units for both:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/dco2_em2.jpg
By using different units you give the impression that there can't be any influence of the emissions at all, while the variability in the increase of the rate of change is clearly a temperature induced variability in the sink rate of change, not source rate of change.
Further, as said before, an increase in CO2 in the atmosphere will influence the infuxes and outfluxes between the oceans and the atmosphere. That is a dynamic process that currently is far above its equilibrium point. Claiming that there is no influence of the CO2 increase in the atmosphere on input and output fluxes is claiming that Henry's law doesn't count and that there are no dynamic processes at work, only one-way input from the (deep) oceans...
BTW, you are far better in theory, I have a lot of practice in solving problems caused by theories and theoretical models that didn't work in reality...
Bart:
But, even if Ferdinand were right, he would have a big problem. The slope of the affinely mapped temperature trend is higher than the slope of the CO2 derivative. So, the human impact would have to be to draw CO2 out of the air, not put more into it.
I have no problem at all with the slope of temperature, because temperature itself has zero influence on the slope: dT has zero trend, thus doesn't affect the slope of dCO2, while the slope of dCO2(emissions) is twice the slope of dCO2.
The only way that temperature can induce a slope in dCO2 is if it has a non-linear effect on some CO2 emissions, causing a threefold increase in total natural emissions (and sinks) in the period 1960-current (all based on 0.5 K increase in temperature!), completely in lockstep with human emissions. Without affecting any observations like residence time, isotope ratio's etc. Such a process doesn't exist...
Oct 21, 2013 at 1:52 PM | David Jones:
Have a ticket for Nov. 6. If anyone wants to meet up before or after, let me know.
Have a ticket for Nov. 6 too, meeting before will be difficult, arrive with the Eurostar around 12 PM at Paddington, so will be short to be on time (you know the Eurostar...). After is possible, will stay one night, if I find an affordable hotel (not so easy in London...). Some other known people coming in and join us?
"Everybody agrees that the short term variability in the CO2 rate of change is caused by temperature variability. But that doesn't say anything about the cause of the long term trend."
When you match the variability in dCO2/dt with variability in the temperature, you also match the trend. If the temperature variability causes the variability in dCO2/dt, then the trend in temperature causes the trend in dCO2/dt, too. That leaves no room for significant human influence.
You cannot just mix and match - taking the trend out of the temperature and saying the remainder causes dCO2/dt variability, and substituting in the human inputs for the trend. In the first place, Nature has no mechanism for performing such a detrending. In the second, why? The trend in temperature already fits the dCO2/dt trend. It is a gross violation of Occam's Razor to hypothesize some exotic process for high pass filtering the temperature input and low pass filtering the human inputs, blending them together with no phase distortion whatsoever, when the entire thing is already explained by the temperature relationship alone.
Bart:
If the temperature variability causes the variability in dCO2/dt, then the trend in temperature causes the trend in dCO2/dt, too.
As I have shown with my simulation and as the observations show: the variability in dCO2 is as good performed by dT as by T. The perfect match in timing between T anomaly and dCO2 is simply the result of the lag between CO2 and T, which gives a lag of dCO2 vs. dT and a perfect lining of T and dCO2:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/sim_co2_temp.jpg
and
http://www.ferdinand-engelbeen.be/klimaat/klim_img/sim_dco2_dT_Tanom.jpg
Thus the perfect line up of T anomaly and dCO2 has zero predictive power in showing the cause of the slope of dCO2 and thus the increase of CO2 in the atmosphere.
"As I have shown with my simulation and as the observations show: the variability in dCO2 is as good performed by dT as by T."
No. It isn't. You cannot match T and dT. They are 90 deg out of phase with one another. Hence, since dCO2 matches T, you cannot match it with dT - they are 90 deg out of phase with one another.
Look at your plot. your dT/dt is peaking when your T(anom) is going from below to above the trend line. Your T(anom) is peaking when dT/dt is crossing zero. They are out of phase. They do not match.
This is elementary calculus. The derivative of a function leads the function by 90 degrees, and the integral of the function lags it by 90 deg. CO2 lags mean temperature anomaly by 90 deg. Hence, it has an integral relationship with it.
You can't just shrug off phase mismatch as if it means nothing, and can be accounted for by some arbitrary means. In natural systems, there is an inherent linkage between phase response and amplitude response. They are not independent of one another, and the kind of arbitrary phase manipulation you are suggesting can occur without any other impact has absolutely no physical basis.
The inextricable interdependency of the gain and phase of a minimum phase system is known as the Bode gain-phase relationship. You can find information on it via the search engine of your choice. Generally speaking, a large change in phase must be accompanied by a large gain. You cannot simply arbitrarily shift the phase of your function without markedly changing anything else about it.
Bart:
No. It isn't. You cannot match T and dT. They are 90 deg out of phase with one another. Hence, since dCO2 matches T, you cannot match it with dT - they are 90 deg out of phase with one another.
Bart, why is there a phase match between T and dCO2? That is because there is a 90 deg. phase shift between T and CO2. Thus the exact phase match between T and dCO2 is the result of the lag of CO2 after T. CO2 lags T on about all time scales (except for the trend over the past 160 years). That means that the exact match of T and dCO2 doesn't prove that T is the cause of dCO2 and says absolute nothing about the cause of the trend.
Because CO2 lags T, dCO2 lags dT and it is the latter that is the cause of the variability of dCO2, without influence on the slope of dCO2 as the slope of dT is near zero.
Moreover, if you choose the right factor to match the amplitude of the variability in dCO2, then either the slope or the offset are too high compared to dCO2, or if you match the two slopes, the amplitude is too low.
If the slope and the variability are caused by different processes, then there is no problem at all to match the slope of dCO2 and the amplitude of the variability. That is the case if temperature is the main cause of the variability in sink rate and pressure is the main cause of the increase in sink rate, where human emissions deliver the bulk of the increase...
Oct 23, 2013 at 5:59 PM | Unregistered CommenterBart
OK, Bart, I'll try to keep an open mind. I cannot explain your plot showing the post 1990 divergence to my own satisfaction.
At the same time, your theory suggests a suspension or swamping of the effect of increasing the atmospheric partial pressure of CO2, which I find difficult to accept. It also wants an explanation for what did happen to the additional CO2 which we put into the atmosphere. About half of it vanished. If there was no net uptake by the oceans, where did it go and under what form of governance? Land-based bio-regulation seems most unlikely. Additionally, I would point to the fact that high northern latitude CO2 measurements are slightly higher than high southern latitude measurements, and diverging, despite very similar sea-surface temperatures. It seems to me to be difficult to explain this if you believe the oceans are not only a source of CO2, but the dominant source. It is a lot easier to explain if the oceans are a net sink and anthropogenic additions are the main culprit.
I've just seen this debate and hope to get to Professor Salby's presentation in Huntingdon. However, I am deeply sceptical that it offers anything as far as debate about the carbon cycle goes. There is abundant isotopic and elemental evidence that the major contributor to CO2 rise in the atmosphere are anthropogenic inputs. Not least amongst these are : (i) the inventory of anthropogenic outputs, (ii) the O2/N2 ratio variation in the atmosphere and (iii) the d13C and 14C data for atmospheric CO2.
At the same time I take on Bart's mathematics and the close corrleation between dCO2/dt and the temperature and the quadrature relationship between dCO2/dt and dT/dt.
However I would like to propose the following model and perhaps Bart would be so kind to think about it. The key elements of the model are:
(1) The atmosphere and ocean are in disequilibrium with PCO2 in the atmosphere greater than PCO2 in the surface ocean.
(2) There is an anthropogenic input of CO2 into the atmosphere (just humour me for a moment Bart). The rate of input is greater than the transfer flux between atmosphere and ocean such that the CO2 level is rising. The exact mathematical function of the rise is not critical (at least I don't think so).
(3) The rate of net transfer of CO2 into or out of the ocean is inversely proportional to temperature. Thus dCO2/dt correlates with T.
I haven't had a chance to sit down with a piece of paper and write out the equations but I have a hunch that such a scenario might produce the graph that Bart referred to earlier. Your thoughts would be appreciated Bart.
Anyway just a thought and I'll look forward to Murry Salby's talk in a few weeks time.
Oct 24, 2013 at 11:28 AM | Ferdinand Engelbeen
"Because CO2 lags T, dCO2 lags dT and it is the latter that is the cause of the variability of dCO2, without influence on the slope of dCO2 as the slope of dT is near zero."
You do not seem to comprehend that it does not matter. These quantities are inextricably intertwined through integration. The slope of the integral of dT is not near zero. It "perfectly" matches the slope of dCO2/dt when it is scaled to match the variation. And, I put "perfectly" in quotes to mean, as perfectly as it can given measurement errors and external perturbations. Leading into your next comment:
"Moreover, if you choose the right factor to match the amplitude of the variability in dCO2, then either the slope or the offset are too high compared to dCO2, or if you match the two slopes, the amplitude is too low."
You are imbuing these data with too much determinism. These are stochastic signals. You cannot say of them, "this indicates this or that", you have to say "this indicates that it is likely that..." The deviations from perfection are well within bounds of statistical significance. In fact, the match is amazingly good, given the bulk quantities tabulated over many years.
Your point of view is, quite simply, mathematically and physically impossible.
Oct 24, 2013 at 12:01 PM | Paul
"At the same time, your theory suggests a suspension or swamping of the effect of increasing the atmospheric partial pressure of CO2, which I find difficult to accept."
An upwelling of CO2 rich waters will produce a relentless pump of CO2 into the atmosphere. I discussed it a bit here. There is really nothing extraordinary about it.
"It also wants an explanation for what did happen to the additional CO2 which we put into the atmosphere."
Biological and mineral sinks. They are apparently more powerful than is currently presumed.
"Land-based bio-regulation seems most unlikely."
What makes you so sure? Here's a question for you: what is the surface area of all the plant leaves on the Earth? What is the overall limiting factor for plant growth on the Earth?
"Additionally, I would point to the fact that high northern latitude CO2 measurements are slightly higher than high southern latitude measurements, and diverging, despite very similar sea-surface temperatures."
The oceans are not uniform. Some areas are emitting CO2, and others are absorbing it. It is the imbalance between the two which drives CO2 higher.
Oct 24, 2013 at 12:48 PM | Paul Dennis
"Not least amongst these are : (i) the inventory of anthropogenic outputs, (ii) the O2/N2 ratio variation in the atmosphere and (iii) the d13C and 14C data for atmospheric CO2."
This is only narrative. It is a type of scientific reasoning which I detest, which I call Narrative Science. In this paradigm, you make an observation and, if you can think of some seemingly plausible way in which that observation could have come about, and you can get a few others to agree that it is plausible, then that is how it came about. This method of science requires no evidence to back it up, merely mocking of dissenters for not seeing what is "obvious". This is the same type of "science" which, for ages, had authority figures claiming that heavier objects fell faster than lighter ones, and that the Sun revolved around the Earth, and that bloodletting released the "ill humours" from the body so it could heal.
The inventory of anthropogenic outputs means nothing without knowledge of the feedback factors - the so-called "mass-balance" argument is an empty assertion made by people who do not understand feedback systems. There are alternative arguments for how observed chemical and isotopic ratios could have come to be what they are which have not been disproved.
"(1),(2),(3)"
It does not work, because the temperature has been increasing, at least up until 1998 or so. Ocean rejection of CO2 should have increased over that time, so that the rate of change of CO2 in the atmosphere would have been increasing above and beyond the rate at which anthropogenic inputs increased. There would be significant curvature to the dCO2/dt plot. This is not observed.
For more info on a very simple model which is consistent with all the observations, see the link I pointed to above here.
Oct 24, 2013 at 12:48 PM | Paul Dennis
Dear Paul, it is a pity we can't meet in London or Huntingdon...
The lecture of Salby in Hamburg where he explains his theories can be seen here.
I have made aleady a lot of comments on that at WUWT:
http://wattsupwiththat.com/2013/06/21/nzclimate-truth-newsletter-no-313/#comment-1346717
and following parts.
Just like Bart, Salby integrates T anomaly as the cause of the CO2 increase in the atmosphere, thus assuming that temperature is the sole cause of the CO2 increase. Moreover, he calculates that ice cores must be smoothing out all peaks in CO2 over time (factor 10 after 100 kyr), thanks to a huge migration which he calculated theoretically to fit his theory... Not only would that imply below zero CO2 levels during glacials to maintain the average CO2 level, but if there was such a huge migration, that would lead to the fading out of the CO2/T ratio for each glacial/interglacial period back in time. Which is not seen at all...
Bart:
The slope of the integral of dT is not near zero. It "perfectly" matches the slope of dCO2/dt when it is scaled to match the variation.
Again you are matching variables of different order. The integral of dT gives a small increase of CO2 over time of maximum 16 ppmv/K, thus maximum 8 ppmv for the 0.5 K increase over the period 1960-1912, according to Henry's law (which isn't visible in the plot of WFT, as the small offset from zero of dT is not shown). That is all. By using T anomaly you attribute the slope of dCO2 to T, while there is no connection between T and dCO2.
The connection is between dT and dCO2, which gives the variability in CO2 sink rate around the slope with a lag. The slope itself is entirely from the slight curvature of human emissions.
Again, there is no process on earth based on a sustained small temperature difference that may have increased in lockstep with human emissions without violating at last one of the observations...
Further, some remarks on your comments to others:
An upwelling of CO2 rich waters will produce a relentless pump of CO2 into the atmosphere. I discussed it a bit here. There is really nothing extraordinary about it.
You completely forget that any increase of CO2 in the atmosphere will decrease the "relentless" influx of CO2 from the oceans into the atmosphere and will increase the outflux into the cold polar waters. An increase of the CO2 concentration in upwelling waters will increase the CO2 level of the atmosphere until a new equilibrium between fluxes and concentration is reached:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/upwelling_incr.jpg
the same for a temperature increase:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/upwelling_temp.jpg
and both influences are only additional in effect:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/upwelling_incr_temp.jpg
There is no way that a small increase in temperature leads to a sustained increase of CO2 in the atmosphere without a counter effect of the increasing CO2 levels on the in and out fluxes between atmosphere and oceans.
Are you really sure that you understand feedback systems?
what is the surface area of all the plant leaves on the Earth? What is the overall limiting factor for plant growth on the Earth?
No need to know the surface of all plant leaves: the total biosphere (land + sea plants, insects, bacteria, animals,...) is a net sink for CO2 of currently 1 GtC/year, based on the oxygen balance:
http://www.bowdoin.edu/~mbattle/papers_posters_and_talks/BenderGBC2005.pdf
Far from removing all human CO2 + the extra release from the oceans as your theory says.
Rock weathering still is much too slow to have an impact.
The oceans are not uniform. Some areas are emitting CO2, and others are absorbing it. It is the imbalance between the two which drives CO2 higher.
And the observed imbalance is in area weighted average some 7 microatm more CO2 pressure in the atmosphere than in the ocean surface. Thus driving CO2 into the oceans, not the reverse:
http://www.pmel.noaa.gov/pubs/outstand/feel2331/exchange.shtml
Oct 24, 2013 at 9:15 PM | Ferdinand Engelbeen
"Again you are matching variables of different order."
No, Ferdinand. That is not how things work. You are arguing mathematical nonsense.
You do not have to match derivatives to derivatives and so forth. For example, when an object is in an unpowered single dimensional trajectory opposed by viscous friction, its velocity obeys the equation
dv/dt = -k*v
where v is the velocity, and k is the coefficient of friction. It does not matter that you have dv/dt on one side and v on the other. The constant k has the proper units to match the two sides.
dCO2/dt does NOT depend on dT/dt. It depends on T, with a coupling constant which has the proper units. That is what is required to get things to work in phase. You absolutely MUST match the phase. If you are out of phase, then your model is wrong.
"You completely forget that any increase of CO2 in the atmosphere will..."
I forget nothing. A relentless upwelling of CO2 rich waters will relentlessly raise the pCO2 of surface waters, and that in turn will relentlessly raise the pCO2 of the atmosphere. There is no way around it. In a world such as you describe, nothing could raise the CO2 content of the atmosphere, not oceanic upwelling, nor human inputs. It would all immediately build into a pressure which would force the greater part into the oceans.
Your argument is obviously absurd. Think it through more carefully.
"Are you really sure that you understand feedback systems?"
Exceedingly well. You, on the other hand, do not. Out of politeness, I have avoided making personal comments on your demonstrated and embarrassing lack of mathematical and technical skill. Do not make me come after you.
"No need to know the surface of all plant leaves..."
Sure. Not when you have a rehearsed narrative, and a bunch of made up numbers underlying it. You don't need to actually understand anything about the system in that case.
I am going to have to avoid responding to you any further as I am getting heated. You are pushing narrative, Ferdinand. You do not understand physical systems or the mathematics which govern them.
"...with a coupling factor..."
The factor which we observe which couples dCO2/dt to T is more-or-less constant over the period of observation, but it is not necessarily so, and may change over time.
Bart,
dCO2/dt does NOT depend on dT/dt. It depends on T, with a coupling constant which has the proper units. That is what is required to get things to work in phase. You absolutely MUST match the phase. If you are out of phase, then your model is wrong.
The proper units of the coupling constant according to your theory are ppmv/K/year. In the real world, that needs a varying "constant" from zero during glacials to 3 ppmv/K/year for the most recent 50 years.
In reality, CO2 depends of T with a real coupling constant between a transient 4 ppmv/K on short term (seasons to a few years) to 8 ppmv/K on longer term (a few decades to multi millennia), with a variable lag. There is no need at all to match the phase between T and CO2, because T drives CO2 levels. The perfect match of T and dCO2 simply is the result of the phase difference between T and CO2. And there is no need for changing the coupling constant for each period again, it works for all periods in time, taking into account the transient time.
In a world such as you describe, nothing could raise the CO2 content of the atmosphere, not oceanic upwelling, nor human inputs. It would all immediately build into a pressure which would force the greater part into the oceans.
Sorry, but it is your idea that the decay rate of CO2 above equilibrium is very fast. It is not. The observed sink rate is 2 ppmv/yr for 100 ppmv above equilibrium, or an e-fold decay rate of ~50 years. Too slow to remove all human emissions of each year, but fast enough to compensate for halve of a 10% extra oceanic CO2 upwelling or 1 K ocean temperature increase or both in about 10 years time.
Sure. Not when you have a rehearsed narrative, and a bunch of made up numbers underlying it. You don't need to actually understand anything about the system in that case.
Bart, you have a nice theory, but that is not more than a theory. If one and only one observation doesn't fit your theory, then your theory is probably wrong. If many observations don't fit your theory, then your theory is simply wrong. It doesn't help your theory if you dismiss all observations that don't fit your theory as "a bunch of made up numbers", only because you don't like them...
"It doesn't help your theory if you dismiss all observations that don't fit your theory as "a bunch of made up numbers", only because you don't like them..."
No, not because I do not like them, but because they are narratives, not observations. They are storylines, with a few facts consistent with the narrative patched in to give them the patina of verified truth. But, consistency with a few carefully selected observables is not proof of anything.
It only takes one unequivocal contradiction to torpedo a theory. The notion that human inputs are responsible for atmospheric CO2 levels is directly and unequivocally contradicted by the observed dependence of CO2 on temperature.
You don't see it, but you will. The divergence between reality and your narrative is only going to get worse, and it is already very bad.
"There is no need at all to match the phase between T and CO2, because T drives CO2 levels. The perfect match of T and dCO2 simply is the result of the phase difference between T and CO2. And there is no need for changing the coupling constant for each period again, it works for all periods in time, taking into account the transient time."
BTW, this is absolute gibberish.
Bart
Under natural conditions gradual temperature changes do drive CO2.
However there are exceptions. The volcanic CO2 induced recovery from snowball earth conditions, shield volcanoes which produce an initial cooling from aerosols and then a warming from the CO2; and events such as the Pliocene-Eocene temperature maximum which produced a temperature rise on a scale and timescale comparable to present changes, possibly in response to a cometary impact.
I would class CO2 release from an industrial civilisation as another example of a CO2 increase causing a temperature increase.
"Under natural conditions gradual temperature changes do drive CO2....I would class CO2 release from an industrial civilisation as another example of a CO2 increase causing a temperature increase."
You can class it however you like, but these claims are mutually inconsistent, as they define a positive feedback loop which would be inherently unstable.
Bart,
If you start with Henry's Law and assume instantaneous equilibration, then holding partial pressure constant and varying the temperature yields a dCO2/dt curve which is pi()/2 out of phase with the temperature input. The CO2 time-derivative curve leads the temperature curve - basically because it includes a factor of dT/dt.
If we now move to a non-equilibrium system, then I can slide the CO2-time derivative curve exactly into phase with the temperature curve by postulating a response time for the system which is significantly greater than the periodicity of the temperature oscillations. This actually seems like a fairly robust assumption to make. Has this argument been presented and rejected for some reason already? If not, then I will model it against actual data and show you the results, and perhaps drop a line to Professor Salby. If it has already been modeled, then it would save me some brain-time if you could let me know what the objections are to this argument.
You see, I agree with you about where your conclusions take you from your starting premise. You are forced to conclude that CO2 has no warming effect to avoid an open feedback loop. It seems more likely to me that the coincidence of phasing between dCO2/dt and T is brought about by a less radical interpretation.
Bart:
No, not because I do not like them, but because they are narratives, not observations.
If you call the d13C measurements in air and oceans "narratives", you may fool yourself, but that doesn't fool anyone who has some knowledge of natural processes. The same for d14C in the atmosphere, DIC in the oceans, the oxygen balance for the biosphere, the residence time which increases over time etc... All these, except the last, are based on direct measurements and/or calculations from observations.
the observed dependence of CO2 on temperature.
That is only true for the short term variability, temperature is not responsible for the trend in dCO2, that is a different process at work.
The divergence between reality and your narrative is only going to get worse, and it is already very bad.
It is far from bad:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/base_dco2_Tanom_1960-cur.jpg
where the trend in dCO2(observed) follows dCO2(emissions) and still is completely within the natural variability band caused by the temperature variations.
And after two times been objected, now you are willfully misleading readers by using a graph that shows two similar variables by different units.
Oct 25, 2013 at 6:32 AM | Paul_K
"The CO2 time-derivative curve leads the temperature curve - basically because it includes a factor of dT/dt."
Incorrect. I showed the math at the link. It leads to an equation of the form
dCO2/dt = k*(T - Teq)
which is precisely the form of the relationship we see in the actual data.
Oct 25, 2013 at 1:22 PM | Ferdinand Engelbeen
"That is only true for the short term variability, temperature is not responsible for the trend in dCO2, that is a different process at work."
The trend matches perfectly. There is no need to postulate additional human responsibility. I show how the relationship can come about at the link I referenced to Paul above. My math beats your assertion.
"It is far from bad:"
It is very bad. I do not know where you are getting your data, or how you are manipulating it. It does not look anything like the WoodForTrees plots, which is where I got mine for the MLO CO2 data. I got the emissions data from the CDIAC site.
You plot looks like your data are severely aliased, with lots of jagged peaks. Given that you do not even know how to calculate the derivative of a function using the chain rule, I can't say I have a lot of faith in your data handling. When you do it right, you should get a plot which looks like mine.
Try again, with proper tag closure. Preview is our friend...
Oct 25, 2013 at 6:32 AM | Paul_K
"The CO2 time-derivative curve leads the temperature curve - basically because it includes a factor of dT/dt."
Incorrect. I showed the math at the link. It leads to an equation of the form
dCO2/dt = k*(T - Teq)
which is precisely the form of the relationship we see in the actual data.
Oct 25, 2013 at 1:22 PM | Ferdinand Engelbeen
"That is only true for the short term variability, temperature is not responsible for the trend in dCO2, that is a different process at work."
It matches perfectly. I show how it can come about at the link I referenced to Paul above. My math beats your assertion.
"It is far from bad:"
It is very bad. I do not know where you are getting your data, or how you are manipulating it. It does not look anything like the WoodForTrees plots, which is where I got mine for the MLO CO2 data. I got the emissions data from the CDIAC site.
You plot looks like your data are severely aliased, with lots of jagged peaks. Given that you do not even know how to calculate the derivative of a function using the chain rule, I can't say I have a lot of faith in your data handling. When you do it right, you should get a plot which looks like mine.
Oct 25, 2013 at 6:32 AM | Paul_K
"If you start with Henry's Law and assume instantaneous equilibration, then holding partial pressure constant and varying the temperature yields a dCO2/dt curve which is pi()/2 out of phase with the temperature input. The CO2 time-derivative curve leads the temperature curve - basically because it includes a factor of dT/dt. "
Let me amend what I said above, because you did explicitly say "holding partial pressure constant". My claim is that partial pressure in the oceans is not constant. There is no reason it should be. Waters are upwelling and downwelling all the time. Who knows the CO2 content of waters which downwelled centuries ago, went on an unobserved trek through the depths, and recently started reemerging at the surface?
What we see in the data is a curve of the form
dCO2/dt = k*(T - Teq)
Powerful upwelling of richly CO2 laden waters leads directly to this kind of relationship.
Mind you, it is not a slam dunk that this is the way the relationship is coming about. It is only my current operating hypothesis (and maybe others - I am not sure of Salby's take on this). But, it stands apart from the conclusion that humankind is not responsible for the rise in atmospheric CO2. The relationship itself, which is a direct observation, proves that.
I want to highlight the point I made to Oct 25, 2013 at 1:42 AM | entropic man above.
The data absolutely show that increasing temperature increases atmospheric CO2. If CO2 also increases atmospheric temperature, then there is a positive feedback loop.
Atmospheric CO2 increases.
Atmospheric temperature increases in response to increased CO2.
Atmospheric CO2 increases in response to increased temperature.
Atmospheric temperature increases in response to increased CO2.
Atmospheric CO2 increases in response to increased temperature.
Atmospheric temperature increases in response to increased CO2.
Atmospheric CO2 increases in response to increased temperature.
Atmospheric temperature increases in response to increased CO2.
Atmospheric CO2 increases in response to increased temperature.
Atmospheric temperature increases in response to increased CO2.
Atmospheric CO2 increases in response to increased temperature.
Atmospheric temperature increases in response to increased CO2.
Atmospheric CO2 increases in response to increased temperature.
...
On, and on, and on, in an ever upward spiral. Until some moderating influence asserts itself, there is no end to this progression.
Some would claim, well, temperatures decrease less and less because of SB radiation. But, that doesn't work, because the system
dCO2/dt = a*T
dT/dt = b*CO2 - c*T^4
for factors a, b, and c, is still unstable for a and b positive. One of the factors a or b has to go to zero for stability to be achieved. Since we can see directly that a is not going to zero, we must conclude that b, the sensiticity of temperature to CO2, is essentially zero. This is why natural factors have combined to create the "pause", despite the fact that CO2 has roared ahead. CO2 has either net zero or negative (as in a negative feedback) impact on surface temperatures. It must, because otherwise, the Earth would have become like Venus ages ago.
This does not say that CO2 has no effect on surface temperatures at all. It merely says that the effect stalls at some point to achieve stability, and we are at such a point in the Earth's history.
You plot looks like your data are severely aliased, with lots of jagged peaks. Given that you do not even know how to calculate the derivative of a function using the chain rule, I can't say I have a lot of faith in your data handling. When you do it right, you should get a plot which looks like mine.
Bart, your plot uses different units for the same type of variables. That gives a complete false impression of the ratio between the variables. That is what I tried to show.
As the human emissions are only known as yearly emissions I used the yearly averages of the rate of change in the atmosphere too. That gives a less nice plot that the 12 month moving average, but it shows that the dCO2 trend derived from human emissions (but should be derived from the trend of the CO2 levels in the atmosphere above equilibrium) still is within the variability of the temperature influence around the trend.
"As the human emissions are only known as yearly emissions I used the yearly averages of the rate of change in the atmosphere too."
So, basically, you are taking out data which would tend to make your fit look bad.
So, basically, you are taking out data which would tend to make your fit look bad.
To the contrary, if I use the 12-month moving average of the MLO rate of increase, the trend of the emissions * 0.53 (which is about the "airborne fraction" over the past 160 years) is even better inside the temperature induced variability and temperature still may have very little influence on the trend:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/dco2_em3.jpg
The Cediac emissions data are in megatons carbon/year. You can transfer the plot of emissions and CO2 increase per year to the same units as 1 GtC = 2.1 ppmv increase in the atmosphere, which gives the above plot.
Powerful upwelling of richly CO2 laden waters leads directly to this kind of relationship.
As said many times before: an extra upwelling from deep ocean waters, either in concentration or mass leads to an increase in CO2 in the atmosphere which gives a negative feedback, suppressing the extra upwelling and increasing the downwelling. That results in a new dynamic equilibrium with a decay rate of ~15 years, not in a continuous constant inflow without any effect from the increase in the atmosphere.
Moreover, such an upwelling may be the cause of a huge increase of CO2 in the atmosphere, but that is hardly influenced by a temperature increase, so you are attributing the effect of an increase in upwelling to a temperature increase which has not much to do with it...