There has been an interesting discussion concerning the airborne CO2 fraction at WUWT recently which is still ongoing. Our friend ‘Bart’ found an close correlation between the surface temperature of the southern hemisphere and the rate of change of the airborne co2 level, and posits the hypothesis that the source of the increase is due to natural co2 sources dependent for their rate of emission on temperature, rather than human emission of the trace gas through the combustion of fossil fuels. I found that the correlation is even tighter with the UAH global lower troposphere data, see fig 1 below.
This is clearly a very good fit, notwithstanding a couple of aberrations around 1992 and 2010, which we’ll discuss below.
Something worth noting immediately is that whereas the actual changes in levels of CO2 in the atmosphere always lag behind changes in temperature by nine months or so at the annual scale, the rate of change of co2 levels sometimes lags and sometimes leads the changes in lower tropospheric temperature. There is not a simple relationship here.
what clues might we gain by looking at the two aberrations in the relationship visible around 1992 and 2010?
These were both El Niño years, but we see no similar divergences during the big El Niños of 1988 or 1998. So should we discount this phenomena as the possible cause? Not yet, because El Niño and its aftermath varies in terms of its effect on ocean circulation. Bob Tisdale may want to weigh in on this question. Bob noticed step changes in global surface temperature occuring following the El Niños of 1988 and 98, but not in 1992. The picture is likely complicated by the eruption of Pinatubo, but it should be noted that LT and co2 ROC go their separate ways for a while well before Pinatubo erupted in 1991 (june 14) and before Dubbi erupted in Eritrea 20 years later to the day.
What else might be involved? The solar cycle was just past maximum (and spiking) when Pinatubo erupted, and just past minimum when dubbi erupted, so in both cases the changes in solar activity levels had recently reversed. Still nothing very slam dunk here though. Another possibility is that the very large source of co2 outgassing, the volcanic soils, got less sunshine on them during those periods, or that major blooms of plankton occurred, absorbing more co2 than usual. We might just have to accept that we don’t know enough about the carbon cycle yet to nail this one, but I hope further ideas will arise in comments to help us piece this together.
Despite those uncertainties, I think the correlation Bart has found and the recent discovery of the much larger than previously estimated volcagenic sources have moved this debate along, and a serious reassessment of the arguments supporting the idea that human emissions are largely responsible for the rise in co2 seen since records began at Mauna Loa is in order. The closeness of the correlation doesn’t appear to leave much room for sources which are not temperature dependent.
Another conversation which can be had here is the question of how well the rate of co2 change serves as a proxy for checking the veracity of various global temperature indices… To have a play with this, try this URL and then change the dataset used, along with the offset if required to get the best fit. No sneaky detrending allowed. 🙂
UPDATE: Contributor Les Johnson has kindly sent me his spreadsheets from a similar study he did in 2010. You can download them below.
Tallbloke's Talkshop 
Tallbloke: I did the same thing in 2010, with both HadCRUT and UAH global temperatures. The UAH/CO2 rate has high correlations: 0.489 for all raw data; 0.668 for 1979-2010 data, and 0.970 for 1998-2010 data (with a 24 month MA). Even the raw data post 1998 is high, at 0.75.
The best fit is with a 9 month lag between temps and CO2.
If you want, I can send the spreadsheets, for you to post.
[Reply] Thanks! Email coming your way.
The HADCRUT data vs CO2 delta actually shows a divergence before the 1991 Pinatuo eruption, starting in early 1990. The Chichon 1982 eruption also shows up early. This is probably a function of the lag time used, vs 24 month moving average.
Either that, or falling CO2 rates of increase predict volcanic acitivity…. CO2, is there nothing it can’t do? 🙂
On the way, Roger.
Hi Les, there is no lag built into this plot, so are you saying one should be introduced to allow for the effect of the 24 month moving average, or the ~nine month lag of co2 behind temperature, or both, in order to ‘calibrate’ the rate of change of co2 vs temp data plot?
Sorry, I shoudl have prefaced the last post, that it wa sin refernce to my sheets, not the one here.
Yes, I believe that you need to introduce a lag, to get a better match. Temperatures go up, then 9 months later, CO2 follows.
I put a lag in my sheets. I found 9 months the best, as did Humlum et al.
Roger: Let me know if you got that email. The mail with the spread sheets shows up in my “In Box”, rather than my “Sent”.
It has your email though, its just not in the right place.
Thanks Les. I’ve added trends to the Woodfortrees plot and adjusted the offset slightly for easier comparison with your spreadsheet.
http://www.woodfortrees.org/plot/esrl-co2/derivative/mean:24/from:1979/plot/uah/scale:0.15/offset:0.143/plot/esrl-co2/derivative/mean:24/from:1979/trend/plot/uah/scale:0.15/offset:0.143/trend
Yes, I agree. I haven’t updated mine since 2010. I would also like to do a SST sheet, but that will have to wait a bit. Also, note that one sheet is HADCRUT, and the other the UAH lower troposphere.
I am in the processing of moving to Scotland, so not much likely to happen from me for a bit…..
One thing I notice in all the charts, is that the slopes are quite different, with the temperature anamoly rising much faster than the delta CO2.
As CO2 lags temperature, this would indicate that you need more temperature to get an equivalent volume of gas released from (the oceans, an unknown sink, Area 51).
No problem, interested parties can build on what you’ve done, thanks a lot for sharing data.
Heh, Area 51 🙂
Maybe the trend difference despite the accelerating human emission is another indication that the culprit is not human emissions? I guess we can’t be sure without knowing more about the sources and sinks. It’s a not unreasonable assumption that the Earth’s biosphere is relatively starved of co2 at the moment. In geological terms it’s near a historic low point. So you’d expect the biosphere to expand and gobble up co2 not long after it becomes increasingly available. Maybe that’s why the trend in rate of change is lower than the temperature trend.
Roger: Yes, thats my guess. The sinks are probably plants (up 6% globally in the satellite record), and the ocean uptaking nearly the rest through an increase of CO2 partial pressure.
This paper suggest that the uptake of Anthro CO2 is about 46%, and this ratio has not changed since 1850. There is no trend.
My take is that there must be a total lack of feedbacks in the sinks, as the system does not seem to get greedy, and over consume, or get full and stop consuming CO2.
http://www.agu.org/pubs/crossref/2009/2009GL040613.shtml
Roger: The recognisable variations in the 1st Deriviative of CO2 are related to sea surface temperatures anomalies, not lower troposphere temperature anomalies. It’s a function of–I believe the term is–outgasing from the oceans. There are a number of papers that discuss its relationship with ENSO and NOAA may have a webpage about it.
Regards
Bob: Thats why I want to do a SST comparison. Using HADCRUT global and UAH lower Troposphere gives a pretty good match as it is, especially after 1998. Its nearly unity.
If it gives a better correlation to SST, especially using raw data instead of smoothed, then its probably mostly outgassing.
Roger Andres: If you go to the Mauna Loa CO2 sheet in my files, you will see exactly the same numbers. Whats your point?
Or to the Wood For Trees data. You do know we are looking at the first derivative of the CO2 increase, right?
Les Johnson:
My point is that the red line on the graph suggests that the increase in atmospheric CO2 since 1900 is mostly if not entirely a result of anthropogenic carbon emissions. However, I would be delighted if someone could prove me wrong.
Roger A.,
I thought Professor Murry Salby had found that mankind’s contribution to the increase of co2 was just 4%
Roger Andrews: your
My point is that the red line on the graph suggests that the increase in atmospheric CO2 since 1900 is mostly if not entirely a result of anthropogenic carbon emissions. However, I would be delighted if someone could prove me wrong.
The increase is almost certainly due to anthro emissions.
What we were discussing was why the increases in rate in atmospheric CO2 seems to to follow temperature. And why the amount of CO2 in the atmosphere only increases by about 46% of emissions.
Please pay attention.
J. Martin
I believe Salby “found” that mankind’s contribution to the CO2 increase was about 20%, but it took him an hour and numerous slides to explain how he figured this out. The graph I present above satisfactorily explains how mankind caused all of the CO2 increase and occupies only half a page in portrait mode. 😉
Les Johnson:
I’m on your side, OK? We may in fact be the only two people on this blog who believe that “the (CO2) increase is almost certainly due to anthropogenic emissions”.
Roger Andrew: LOL. Ok, point taken. My apologies for being a wanker in the last post.
Guys: It took me a lot more than half an hour to wade through the argumention on the linked thread at WUWT. Ferdinand Englebeen was on good form, but Richard Courtney’s argument seemed more cogent to me. Basically, Richard says we don’t have enough evidence to know what proportion of the increase belongs to us, and that Ferdi’s mass balance and d13/d12 arguments are spurious. I think he’s correct in this.
Bart presented some interesting algebra which I had trouble following, and claims the same as Richard, and also claims that the closeness of the relationship between dco2 and temperature is good evidence that there isn’t much room for a source big enough to be important which isn’t temperature dependent.
I weighed in with the new empirical evidence from Italy and pointed out that the IPCC ‘preferred estimate’ of 0.26Gt worldwide for volcagenic emission is ludicrous, considering the fieldwork which concludes that there is 9Gt annually from central Italy alone, and that the rate of emission from volcanic soils is likely to be temperature dependent and affected by the reduced cloud cover (and therefore increased sunshine hours) in a way which fit’s Bart (and Les’) plots. Geologist Tim Casey says we can’t discount the possibility of volcagenic outliers we haven’t found which have a isotopic signature similar to fossil fuels. The isotope argument is shaky in other ways too: the C3 C4 plant argument for example is not uncontroversial.
So what I’m saying with this post is that science has moved on since it was thought that this issue was ‘settled science’ (Cardellini et al 2011 on soil emission, and the 2012 Spanish, Chinese and global cloud cover studies), and that it’s time to reassess the evidence and the conclusions which have been drawn.
Salby is a clever guy, and if he says only 20% of the increase in airborne co2 is from human emission, I’m willing to pay attention and try to understand how he has reached that conclusion. So if Roger A would like to summarize the salient points from Salby’s study which lead him to the conclusion, I’m all ears.
Roger Andrews says:
September 8, 2012 at 4:41 pm
http://oi46.tinypic.com/bimis5.jpg
Last time we took a look at this graph I came to the conclusion that the 5 year residence time is a cherry pick. 😉
Les Johnson says: “Bob: Thats why I want to do a SST comparison. Using HADCRUT global and UAH lower Troposphere gives a pretty good match as it is, especially after 1998. Its nearly unity.”
As you know, HADCRUT data includes sea surface temperature data but it is not sea surface temperature data. So if you’re planning to use HADSST2, it has a problem: a significant upward shift in 1998 when the Hadley Centre spliced two source SST datasets together. Your best bet is to use satellite-base Reynolds OI.v2 data. It’s available through the KNMI Climate Explorer in easy-to-use form.
I’ve also looked at the Mauna Loa CO2 and HadCrut temperature data, and found essentially the same results. It is easy to find many “superb” fits of temperature to the first derivative of CO2 concentrations. Much more so than the other way round.
[The free software “Eureqa” [http://www.nutonian.com/] from a Cornell University spin-off, allows such curve-fiting using differential equations and many other things, more easily than with Excel. 🙂 I posted that link once before, but I think Tim Channon was less than impressed. However, I am well aware of the perils of an idiot with some squiggle-matching software that ought to come with a scientific-health warning. Otherwise I would not be as sceptical about the IPCC models as I am.
It’s worthwhile as a thinking aid, if for no other reason that it is an excellent tool with which to experience the seductive dark side of computer modelling. That’s a rabbit hole I’ve been down, but at least I didn’t get stuck in the door after eating too much honey, like Winnie-the-Pooh did 🙂 ]
I would add that in one respect we are also lucky, in that since the inception of the Mauna Loa CO2 record we have the 1991 Pinatubo eruption that appears to have produced such a pronounced inflection on CO2 rise.
I would love to know if there is some more detailed frequent CO2 data available either side of the event.
Tmperature varies more in the Northern Hemisphere than the Southern Hemisphere due the damping effect of a greater proportion of ocean in the South, which means that the global average peaks in July or August.
Carbon dioxide levels vary annually as the proportion of photosynthesis and respiration change. In the northern Hemisphere Summer photosynthesis on land removes CO2 from the air faster than respiration replenishes it, so minimum C02 occurs at the end of the Northern Hemisphere growing season in September. It then increases as Southern Hemisphere photosynthesis is less than worldwide respiration.
Since both photosynthesis and warming are functions of solar insolation, I suggest that CO2 and temperature are both responding to the seasonal variation in light and heat input, rather than to each other. The phase difference is due the maximum CO2 occuring in the Northern Hemisphere Spring and maximum temperature in midsummer.
TB:
“Last time we took a look at this graph I came to the conclusion that the 5 year residence time is a cherry pick.” Sure it is. But you have to admit it’s a hell of a good one. And a straight-line CO2
decay over ten years fits ever better. 🙂 🙂
Back later with more
“Tmperature varies more in the Northern Hemisphere than the Southern Hemisphere due the damping effect of a greater proportion of ocean in the South,…”
Remember that there is another asymmetry to be considered: The orbit of the earth about the sun deviates significantly from circular.
It does indeed. We are in a waning interglacial, with the Northern Hemisphere still receiving more energy than the South.
Earth is at perihelion in northern winter, but this doesn’t mean the NH receives more energy than the south, just that it’s winters and summers are more moderate.
You may well be right. The glacial/ interglacial cycle for the last few hundred thousand years correalates best with orbital eccentricity. A Winter perihelion would limit minimum winter temperatures and limit snow accumulation by increasing winter insolation without necessarily increasing the overall total for the Northern Hemisphere.
If I was constructing a model of the carbon cycle from scratch, I would not ignore the-elephant-in-the-biochemical-room which is the role of carbonic anhydrase in the 12C/13C isotope ratio.
In my view the direct focus on photosynthesis is probably mistaken. Close to 100% of living organisms express this enzyme in large amounts as it is used to control pH and to enhance the rate of CO2 absorption, excretion, and exchange via the CO2 + H2O=H2CO3 equilibrium in water, which is slower than a diffusion-limited chemical reaction.
The ‘normal’, thermal, uncatalysed reaction is slower by about a factor of 10e-7. This reaction takes place at all times over most of the planet. Just because there are no visible forests at the ocean surface, it doesn’t follow that they are not under the influence of the agents of the biosphere.
All oceanatmosphere CO2 diffusion exchange must pass through the boundary layer at the surface, and much carbonic anhydrase is not limited to intracellular equilibration. Carbonic anhydrase is actively secreted in to the surrounding milieu be it the ocean or the droplets of rain/dew on the surface of vegetation or soil particles. This reaction happens in locations and times of photosynthesis and decay of organic material, during the daytime and night time, on land and in the oceans and does not require any significant input of energy.
I’ve not yet put any flesh on these bones, and maybe the appropriate efforts have been made to address this issue elsewhere. But I haven’t seen it yet. Kind of hoping that someone else will do it because it’s significant work to do properly, and there is no shortage of other chemists/biochemists who are equally competent to do it if motivated… So I’ll wait until at least after AR5 [and references] is published before I consider putting in that effort.
I suspect more than one milankovitch cycle is involved. 65N insolation correlates well with glacial/interglacial cycles, so this would involve precession too. We’ve done a couple of in depth posts here on this if you want to have a look for them.
You can google this term to help find them
site:tallbloke.wordpress.com milankovitch
With axial tilt varying on a 41,000 year cycle, precession varying over 26,000 years and eccentricity over 100,00 years the glacial cycle matches eccentricity better than the others, despite its lower apparant effect . More work to be done here to improve our understanding.
Over a Milakovich Cycle timescale the link between CO2 and temperature starts with the temperature rise releasing CO2 stored in tundra and water, followed by a positive feedback as each increases the other until approximate stability is reached. CO2 certainly follows temperature during the initial warming stage, though it gets hard to separate cause and effect once the feedback starts. The CO2 in past cycles has mostly oscillated between 200 and 280ppm, while temperature varied 5C with occasional excursions. The change would correalate as about 16ppm of CO2 per degree C, whichever was leading.
On an annual basis the effect of eccentricity is small compared with the primary effect of axial tilt, and both temperature and CO2 are probably seasonally driven.
Over the last 120 years something unusual seems to be happening. Increased CO2 from 280 ppm to 390ppm, accompanied by a temperature rise of less than 1C is outside our civilization’s, and possibly our planet’s, recent experience.
I incline towards the cAGW interpretation for recent changes.
Heh, I’ll be back, as Arnie used to say. 😉
The title of this thread is “Is the airborne CO2 fraction temperature dependent?” So let’s compare some actual CO2 and temperature data.
The first graph compares ICOADS global SST anomalies with the Mauna Loa CO2 record, with both records 12-month smoothed to remove seasonality. The scales are adjusted so that CO2 visually tracks SST.
Could the highly erratic +/- 0.4C increase in SST since 1958 explain the remarkably regular +/-80 ppm increase in atmospheric CO2 over this period? I don’t think so, but others may see it differently.
The second compares monthly changes in SST and CO2 using the above data, with CO2 advanced by 7 months to allow for a 7 month global average lag relative to SST:
Here we see a respectable direct correlation between changes in CO2 and SST (R=0.66) between 1965 and 2000 (although not before 1965 or after 2000), thereby demonstrating that the airborne CO2 fraction is to some extent temperature dependent. However, a regression fit gives an increase of only 3ppm CO2 for each degree C increase in SST, which is far too small to explain the observed +/-80 ppm CO2 increase.
Note that we get substantially the same results when we use SAT or TLT rather than SST.
Entropic Man: The precession cycle actually varies between around 19k and 26k years and similarly, axial tilt isn’t regular either. This is partly because the depth of ice ages varies, which affects the mass balance of accreted ice and thus the polar motion of the Earth’s axis. So as you say, much more work to be done. But the larger point in your comment was about co2 amplifying the initial temperature change. I doubt it myself, partly because you’d expect the lag of co2 behind temperature to be different at the top and bottom of the glacial/interglacial temperature range due to the logarithmic nature of the alleged forcing. This isn’t observed so far as I know.
Roger A: nice graphs thanks. But are we comparing apples with oranges here? Wil your monthly differences match the derivative function used in the head post’s fig1?
It’s interesting that the aberration between LT UAH data and Mauna Loa around 1991-3 isn’t apparent in the sst comparison. What is that telling us? Is there a spike in global OLR at that time that is bigger than in the ’88 and ’98 El Nino’s?
TB.
Although I’ve looked into the Mauna Loa co2 data some time ago, I didn’t find any correction for changing condensate and precipitation levels to up-wind regions that may affect the data as a ‘global reference’. Do you know if this exists?
I’d like to add to Michael Hart’s post, but for atmosphere. Mauna Loa/Hawaii climate:
http://en.wikipedia.org/wiki/Hawaii#Climate
There’s a lot of rain there and the science behind precipitated water is that it’s ‘pure’ (without contaminants). However, co2 is the first ‘gas’ to diffuse into pure water and can be considered to be virtually the ‘only’ ionic contaminant above a 4M ohm/cm^2 resistance (current resistance being taken as an inverse measure of ionic contamination [the greater the resistance, the lower the ionic contamination]). Just a point in physics. 🙂
Any change in precipitation ‘up-wind’ of the Mauna Loa site, or change in wind direction, for a ‘long time-scale’ (El Nino/La Nina, etc.) will alter, IMHO, the bulk residue of co2 that ‘has/hasn’t’ been ‘washed out’ of the atmosphere for the down-wind region.
Could this effect be an intermittent effect that upsets the graphs during an aberration excursion?
Best regards, Ray.
TB
“Wil your monthly differences match the derivative function used in the head post’s fig1?”. The CO2 plots both show changes in CO2 and as far as I can see they are the same. The temperature plots are different because I compare CO2 changes with SST changes while fig 1 compares them with absolute TLT values. I think this makes my plot less apples and oranges, but I don’t insist on it.
The reason you see differences in the early 1990s is that Pinatubo caused (?) about 0.5C of TLT cooling but only 0.1-0.2C of SST cooling. Not sure what that tells us.
The impact of Pinatubo on CO2 is also curious. It seems to have shifted the CO2 record bodily downwards without affecting the rate of CO2 increase. I’ve used regression lines to illustrate this in the graph below:
Ray:
Your question about whether the Mauna Loa CO2 record is a reliable “global reference” keeps coming up in one form or another, so I thought this might be a good time to see how it fits in with CO2 records from other parts of the world
So here’s a plot that compares it with 13 other CO2 records from stations that extend from Alert at 82 degrees N down to the South Pole, with Northern Hemisphere records shown in red, Southern Hemisphere records in blue and Mauna Loa in thick black.
Based on these results I would guess that the Mauna Loa record overestimates the increase in global average CO2 by about 4ppm since 1958, but as a practical matter this isn’t going to make much difference.
Note also how CO2 runs consistently higher in the Northern Hemisphere than in the Southern (the pole-to-pole difference is about 4 ppm). This could be because more CO2 is emitted in the NH, but I think it’s because there’s more CO2-absorbing ocean in the Southern Hemisphere. As shown in the plot below there’s a pretty good correlation between atmospheric CO2 levels and ocean extent in ice-free latitudes:
Over on the WUWT thread STephen Wilde and I have been battling Ferdinand Englebeen’s logic, resulting in this concise and clear summary from Stephen:
Stephen Wilde says:
September 9, 2012 at 3:27 pm
Ferdi says:
“Thus on average, the human contribution is less than 5% of the natural flows. No wonder that the human contribution isn’t visible in the AIRS data, simply because it is too small, compared to the natural flows.”
CO2 being a well mixed gas the AIRS sensors are actually displaying very small variations so my understanding is that they do reveal regions involving 5% variations or less.
http://airs.jpl.nasa.gov/AIRS_CO2_Data/About_AIRS_CO2_Data/
“The high spectral resolution and stability of AIRS allows a measurement accuracy between 1.5 ppm and 2 ppm, making it ideal for mapping the distribution and transport of carbon dioxide levels in the free troposphere.”
2 ppm is less than 5% of 390 is it not ?
Ferdi says
“humans add 8 GtC/year. In average the measured increase is 4 GtC/year. Thus some 4 GtC/year extra must be sequestered somewhere.”
Or:
i) The entire 8GtC/year is rapidly absorbed locally by a biosphere response that only occurred because of the presence of that human CO2
and at the same time:
ii) The ocean sources are running ahead of the biosphere sinks to the extent of about 4GtC/year as a result of decreased cloudiness, wider tropics and more sunlight into the oceans.
Ferdi says:
“The net result must be that the total natural cycle is a net sink for CO2, not a source. Thus nature doesn’t contribute to the observed CO2 increase…”
Or:
The total natural cycle is currently a source because of increased solar input to the oceans and humans don’t contribute to the observed CO2 increase because it gets sequestered locally by an enhanced local biosphere response.
Now we do know that there was reduced cloudiness and wider tropics during the late 20th century and we know that warmer water holds less CO2.
Is it actually proposed that the increased sunlight into the oceans during that period had no effect ?
Or that the effect was actually negative so as to absorb half the human output ?
I don’t think so.
tallbloke says:
September 10, 2012 at 7:30 am
Very cogent Stephen, I agree. I think Ferdinand is making several unsupportable assumptions in his argumentation, and keeps repeating them after they’ve been pointed out to him. A couple of points worth flagging up again:
1) The stomata data from Tom van Hoof agreed well with the Greenland cores *before* they got recalibrated to match the Antarctic cores. This suggests co2 variation is considerably higher than the currently preferred estimates. This means that Ferdi’s Henry’s law argument could be out by a large factor.
2) The volcagenic estimate preferred by the IPCC of 0.26Gt/year globally is based on notoriously unreliable ways of treating data which contains a lot of uncertainty due to inadequate sampling. Recent direct empirical measurement suggests central Italy alone is degassing 9Gt/year. The Mass balance argument is out of the window.
3) If the sources are much (much) bigger than previously thought, then so are the sinks. Since we don’t know what their relative contributions to absorption are, the isotope ratio argument is out of the window too.
4) Since the cloud data as measured by ISCCP shows a bigger drop in the cloud over the tropics than the high latitudes, it is to be expected that the loss of shade in the warm areas will have a bigger effect on the outgassing of co2 than on its re-absorption in the Antarctic southern ocean.
5) Since the fall of co2 behind temperature change at the end of interglacials lags the temperature by as much as the lag of the rise of co2 at their beginning, it is clear that major sinks and sources take a long time to react to temperature change. Clearly, the outgassed co2 is going to hang around in the atmosphere at the end of a warming period of whatever length, before the system restores the balance through re-absorption.
Incidentally, if the warming effect claimed for co2 were real, then due to the approximately logarithmic response of temperature to co2 this should mean the lag at the end of the interglacial will be different to the lag at the start, but so far as I know, it isn’t.
This may be of interest.
http://carboncycle2.lbl.gov/resources/experts-corner/annual-cycles-of-atmospheric-co2-concentration.html
Thanks E.M.
I disagree with the interpretation, but the last graph is what Stephen Wilde and I were looking for.
As Stephen points out, the new AIRS data shows that most emissions are coming from the downwind side of the oceans in the tropics. This indicates that reduced cloud cover since the 1960’s, according to the new Chinese and Spanish Studies, has caused the tropical oceans to outgas a lot of co2 in th warm plentiful sunshine, and the cold antarctic southern ocean sinks haven’t caught up. Result: higher atmospheric levels.
Meanwhile, the AIRS satellite doesn’t find any significant plumes of co2 downwind of big conurbations even though it easily has the resolution to do so, indicating that enhanced growth of biomass caused by the ready supply of human emissions is cancelling out the human contribution.
Strong evidence for our hypothesis, thanks again.
Your hypothesis would predict that most of the CO2 increase, and therefore the maximum concentration, would occur in the Tropics. The biological hypothesis descibed in my link would predict that the maximum concentration would occur at high latitudes in the Northern Hemisphere.
Watch this video.
E.M. No, what we’re saying is that the tropical oceans is where most of it gets released, not where it seasonally accumulates in the following months. And your video is a composite of AIRS data overlayed with mauna loa co2, complete with scary red increasing as the co2 level rises a whole 6 parts in a million of the atmosphere.
STephen Wilde uses this graphic of July AIRS data
and make these observations:
All the highest concentrations are downwind of warm water.
The Mediterranean gets very warm in summer so you can see the plume across the Middle East.
Australia gets CO2 from the ocean between it and South Africa.
South America gets CO2 from the Pacific upwind.
Western USA from the Pacific, upwind.
Southern Asia gets CO2 from the Indian Ocean, upwind.
There is a plume of CO2 downwind of the warm Gulf of Mexico.
and so on.
There is little or no significant excess CO2 above or downwind of major population centres such as Western Europe or the North Eastern USA.
The relatively low CO2 quantities above the equator are due to the clouds and rain of the Intertropical Convergence Zone.
The two main bands of higher CO2 concentration are under the subtropical high pressure systems in each hemisphere where most sunshine gets into the oceans to warm the sea surfaces.
What is the mechanism by which the CO2 you suggest is produced by tropical seas is transferred to, and accumulates, at high latitudes only in the Northern Hemisphere?
I phrased badly, sorry. Meridional flows in the atmosphere ensure co2 rapidly becomes a ‘well mixed gas’ but local differences between the hemispheres mean the annual swings in co2 level are greater in the north.
It isn’t only in the northern hemisphere that co2 is most strongly emitted from the ocean, but under both subtropical zones where cloud is at a minimum and the Sun heats the ocean surface strongly. The ITCZ is cloudy near the equator and the ocean surface stays cooler there.
Your video masks this by radically changing the colour to eyecatching reds for a tiny 2ppm difference in concentration. Furthermore, a lot of co2 is emitted from decaying biomass on land and land is predominantly in the northern hemisphere. But this is only for a brief spell in the year. Because of the way the data is presented in your video, the eye is distracted north and ‘blinded’ by the glare of the strong colouration of those brief intense periods.
I cant help the colour code, which is often used to show greater or lesser values in graphics. Try not to get distracted by the colours and the conspiracy theories; concentrate on the concentrations.
The data shows increased CO2 concentrations initially at high Northern latitudes, which then even out as mixing spreads the gas into the rest of the atmosphere.
If you watch for blue, you see either no change, or decreased CO2 at the equatorial and subtropical latitudes where you suggest most CO2 is released.. The other graphs in my earlier link also showed that the variability was greatest in the North, which would be expected if this was where the gas was produced.
Two competing theories here:
“reduced cloud cover since the 1960′s, according to the new Chinese and Spanish Studies, has caused the tropical oceans to outgas a lot of co2 in the warm plentiful sunshine …”
and:
“the increase in atmospheric CO2 since 1900 is mostly if not entirely a result of anthropogenic carbon emissions”
Nice graphing Roger A, thanks. I do think you should be more explicit about the caveats though. Like the fact that your neat looking correlation for emissions and airborne co2 uplift only works for a narrow range of ‘residence time’ around five years. Given that the lag behind temperature at this scale is around 15 years, it is questionable.
Also, the obvious inverse of less cloud is more sunshine reaching the surface, that’s where you’ll find the close correlation between sunshine hours and temperature. As you can see from fig 1 in the headline post, there is a very close correlation between temperature and rate of change of co2 level.
So I think the two competing theories are more evenly matched than your graph comparison implies.
By the way, it’s not just oceans which have outgassed more because of reduced cloud. Don’t forget Cardellini’s volcanic soils too.
You can’t rely on cloud cover for a single region in the way Roger A has done. Each region is affected differently as it is moved towards or away from the main cloud bands when the climate zones shift latitudinally.
Using the global cloudiness index gives a much better fit than Spain alone.
The best fit of all is for cloudiness under the subtropical high pressure cells which is where all the action occurs.
Sometimes the level of solar input causes more CO2 to be released than the global sinks can absorb and sometimes not.
The arbiter would be solar effects on total global cloudiness which seem to be influenced by the degree of jetstream zonality / meridionality because the length of lines of air mass mixing increases with more meridionality causing more clouds globally.
Evidence is accumulating to the effect that solar variability affects jet stream behaviour by acting on the size and intensity of the polar vortices from top down and that in turn allows contraction or expansion of the subtropical high pressure cells which closes the circle of cause and effect.
Volcanic soils would be affected too as would soil moisture CO2 content on all the landmasses.
More or less sunlight onto the surfaces and into the oceans especially in equatorial regions must be what switches the balance of the carbon cycle in the atmosphere from net CO2 gain to net CO2 loss and back again.
It would happen on a 500 year cycle trough to peak as per LIA to date or peak to trough MWP to LIA but the ice cores just don’t record it for some reason.It is that long cycle period that appears to give a relatively monotonic trend over shorter periods of time.
Very similar findings are here
Click to access CO2-MEI-LOD.pdf
Mazzarella A., A. Giuliacci and N. Scafetta, 2012. Quantifying the Multivariate ENSO Index (MEI) coupling to CO2 concentration and to the length of day variations. Theoretical Applied Climatology, in press. DOI: 10.1007/s00704-012-0696-9
Abstract The El Niño Southern Oscillation (ENSO) is
the Earth’s strongest climate fluctuation on inter-annual
time scales and has global impacts although originating
in the tropical Pacific. Many point indices have been
developed to describe ENSO but the Multivariate ENSO
Index (MEI) is considered as the most representative
since it links six different meteorological parameters
measured over the tropical Pacific. Extreme values of
MEI are correlated to the extreme values of atmospheric
CO2 concentration rate variations and negatively correlated
to equivalent scale extreme values of the length of
day rate variation. We evaluate a first-order conversion
function between MEI and the other two indexes using
their annual rate of variation. The quantification of the
strength of the coupling herein evaluated provides a
quantitative measure to test the accuracy of theoretical
model predictions. Our results further confirm the idea
that the major local and global Earth–atmosphere system
mechanisms are significantly coupled and synchronized
to each other at multiple scales.
Slightly off-topic [but worth mentioning because I have a hunch that it will crop up again], is that a model which has tropical oceans being net contributors to atmospheric CO2 concentration, cannot [should not] simultaneously experience lower pH as a consequence of increased absorption of CO2 from the atmosphere.
Ocean “acidification” does not sit well with my current perception of an IPCC carbon-cycle model. It seems more credible that bulk pH changes may arise from beneath, not above, the ocean-atmosphere interface [as far as CO2 is concerned at ‘non-sink’ locations].That might also account for some portion of oceanic calcium supersaturation wrt carbonate.
TB:
“… your neat looking correlation for emissions and airborne co2 uplift only works for a narrow range of ‘residence time’ around five years.”
The “neat looking correlation” between CO2 and fossil fuel emissions in the second graph is in fact statistically independent of residence time. (You can argue that the scaling implies a residence time assumption, but R^2 = 0.97 regardless of how you scale the data on a graph.)
Stephen W:
“You can’t rely on cloud cover for a single region in the way Roger A has done.”
Objection! I didn’t come up with the Spanish cloud cover theory; TB did.
This was recently published, relating temperature and CO2
http://www.sciencedirect.com/science/article/pii/S0921818112001658?v=s5
And this was Realclimate’s response.
http://www.realclimate.org/index.php/archives/2012/09/el-ninos-effect-onco2-causes-confusion/comment-page-1/#ITEM-13053-0
[Reply] I’ve been involved in a long discussion on this paper here: http://wattsupwiththat.com/2012/08/30/important-paper-strongly-suggests-man-made-co2-is-not-the-driver-of-global-warming/ Ferdi Englebeen has been doing his best to stem the tide of heretical thought, but it’s a losing battle for him. He’s up against some very acute minds who see through his and norealcluemate.org’s arguments on the mass balance, d13/12 ratio, assumed sources and sinks and rates of release/absorption etc.
This is the caption for the graph you highlighted in your September 10th, 8.04pm comment.
CO2 emission from land use modification. Figure taken from Global Carbon Project 2010. Data source: R.A. Houghton 2010 (personal communication); GFRA (2010).
It does not show the total change in CO2 from tropical and temperate latitudes, which I think you assumed when talking about CO2 release from oceans. It indicates the change in net CO2 release and absorbtoin due to deforestation in the tropics and reforestation at higher Northern latitudes.
[Reply] Yes, I agree that is what they think it shows.
Ferdi seems to think that the rate of CO2 release from the oceans is limited to 16ppm per 1C of surface warming however much energy the sun pumps into the ocean beneath the evaporative layer (often down to 200 metres).
He sees no problem that the surface temperature rise is limited to 1C by faster evaporation and that the solar energy pumped in would result in a much higher temperature rise than that were it not for faster evaporation.
I’m not an expert on the response of the oceans to solar energy but it seems implausible to me that there could be such a limitation however much sunlight gets in.
Any ideas on how to resolve the issue ?
The significance being that if one were to multiply that 16ppm by 5.39 which is the enthalpy of vaporisation one gets pretty darned close to the actual observed rate of increase in atmospheric CO2 in response to the observed ocean warming.
It seems to me that the increased evaporation keeps the surface temperature lower than it otherwise would be and that the oceanic CO2 emissions to be expected should be linked to the amount of extra solar input and not the change in surface temperature.
But I’m not sure and need help on the point even if I just need to be told that I’m barking up the wrong tree. Unfortunately I don’t trust Ferdi’s input on the point.
The importance of the enthalpy of vaporisation being that the ocrean surface temperature would be 5.39C higher from the extra solar input if the rate of evaporation had not accelerated to reduce it to 1C.
So should one actually calculate the expected emissions as 5.39 times 16ppm or not ?
“rate of CO2 release from the oceans is limited to 16ppm per 1C of surface warming ”
It may be coincidence, but the paleoclimate data shows two typical changes between glacial and interglacial conditions. CO2 rises from 200ppm to 280ppm and temperatures riseby 5C.
That’s 16ppm/C.
Friends:
I interrupt your interesting conversation to remind about the limitations of the data you are trying to interpret.
The annual rise in atmospheric CO2 is the residual of the shorter term fluctuations over a year and is provided by the Mauna Loa data and some other stations. The longest time series of this annual CO2 data is obtained from Mauna Loa and only exists since 1958.
This Maun Loa data set of annual rise in atmospheric CO2 is close to linear with time Therefore, within the measurement errors, almost any curve can be fitted to the data. In one of our 1995 papers we fitted six different relationships to the data:
3 assumed a natural cause of the CO2 rise (i.e. global temperature increase),
and
3 assumed the anthropogenic emission is the cause of the CO2 rise.
(ref. Rorsch A, Courtney RS & Thoenes D, ‘The Interaction of Climate Change and the Carbon Dioxide Cycle’ E&E v16no2 (2005))
In each case the data were a perfect fit (within the measurement errors) for each annual datum.
These perfect fits were achieved by assuming the system of the carbon cycle is adjusting towards an altered equilibrium. And 3 of our models each postulated that the rise in global temperature had induced the changed equilibrium in a different way, while 3 of our models each postulated that the anthropogenic emission had induced the changed equilibrium in those different ways.
Data indicates nothing when it can be demonstrated to represent anything.
Also, our referenced paper analyses the dynamics of seasonal sequestration at each measurement site. These dynamics demonstrate that local sequestration processes can easily sequester all the emitted CO2 (both natural and anthropogenic) of each year. This provides a paradox: the atmospheric CO2 is rising but the sequestration processes have dynamics which indicate all the emitted CO2 could be expected to be sequestered.
A solution of that paradox may (I think probably would) indicate the cause of the recent rise in atmospheric CO2 concentration.
Investigation of the AIRS CO2 spatial distribution plots may indicate sources and sinks so may afford a solution to the paradox. Hence, I am interested in the hypothesis suggested by Stephen Wilde.
Richard
Entropic Man:
You say;
“It may be coincidence, but the paleoclimate data shows two typical changes between glacial and interglacial conditions. CO2 rises from 200ppm to 280ppm and temperatures rise by 5C.
That’s 16ppm/C.”
It is not a coincidence. It is how the 16 ppmv/deg.C was determined.
Richard
The AIRS graphic in my September 10th, 10.09 comment shows the highest CO2 concentrations in Spring and early Summerin Northern Temperate latitudes. The lowest concentrations consistently occur over the tropical oceans.
High concentrations should occur over sources, and low concentrations over sinks. This is consistent with the annual cycle of CO2 concentration being driven mainly by photosynthesis and decay in Northern hemisphere temperate and taiga forests.
With La Nina conditions predominating over the period shown, cool ocean water surfacing in the Pacific would be expected to absorb CO2, explaining the low air concentrations observed there.
It will be interesting to see what the next El Nino does to AIRS measurements of CO2 levels over the Pacific.
“It is not a coincidence. It is how the 16 ppmv/deg.C was determined.
Richard”
Which leaves us with the dilemma of trying to explain why the relationship seems to have become so distorted in recent decades. We have an increase of 80ppm in 55 years, without the 5C temperature rise which, from past paleoclimate precedent, would normally be required to drive the change.
Richard:
Could you supply a link to your E&E paper that isn’t behind a paywall? It sounds interesting and I’d like to read it. Thanks.
Entropic Man:
You say
“Which leaves us with the dilemma of trying to explain why the relationship seems to have become so distorted in recent decades.”
No. It gives us the dilemma of trying to understand why anybody accepts the ice core data.
Richard
Entropic Man:
It’s a dilemma only if you believe that the 1.3 trillion tons of man-made CO2 emitted over the last 55 years had nothing to do with the 80 ppm CO2 increase. 🙂
Roger Andrews:
The presentation I gave at Heartland 1 is almost entirely ‘ cut & paste’ from the paper and I can send you a copy of that if you email me. Alternatively, Tallbloke could email me so I could send it to him and he could forward it to you.
But I regret that I cannot send you a copy of the actual paper. This is because I am now on the Editorial Board of E&E and the publisher would take a dim view of my breaching publication policy. Sorry.
Richard
Thanks Richard. Your offer is gratefully accepted. Easiest way is probably to have TB send you my email address so you can send it directly, but having him forward it would be fine too.
TB are you there? Are you still speaking to me? 🙂
Roger Andrews:
TB has probably gone to bed (it is nearly half past one in the morning here in the UK). I am going to bed now and will return to this in the morning.
Richard
Roger Andrews says: September 10, 2012 at 4:12 am
“Ray:
Your question about whether the Mauna Loa CO2 record is a reliable “global reference” keeps coming up in one form or another, so I thought this might be a good time to see how it fits in with CO2 records from other parts of the world”
Thanks, but your links are to static data comparing co2/temp for a ‘snapshot’ of live, or averaged, data. I tried to imply a more dynamic form of data that would include the regional hydrocycle affecting a node of data collection. ‘Water’ can be defined as the atmospheric ‘atractor’ for co2, with two opposing systems of both ‘ocean release’, and ‘atmospheric scrubing’, with temperature being the main determinant that affects both systems.
I agree with the offset for global co2 average, but was considdering the dynamic ratio between ‘atmospheric scrubing’/’ocean uptake’ against ‘atmospheric scrubing’/’ocean outgassing’ with temperature changes up-wind of the data collection node for near surface co2 measurement.
There’s no mention of the main atractor for co2 in your links. ‘H2O’ (in liquid phase [water]) is important in that its precipitation brings co2 to a scenario where it can be sequestered to ocean flora, land flora, or land-based erosion sequestration, by way of precipitation striking Earth’s surface (don’t confuse this with ‘insolation’ 🙂 ).
I was looking for a way to explain the lead/lag anomalies in TB’s first graph. 🙂
Best regards, Ray.
Roger A: Sorry, got an early night. I’ll email Richard today. I think he makes a strong point when he says: “Data indicates nothing when it can be demonstrated to represent anything.”
He also says: “Investigation of the AIRS CO2 spatial distribution plots may indicate sources and sinks so may afford a solution to the paradox. Hence, I am interested in the hypothesis suggested by Stephen Wilde.”
To which I have given some thought, and will say something about in my next comment. Hopefully you’ll still be awake when I finish writing it. 🙂
Entropic Man said:
“It may be coincidence, but the paleoclimate data shows two typical changes between glacial and interglacial conditions. CO2 rises from 200ppm to 280ppm and temperatures rise by 5C.
That’s 16ppm/C.”
Richard replied
“It is not a coincidence. It is how the 16 ppmv/deg.C was determined.”
In that case the paleo data doesn’t resolve the issue since it appears that the ice core records
may be failing, for whatever reason, to record past short term CO2 swings of the scale we are currently observing.
It could be that human emissions have been causing the large recent changes or it could be a natural process that ice cores fail to record.
Given the AIRS data showing no sign of a human CO2 source I suspect the latter.
My favoured explanaton will remain changes in sunshine hours in the subtropics unless that inferred limit of 16ppm per 1C ocean surface warming can be better substantiated.
Entropic Man said:
“The AIRS graphic in my September 10th, 10.09 comment shows the highest CO2 concentrations in Spring and early Summerin Northern Temperate latitudes. The lowest concentrations consistently occur over the tropical oceans.”
TB has explained why that AIRS graphic is not suitable for the current discussion. The one I have been working from is a far more useful graphic.
Stephen Wilde says:
September 11, 2012 at 10:53 pm
The importance of the enthalpy of vaporisation being that the ocrean surface temperature would be 5.39C higher from the extra solar input if the rate of evaporation had not accelerated to reduce it to 1C.
So should one actually calculate the expected emissions as 5.39 times 16ppm or not ?
Hi Stephen. I think Ferdi is correct in saying that for the instantaneous snapshot of the flux between air and water, what matters are their actual respective temperatures, not “what the temperature would have been if…”
However, where Ferdi goes wrong is in falling victim to thinking statically and in averages, instead of considering actual local conditions over time. I’ll explain:
In the tropics, the SST max’s out at around 30C. It can’t get warmer than this no matter how much sun hits it. At that T, the evaporation is really fast and the surface waters soon become salt heavy and sink. This brings new, more heavily co2 laden water to the surface, which will outgas. So your ‘faster rate of evaporation’ will have an effect. This will have speeded up the upwelling and therefore the Thermohaline Overturning Circulation in general. That leads to a general increase in deep co2 enriched water coming back to the surface.
The decrease in cloud cover will have meant more ‘faster evaporation’ over more areas of ocean too. This may be the critical factor. at the local level, adjacent sunlit areas of ocean and cloud shaded areas of ocean are contantly emitting/absorbing co2 and the balance between the two processes will constantly be changing as clouds come and go. If the cloud amount falls over decades as the Spanish data shows, then you’d expect emission to gain the upper hand and airborne levels to increase. So I think the crucial point is that the rate of change of co2 level is directly affected by sunshine hours, rather than sst, though the two correlate well anyway, but more-so over land, as Doug Proctor and Willie Soon showed us.
Now add in the positive phase of the PDO and generally warmer sst’s over vast areas of the Pacific over the last 40 years.
Then consider the dodgy nature of the calibration of the Antarctic ice cores, and diffusion smoothing out the peaks in co2 spikes during the top of ~60 year cycles which are evident, particularly in the Greenland data. As Entropic Man says, the Northern Hemisphere is more strongly affected, and has a bigger seasonal swing in co2 emission/absorption. That’s why the stomata data fits Greenland ice core co2 estimates *before they got recalibrated to the Antarctic series*.
We’ve been sold a pup.
Stephen says:
“My favoured explanaton will remain changes in sunshine hours in the subtropics”
Heh, beat me to it while I was writing my comment. 🙂
Not that we haven’t both been saying this for some time now. I think our new insights here help us understand the apparently non-linear effect though, and why Lower Tropospheric temp increases more than sst, even while global OLR increased 5W/m^2 since the late ’40’s according to the NCEP reanalysis of radiosonde data which might not be so bad after all.
All of which adds up to solar variation having a much bigger effect on climate than recognised, because cloud cover change is linked to solar variation.
Now we need a competent and patient programmer to help us model our findings.
Thanks TB.
A focus on more evaporation pulling the cooler CO2 rich waters upwards works well enough to deal with the Henry’s Law problem.The question of how the CO2 from lower down gets to the surface is one which Ferdi mentioned.
If the primary factor is more evaporation pulling up the CO2 rich cooler water then the relationship between the surface temperature rise and the amount of CO2 coming up could still be linked to the enthalpy of evaporation at 5.39 to 1 though couldn’t it ?
Due to that faster evaporation pulling water upwards faster the surface rise of only 1C would result in a similar rate of CO2 release as would otherwise have needed a surface temperature rise of 5.39C to support it in the absence of faster evaporation.
I’m pushing that relationship because a multiple at that level gets so close to real world observations that I need some convincing that it is a coincidence.
A lot of stuff is fitting together nicely in my view from the top of atmosphere to depths of oceans and all due to the sun interfering with air and water circulation from the top down when wavelength and particle proportions change rather than raw TSI.
Hi Stephen,
we’re pulling the cart well together on this one, as usual. I think you need to be cautious about pushing the enthalpy energy times 16ppm argument though. Firstly the 16ppm is likely wrong, it should be higher. Secondly, I don’t see a direct relationship between the energy of enthalpy and the increase of evaporation globally under reduced cloud cover. There will be an indirect one though, so your point is important. Also, be careful with terminology. Evaporation doesn’t ‘pull co2 rich water up”, it concentrates salt content which makes surface water sink, thereby pushing the co2 rich water up.
I think for now I’m going with investigating the implication of the increase in OLR seen in the NCEP reanalysis and it’s relationship with the amount of extra energy being shifted from ocean through air to space. That should give us a ball park figure on the increase in thermohaline overturning, and from that, a clue as to the amount of extra co2 released from the oceans. It should be possible to back check the energy figure with Spencer’s estimate of the effect of a 1% cloud decrease on surface temp too.
I think we’re nearly ready for it to be worth firing up the old Casio calculator. I’ll check its (very old) batteries. Otherwise my old slide rule will probably get us close enough.
Old engineer’s joke:
Q: What is 2 x 2?
A: Approximately 4
🙂
Points taken TB.
An indirect relationship is what I had in mind.
The rate of upward drawing of CO2 rich water will be linked to the energy requirement of the evaporative process and thus the enthalpy of vaporisation. Faster cooling at the top from more evaporation will increase the temperature differential through the ocean skin to encourage upward convection and conduction from below the ocean skin.
So a 1C warmer ocean surface which is kept at that temperature by more evaporation could achieve the same outgassing outcome as a water surface 5.39C warmer in the absence of more evaporation.
The 5.39 multiple only gets us to around 86ppm but the real world produced about 100 ppm so other factors are involved including perhaps an initial error with the 16ppm since it seems to be inferred from ice cores rather than measured and also the upwelling from the thermohaline circulation will itself have variations in CO2 content emanating from the time when it was initially subducted.
An interesting point is that a 1C warmer surface (or rather evaporative layer) doesn’t reduce the temperature gradient across the ocean skin otherwise it would REDUCE the energy uplflow from beneath the skin.
The so called AGW theory concerning the ocean skin relies on that proposition,namely that more warmth in the evaporating layer slows down energy loss from ocean to air.
I’ve spent a lot of time trying to work out whether that theory is correct and eventually came to the conclusion that it doesn’t work like that otherwise the cooler ocean skin layer above the slightly warmer ocean bulk below it could not exist.
In practice the warmer the ocean surface/evaporative layer becomes the faster is evaporation so that the gradient across the cooler ocean skin is actually increased (either going deeper or gaining a steeper slope) because of the enthalpy of vaporisation causing evaporation to be an energy hungry net cooling process.
It is the energy hungry nature of the evaporative process that created the cooler ocean skin in the first place so more evaporation from a warmer sea surface / evaporative layer simply intensifies that cool ocean skin layer to increase the rate at which the lower waters are drawn up.
We will get a lot of opposition to the proposal that a warmed surface with more evaporation actually speeds up the water to air energy flow thereby bringing more CO2 up with it but I’m sure that is correct.
It all comes back to my earlier proposition that the oceans will only hold as much energy as the surface air pressure allows when combined with solar input. I know we are agreed on that but many others need it to be otherwise for their climate theories to get off the ground but to my mind there is lots of evidence inconsistent with their view.
“No. It gives us the dilemma of trying to understand why anybody accepts the ice core data.
Richard”
I see a lot of of sceptics complaining about the ice core data, but, among the laymen, their reasoning mostly comes down to ” It disagrees with my world view”. . Could you refer me to a clear analysis of the reasons why sceptic scientists reject the ice core data,pitched at a level suitable for a retired science teacher.
“We may in fact be the only two people on this blog who believe that “the (CO2) increase is almost certainly due to anthropogenic emissions”.
Roger Andrews
Make that three.
EM: their reasoning mostly comes down to ” It disagrees with my world view”
Don’t denigrate entire groups of people here please.
Could you refer me to a clear analysis of the reasons why sceptic scientists reject the ice core data,pitched at a level suitable for a retired science teacher.
Are you a retired science teacher? Try this for starters:
http://www.warwickhughes.com/icecore/
“Could you refer me to a clear analysis of the reasons why sceptic scientists reject the ice core data,pitched at a level suitable for a retired science teacher.”
Because it is inconsistent with clouds of CO2 coming off sunlit oceans (and only sunlit oceans) as per the AIRS data.
Variations in global cloudiness and sunlight appear to be enough to make a significant difference to the outgassing of the oceans in regions where the oceans are sources on centennial timescales such as MWP to LIA or LIA to date but the ice cores appear to show no response to CO2 changes until about 800 years have passed.
For some reason the ice cores are not catching the natural variability in atmospheric CO2 over periods of less than 800 years.
Now it could be that there is no such variability in which case humans would indeed have reversed the natural order of things but applying Occam’s Razor I deem that unlikely.
Apart from the ice cores the standard diagnostic indicators are the mass balance (which can be shown not to apply if the human CO2 is all quickly sequestered locally from extra biosphere activity induced by the human CO2 as seems likely from the AIRS data) and the isotope ratio but there are many questions about the validity of that too because non organic processes have not been fully accounted for.
So, in light of current information it is simply a matter of what one wishes to believe.
However, to my mind the new AIRS data is highly persuasive to the effect that sunshine on water and soil moisture on land is capable of swinging the system periodically from net CO2 release to net CO2 absorption and we are currently in a period of net natural release consistent with that release being a consequence of warming from centennial timescale solar influences on the global air circulation and not a cause.
Entropic man:
At September 12, 2012 at 10:37 am you ask me
“I see a lot of of sceptics complaining about the ice core data, but, among the laymen, their reasoning mostly comes down to ” It disagrees with my world view”. . Could you refer me to a clear analysis of the reasons why sceptic scientists reject the ice core data,pitched at a level suitable for a retired science teacher.”
I think the best summary which fits your request is the Statement written for the Hearing before the US Senate Committee on Commerce, Science, and Transportation from the late Zbigniew Jaworowski. It is titled “Climate Change: Incorrect information on pre-industrial CO2” and is dated March 19, 2004. As Tallbloke has already said, it can be read at
http://www.warwickhughes.com/icecore/
Zeb’s severely failing health meant he could only submit the statement in written form.
Similarly, his failing health prevented his attending Heartland 1 and he asked me to present his paper there so I did.
I consider it a great honour to have been associated with Zeb throughout the final decades of his life. He is the ‘father’ of ice core studies who conducted dozens of field trips to obtain ice cores, and he devised most of the methods now used to analyse ice cores. Hence, his anger when he learned of – what he considered to be – the abuse of the methods he devised for paleo- climate studies.
I first became associated with him when the UN appointed him to determine the world-wide dispersal of radionucleatides from the Chernobyl nuclear disaster. He was a professor from a communist country and the disaster was in a communist country at the height of the Cold War. But nobody questioned his appointment because he was unarguably the outstanding authority in his field. However, he feared that whatever he discovered could be portrayed as having partisan bias so he desired association for review of his work by someone in a coal industry (coal was a competitor to nuclear) from the Western side of the Iron Curtain. I was a material scientist at the UK’s Coal Research Establishment (but not then the Senior Material Scientist) and filled that role. He was a communist and atheist while I am a socialist and Accredited Methodist Preacher, so we could be thought to be natural ‘enemies’, but I consider it a great honour that we were friends.
And that friendship is why I add this background information in hope that I honour the name of the great Zbigniew Jaworowski. A great man, a true scientist and a sorely missed friend.
Richard
tallbloke:
At September 12, 2012 at 6:45 am you say
“Roger A: Sorry, got an early night. I’ll email Richard today.”
I write to ask if you have yet done that.
Following concerted attacks which destroyed two computer systems, I now have severe defences and communications sometimes don’t get through. Sorry for the inconvenience this causes, but if I know you have sent it then I should be able to find it.
Richard
[reply] not yet.
Entropic Man:
You may also want to read the section titled
“The truth About Ice Cores”
in the paper at http://www.warwickhughes.com/icecore/zjmar07.pdf
That section includes the info. in Jaworowski’s submission to the US Senate Committe and much more.
All of that paper is worth a read if you want an overview of main sceptical arguments.
Richard
Thanks, Richard, I haven’t seen that summary before. It certainly articulates many of my objections that I have previously read in various locations.
Entropicman, I’m glad you are returning your attentions to carbon dioxide and the carbon cycle. It is certainly where my chief concerns lie in this ensemble.
[I have to say that I think you took your eye off the ball during the Rio Conference when you announced in you BBC posts that you now considered it subsidiary to the other sustainability arguments.]
Professor Jaworovski’s evidence to Congress recaps his 1997 paper. Anyone who heads a paper – “Another Global Warming Fraud Revealed” implies an agenda beyond the purely scientific. Do you have anything with a more respectable pedigree?
Click to access icecoredatashownoco2increasejaworowskispring97.pdf
[snip] Off topic.
[Reply] Debate reasonably, and on topic, or naff off.
Entropic Man:
I take extreme umbrage at your suggestion that anything on ice cores from the ‘father’ of ice core studies lacks “pedigree”.
Furthermore, I gave you link to a peer reviewed study that includes the same data (and more) as is in his submission to the US Congress. And you link to another paper from the same author claiming its title somehow discredits its contents.
You have posted this nonsense because I offered you the links in response to your having said to me;
“I see a lot of of sceptics complaining about the ice core data, but, among the laymen, their reasoning mostly comes down to ” It disagrees with my world view”. . Could you refer me to a clear analysis of the reasons why sceptic scientists reject the ice core data,pitched at a level suitable for a retired science teacher.”
You have displayed your “world view” for all to see; i.e. you use ridiculous excuses to ignore and discredit whatever does not concur with your prejudices.
Richard
Mr Wilde, look at the October 2005 data. The tropical seas are at their warmest and you would expect them to be releasing CO2 at their maximum rate. By your hypothesis ,these should be the areas with the highest CO2 concentration.
Instead, they are still lower than the planetary average. Your hypothesis would appear to be falsified.
Entropic Man:
You quote the first two sentences of Jawarowski’s paper and assert;
“I find it difficult to take seriously as scientific evidence a paper that begins its introduction with these two sentences.”
Well, I find it difficult to regard with anything but contempt somebody who quotes out of context two sentences and uses that to disparage work of the deceased. The first two paragraphs – which begin with those sentences – are clear, factual and cogent. They say;
“On Feb. 2, 2007, the Intergovernmental Panel on Climate Change (IPCC) again uttered its mantra of catastrophe about man-made global warming. After weeks of noisy propaganda, a 21-page “Summary for Policymakers” of the IPCC Fourth Assessment Report, 2007, was presented in grandiose style in Paris to a crowd of politicians and media, accompanied by a blackout of the Eiffel Tower to show that electric energy is bad. The event induced a tsunami of hysteria that ran around the world. This was probably the main aim of this clearly political paper, prepared by governmental and United Nations bureaucrats, and published more than three months before the IPCC’s 1,600-page scientific report, which is to be released in May. In the words of the IPCC, this delay is needed for adjustment of the main text, so that “Changes . . . [could be] made to ensure consistency with the ‘Summary for Policymakers.’ ” Not a single word in these 1,600 pages is to be in conflict with what politicians said beforehand in the summary!
This is a strange and unusual method of operation for a scientific report, and even stranger is the frankness of the IPCC’s words about the delay, disclosing its lack of scientific integrity and independence. It is exactly the same modus operandi demonstrated in the three former IPCC reports of 1990, 1995, and 2001: First the politics, then the science.”
As an introduction, that is both true and informative. But you choose to ignore what follows because it does not support your “world view”: sad, so very sad.
Richard
[Reply] I’ve binned Entopic’s off topic stuff so this will be the last word on the matter on this thread. We will now return to discussing the relationship of co2 and temperature. If you want to discuss the politics with Entropic, please feel free to continue on the history of the climate scare thread. Thank you. TB
I would have thought the tropical ocean would be at it’s warmest around February-march, a couple of months after perihelion. Where’s the data you’re working from?
Tallbloke:
A reply to a post that I have provided seems to have vanished without appearing here, but I observe you snipped the points which it was answering.
If that was moderation then I accept it. If not, then you may want to find it.
Richard
It has reappeared. I apologise.
Richard
Richard, I just want to keep this useful discussion on topic. Thanks. By the way, I emailed you Roger A’ address.
Tallbloke and Roger Andrews:
The 3 MB file is now downloading as an attachment to an email to you.
Thanks to Tallbloke for his help.
Richard
My apologies, Richard.
As for when tropical surface temperatures peak, it’s an interesting question and much more complex than my simple assumption . A quick scan of a couple of papers
Click to access chang_1994.pdf
http://journals.ametsoc.org/doi/abs/10.1175/1520-0493(1982)110%3C0354%3AVITSST%3E2.0.CO%3B2
suggests that seasonal changes occur in different parts of the tropics at different times, with the dominant influence being wind direction, ratther than direct insolation. Multiyear cycles such as ENSO also distort seasonal patterns.
Building a convincing worldwide model to demonstrate that the annual cycle in CO2 concentration is due to marine, rather than land influence, is likely to be a complex undertaking.
Entropic man:
It is good there is something we can agree about.
You say;
“Building a convincing worldwide model to demonstrate that the annual cycle in CO2 concentration is due to marine, rather than land influence, is likely to be a complex undertaking.”
I agree. And I add that it would be more useful than complex undertakings which are attempted; e.g. building GCMs that cannot be validated and each emulates a different climate system when the Earth only has one climate system.
Richard
TB:
Just got Richard’s stuff in the mail. I’ll look at it later today. My thanks to you and Richard.
In the meantime I have some rogue data to present. Here’s a comparison of monthly changes in SST against monthly changes in CO2 between 1959 and 2011 for the 60-90N, 30-60N, 0-30N, 0-30S and 30-60S latitude zones showing correlation coefficients (R) for each zone over a range of lead and lag times:
The results show that monthly CO2 and SST changes are correlated in the tropics and subtropics (30N to 30S) but effectively uncorrelated at higher latitudes.
And they also show that R values are highest when the CO2 changes lead the SST changes, not when they lag them.
I think this result is most likely some kind of statistical aberration, but on the other hand there’s a lot about the the carbon cycle that we don’t understand, so I’m going to throw out the idea that maybe CO2 changes lead SST changes and not the other way round. Any comments?
The case for why the temperature to CO2 rate of change relationship absolves humans of responsibility for the rise is laid out simply here.
There is a 1-1 relationship between the rate of change and the integrated CO2 level, modulo an constant of integration. If you can match the CO2 rate of change to the temperature, then integrating that matched temperature to the CO2 level will reproduce the CO2 level, as demonstrated here.
It is true that there is a constant offset necessary to match the temperature anomaly data to the CO2 rate of change. But, that is perfectly reasonable, as it is a temperature anomaly, and already has an arbitrary constant offset built in. Human inputs cannot explain a constant offset, as the rate of emissions is not constant.
In the end, the basic relationship, in the modern era since at least 1958 when we started getting good measurements, is that the rate of change of CO2 tracks the scaled and re-baselined temperature anomaly. And, there is no room in that relationship for significant human attribution.
Les Johnson says:
September 8, 2012 at 1:14 pm
“I put a lag in my sheets. I found 9 months the best, as did Humlum et al.”
That is an average lag in the total accumulation. But, the derivative of CO2 and temperature are contemporaneous. There is then necessarily a 90 deg phase (1/4 wavelength) lag due to the integration. I.e., the delay for a particular sinusoidal function in a Fourier basis is 1/4 wavelength, because, e.g., the integral of the cosine is the sine, which lags the cosine by 1/4 wavelength.
Thus, the lag in overall concentration for the pronounced ~60 year oscillation in temperatures is about 15 years. For a ten year oscillation, it would be 2.5 years. When you position the series to get the best overall fit to a constant delay shift, 9 years is probably about what you would expect. But, the actual lag is frequency dependent.
Bart: thanks for joining this thread. I think the units in your graph here:
Are millions of tonnes. 8Gt is the figure given for recent annual emission levels anyway.
Roger A, interesting graph. Maybe what you think is a lead of 8 months is more likely a lag of 4 months from the previous year?
There’s no doubt in my mind that co2 lags T globally by around 6 months.
Ray Tomes confirmed it last year, but I first determined this by using the woodfortrees site in 2009.
tallbloke says:
September 12, 2012 at 8:36 pm
The units don’t concern me as much as morphology. There’s only room to fiddle with the constant offset to match everything. As emission rates are not constant, there’s no room for them.
tallbloke says:
September 12, 2012 at 8:41 pm
Please note my response to Les J. above. The lag is not constant across the frequency spectrum. It comes about because the rate of change of CO2 and temperature are contemporaneous.
TB:
“There’s no doubt in my mind that co2 lags T globally by around 6 months. Ray Tomes confirmed it last year …”
So did I. https://tallbloke.wordpress.com/2011/06/26/which-causes-which-out-of-atmospheric-temperature-and-co2-content/#comment-7250
But now I’m saying that there’s a chance we were wrong and that the lag is the other way round. And if it is we have some serious reconsidering to do.
Take a look at these two graphs for 0-30N. The first assumes that SST leads CO2 and plots the best-fit match, which we get with the CO2 response lagged 6 months (note the inverted CO2 scale). As best-fit matches go, it sucks. The second assumes that CO2 leads SST and plots the best-fit match for this case, which we get with the SST response lagged 7 months. This match is some way short of perfect but it’s a hell of a lot better than the other one, and it isn’t manufactured by lags or leads from the previous year.
Roger A. – Comparing the rate of change of temperature to the rate of change of CO2 shows that CO2 clearly lags. The lag is not uniform, and you are never going to get things to match up well for both long term and short term variations with a uniform lag. The rate of change of CO2 is synchronous with temperature. That explains why the lag is not uniform in CO2 vis a vis temperature.
Bart, I’m trying to get your graphs to appear inline with your comments, but it seems photobucket doesn’t allow it. Sorry.
I don’t understand why the two red co2 curves in Roger A’s graph are different . I feel a d’oh! moment coming my way.
“But now I’m saying that there’s a chance we were wrong and that the lag is the other way round. And if it is we have some serious reconsidering to do.”
When I made my graph it was obvious which co2 oscillation related to which temperature oscillation, and co2 definitely lagged temperature. I’m not meaning to be obtuse, I just don’t understand your graph.
Note please that these variations are typically about a 2 year period. With the 1/4 delay of the integration, there should result in a roughly 6 month lag.
Agreed, that’s roughly what I found
http://woodfortrees.org/plot/esrl-co2/from:1997/to:2006/mean:13/detrend:14/plot/uah/from:1997/to:2006/mean:13/scale:4/offset:364
TB:
The two red co2 curves are different because the top one is inverted – as I took pains to point out – while the bottom one isn’t, and because one of them is shifted six months one way and the other seven months in the opposite direction. If you apply the concomitant eyeball gymnastics you will see that they are in fact the same.
D’oh!
🙂
So which one is inverted relative to the temp rate of change? I ask because it occurs to me that when sst falls, surface air temp falls further, afew months later, and this will make a difference to the air/sea co2 flux balance.
Ah, I see, the y axis in red is opposite in each graph. Hmm, some pondering to do while I sleep. Gunnite.
“Take a look at these two graphs for 0-30N. The first assumes that SST leads CO2 and plots the best-fit match, which we get with the CO2 response lagged 6 months (note the inverted CO2 scale).”
It is perhaps oversimplifying to assume that CO2 always drives temperature,that temperature always drives CO2.
I suggest different situations in which different interactions occur.
1) During the early stages of an interglacial rising temperatures due to .rbital variations warm the Northen Hemisphere , modetrating Winters and leading to CO2 release from warming water, decaying tundra etc. This is definately a case in which temperature leads CO2. Cooling at the end of an interglacial reverses the effect, with temperature again leading CO2.
2) Once the warming is under way, extra back radiation from increasing CO2 amplifies the temperature change. This accelerates further CO2 release and the two drivers generate a positive feedback loop. At this stage it is meaningless to describe one or other as leading when both processes proceed simultaneously.
3) Over the last half century we observe a 21% increase in CO2, withn no matching increase in energy input and a much smaller temperature rise than past experience would suggest. This will certainly generate argument , but in my own view the change is anthropogenic and global temperatures are lagging behind the rate of CO2 increase. If you want a non-anthropic equivalent, think what effect you would expect of a continuing massive CO2 release from a supervolcano.
In these situations CO2 leads temperature.
Roger A. – I would suggest not inverting the axis, as we are trying to align things in phase, and 180 deg out of phase is totally out. Also, it makes it very difficult to compare.
However, if it appears to me that you have the delay and advance directions reversed. The rumpled upward peak in the red CO2 line in the first plot near the end appears to have moved forward in the second plot to cut be off halfway at the end. That would mean you advanced the CO2 line, i.e., gave it some lead. Giving the line lead is what makes up for the lag. So, the better match of the second plot appears to confirm the roughly 6-7 month lag.
tallbloke says:
September 12, 2012 at 11:35 pm
“Agreed, that’s roughly what I found”
But, note how the delay depends on the cyclic period of the formation you are looking at. It is always about 1/4 wavelength.
Humble pie time
Munch munch, chomp chomp, chew chew, swallow, gag.
I just found out that I got the columns on my spreadsheet flipped. What I plotted as SST was CO2, and vice versa.
So yes, SST does lead CO2, and all’s right with the world.
Apologies to all, and I promise not to do it again.
Bart says: September 13, 2012 at 12:38 am
“Roger A. – I would suggest not inverting the axis, as we are trying to align things in phase, and 180 deg out of phase is totally out. Also, it makes it very difficult to compare.”
I concur.
I’d also like to add that we are looking at an ‘ocean-atmosphere’ interface that reacts to both systems of ocean ‘and’ atmosphere, with their inclusions and exclusions of co2 dependant upon the particular state of the scenario for the transition of co2 across the boundary of ocean-atmosphere.
A point of interest here. As ocean surface salinity is increased by evaporation, the co2 density is also increased. Partial pressure may well force co2 from ocean surface via this scenario.
Also, ‘RH’ (relative humidity) provides the key for ‘automatic’ atmospheric acceptance of water vapour from SSTs of ~+28C.
Best regards, Ray.
Roger Andrews says: September 13, 2012 at 1:22 am
“Humble pie time”
Don’t ‘gag’ on this, it’s an example of how ‘anyone could make the same mistake’ (including myself). 🙂
Put this in the box of experience and soldier on! 🙂
Best regards, Ray.
Thanks Ray 🙂
Soldiering on, I hadn’t forgotten your earlier question about “dynamic” CO2 measurements. I don’t know where you would go to get info on something like that, but with gigabucks now being spent each year on climate research one would think that someone had done some work on it.
Regarding your comment on temperature/CO2 lags, no, I don’t know why they occur either, but they do exhibit the following interesting (to me at least) features:
1. They’re quantifiable only in the tropics and subtropics because we don’t see a relationship between temperature and CO2 changes anywhere else.
2. The temperature (SST) changes that trigger the CO2 changes are dominantly ENSO-generated.
3. According to my results (and I think I got it right this time) we see a relationship between SST changes and CO2 changes only when the annual average SST exceeds 20C.
“The temperature (SST) changes that trigger the CO2 changes are dominantly ENSO-generated”
That is the essential first step in the chain of logic.
By my reckoning the changes in the width of the equatorial air masses are caused by solar influences which I have described elsewhere and it is that change in width which allows more or less solar energy into the subtropical oceans beneath the relatively cloud free, descending high pressure cells either side of the ITCZ.
Those changes in solar input then skew the long term balance of the global energy budget so that when warming is in progress warm El Ninos gain dominance relative to cool La Ninas and when the globe is cooling La Ninas gain dominance over El Ninos.
That relationship seems to hold on the timescale of the transition from MWP to LIA and LIA to date.
However there is that 60 year cycling between El Nino dominance for about 30 years and La Nina dominance for about 30 years which tends to obscure the background signal over those timescales.
Nonetheless the background signal comes to the fore from one positive or negative PDO phase to the next in the form of upward temperature stepping every 60 years or so during the warming phase such as LIA to date and presumably downward stepping from MWP to LIA.
Bob Tisdale has good data on that stepping phenomenon over the past century.
The reason for the 60 year cycle is separate and need not concern us here.
The important point for this thread is the link between those ENSO/PDO phenomena and the changing CO2 release/absorption balance from the oceans.
I would expect rising atmospheric CO2 content during the approximate 500 year rise in dominance of El Ninos from LIA to date and falling atmospheric CO2 content for the approximate 500 year rise in dominance of La Ninas from MWP to LIA.
Within such 500 year periods there are peaks and troughs which would fit with shorter term solar variations such as the Dalton Minimum and I note that some say the LIA effectively continued to about 1850 due to such shorter term troughs in solar activity.
The rise since 1850 in both global temperatures and atmospheric CO2 could be explained by such a process.
The fact that the observed rise in CO2 is steady rather than closely tied to the shorter term ENSO / PDO variations would be due to smoothing of the trend arising from the length of time it takes the ocean warmth to circulate through all the ocean basins.
The ice cores do not record changes on those timescales for whatever reason and I think we need to look to the various non organic processes within the carbon cycle to explain the observed isotope variations.
The mass balance argument is dead because there are alternative explanations for the observations as Bart and I pointed out at the WUWT thread.
I think I have presented a plausible complete system hypothesis there which complies with observations and basic physics. It just needs testing.
Stephen says:
The mass balance argument is dead because there are alternative explanations for the observations as Bart and I pointed out at the WUWT thread.
Ahem.
tallbloke says:
September 6, 2012 at 4:05 pm
“All of your mass balance and isotope balance arguments are moot with the new discovery by Cardellini et al. You are going to have to rethink your argument structure in the light of the new empirical evidence.”
Ahem!
For a decade I have been explaining that the mass balance argument is circular.
It uses the assumption that the anthropogenic emission causes the rise to prove that the anthropogenic emission causes the rise.
Richard:
I’ve met Stephen in person and he’s a great guy, even if he does have a tendency to eloquently reformulate prior art and claim it as his own. 😉
I agree your own ‘Ahem’ has a stronger claim to primacy, and thoroughly earned its exclamation mark. 🙂
Anyway, we’re having a really productive time here, so lets put it aside for now.
Something I’s like to discuss is ocean heat content and how much extra energy must have gone into the oceans during the global warming period to generate the rise in steric sea level indicated by the enhanced rate of sea level rise measured by satellites and tide guages. This might be another way of checking Spencer’s estimate of how much forcing a drop in cloud cover produces.
Generally, I’d like to garner ideas for attempting quantification of what we’ve found, and thus getting to a point where we can challenge the ice core derived 16ppm per degree C figure, also taking into account the non-linearity which will be introduced by a speeding up of thermohaline overturning. The MET Office averages post I put up the other day found there had been an increase in rainfall as they shifted the baseline on a decade, and this indicates a speeded up hydrological cycle too.
I doubt we’ll pin anything down to a small margin of error, but it seems a worthwhile exercise to help keep us focused on measured quantities.
“even if he does have a tendency to eloquently reformulate prior art and claim it as his own. ”
ooh, that’s a bit harsh.
I only said that Bart and I had pointed out that the mass balance approach is dead, not that we had originated the thought.
Nor did I claim originality on the isotope issue, merely reiterating it.
Where I do claim originality is in taking disparate bits of prior art and integrating them into a coherent hypothesis that appears to fit with a wide range of observations without offending basic physical principles.
Often those observations are ones which AGW theory struggles with.
“Where I do claim originality is in taking disparate bits of prior art and integrating them into a coherent hypothesis…”
Which grows out of the conversations you have with other people who are also intelligent enough to connect the dots and add integrative insight. 😉
More power to all of us, in equal measure.
Yes indeed.
Without constructive input from others I can’t do much.
Still, I think a couple of my insights as to how it might all integrate are original as far as I know at present but the more I get into it the more overlap I find with things others have said in the past.
The question of originality or not will have to be for others to decide (assuming we are creating something worth while).
This is why I always name the other people involved in my area of research with solar-planetary relations when I post original articles. They can share the blame if it all goes Pete Tong. 😉
Joking aside, it seems to have encouraged those people to join in discussion of my original articles and a constructive, collegial atmosphere prevails in the comments sections.
Stephen Wilde and Tallbloke:
I strongly support your points about the importance of ‘where we go from here’ and the triviality of ‘who said what first’. But I write to explain my “Ahem!” comment. And I know this post is a whinge. I apologise if it offends, but I need to ‘get it off my chest’.
For several years I was widely vilified for pointing out that the ‘mass balance argument’ is invalid and that there exists no clear evidence of the cause of recent atmospheric CO2 increase. And I consider the matter of that cause to be important because I want to know what it is. Perhaps it is the anthropogenic emission: I don’t know.
It pleases me that others are now starting to study the matter because they may determine the cause of the CO2 rise which my efforts have failed to determine.
However, it is hard to take when people who have recently joined the study pretend I have not had to suffer the attacks and vilifications used to demean me because I refused to accept the dogma which had to be exposed before their studies could occur.
I appreciate that this post will be thought an offensive distraction to the thread, but if you had had to put up with what I have then you would understand the strength of my feeling. And I will understand if this post is thought so offensive that my withdrawal from this thread is desired.
Richard
No problem Richard. Going through years of vilification and being subject to irrational attacks is tough work. But I think we’re over the worst, and being openly sceptical of impending catastrophic man made climate change isn’t as hazardous to your career as it was around 10 years ago.
Richard.
As a relative newcomer I had no idea what you had been through in relation to the mass balance issue.
The record no doubt shows what happened and when the truth comes to the fore I hope that you obtain some satisfaction.
I hope you continue here and elsewhere. Although you remained ‘on the fence’ some of your past comments helped give me the confidence to continue in the face of a lot of abuse on other sites.
I was contemplating enthalpy. Changes in energy content of oceans and atmosphere are relevant to their behaviour and would be necessary variables in the algorithms behind any climate model.
Unfortunately enthalpy is not normally measurable, but needs to be calculated after a considerable amount of measurement and analysis or approximated from temperature, RH data and gas composition. As a real-time, measurable quantity it is not a practicable way of describing atmospheric conditions.
Until someone can design an enthalpy measurement device cheap enough to install at every weather station and simple enough to be operated by their readers, enthalpy measures will always be approximations from station data.
After further thought I submit a classification for terrestrial interactions between CO2 and temperature, with some examples.
Which examples are valid or invalid, I leave to you. You can probably suggest more examples and more categories.
Temperature drives CO2
Onset of interglacial.
End of interglacial.
Seasonal and weather related exchange of CO2 between atmosphere and oceans.
11 year solar cycle.
Feedback loops (Warming)
Acceleration of warming and CO2 in the millennium following onset of an interglacial.
Release of CO2 from warming seas around Antarctica.
Long term release of CO2 from tropical oceans.
El Nino
cAGW forcing
Feedback loops (Cooling)
Declining CO2 and temperature approaching the end of an interglacial.
Onset of a Snowball Earth period.
Long term absorption of CO2 by cooling tropical oceans.
La Nina
CO2 and temperature independant
Annual cycle of CO2 concentration.
20th/21st century increase in CO2.
Normal CO2 release from volcanoes.
CO2 release from impact events such as Chixulub.
CO2 drives Temperature
Massive vulcanism such as the Deccan Traps.
Permean –Triassic extinction event
cAGW.
Emergence from a Snowball Earth period.
Gentlemen:
Getting back to the original question of whether the recent increase in atmospheric CO2 might have been caused by temperature rather than by man-made CO2, allow me to draw your attention once again to the SST vs. CO2 graph I posted earlier:
The short-term ENSO-driven temperature fluctuations generate only tiny wiggles in the CO2 plot, and applying regression analysis to these wiggles shows that a 1C SST increase causes an increase of only about 3ppm CO2, far too little to explain the 80 ppm CO2 increase since 1960.
But when we eyeball the wiggles out we see CO2 and SST increasing together in the approximate ratio of 150ppm CO2 per degree C. Could we be looking at a long-term, SST-dependent release of sequestered CO2 here? Well, I’ve argued that the post-1976 increase in SST was caused by the release of ocean heat stored during earlier periods of high solar activity, and TB argues along similar lines, and so does Stephen W, so I guess I have to admit that we could be.
And if we are looking at an SST-dependent release of stored CO2 it should be possible to fit the Mauna Loa CO2 record by applying scaling factors and time constants to the historic SST record. Can we fit the CO2 record this way? Indeed we can.
I’m not sure what the physical significance of the 100 year time constant and the 3.2 ppm/degree conversion factor is, but the main thing is that we get a fit. (And we don’t always get one. I didn’t when I repeated the exercise using the SSN record.)
So do I now believe that the CO2 increase was natural? No, I still think it was anthropogenic. I just thought you might be interested in seeing how incorporating observational data might help you reinforce your arguments. 😉
Hi Roger A: I’m quite willing to accept that some of the increase is due to human emission. Bart seems to think it’s about 5% according to his calcs. I’m willing to give a bit of latitude on that figure. But the other important thing which came out of the WUWT discussion was the realisation that various sinks operate on various timescales, and have more or less inertia.
So that’s why I agree with Richard that we don’t have enough data to settle the issue yet. However, if temperature drops over the next decade (or more), and Bart’s estimate is correct of ~15 years lag, and co2 starts to fall too, then nature will have performed the crucial experiment and we can all lament the fact that burning more coal is only going to warm our hearth’s and not the atmosphere.
RA:
In the light of that data presentation what makes you still think it is anthropogenic ?
You should at least be undecided.
TB:
I judge the inflection point when many disparate phenomena changed trend to have been around 2000.
On that basis we must start to see a more definite system response to reduced solar activity from 2015 onwards.
Since the warm AMO phase will be fading by then and the current trend of cooling in the tropics will have consolidated a bit more that timescale is plausible.
Of course, if things start moving the other way again despite a continuing quiet sun then we will have to consider munching our hats unless some other factor can be shown to be confounding our expectations.
At least we have put forward propositions that can be tested, unlike AGW theory which seems to accommodate every possible combination of climate phenomena and to morph as necessary over time to try and avoid discredit.
Nice of the sun to give us a global testbench such that all we now need to do is observe for a few years.
In the meantime we can polish up the details and create better data in support of the various components as RA suggests.
Stephen,
I think it’ll be nearer 2020 before the co2 responds if Bart is right. The Sunspot number didn’t drop below the ocean equilibrium value I heuristically determined until 2004. That’s when ocean heat content started to fall.
I agree that the time is approaching when we need to consolidate and quantify. The freeform period where the ideas have been thrashed out has been so enjoyable I don’t want it to end though. 🙂
Stephen:
“In the light of that data presentation what makes you still think it is anthropogenic ?”
The good fit in the graph shows that SST changes could have caused the CO2 increase provided the numbers I used to obtain the fit – the 100 year time constant and the 3.2 ppm/degree C conversion factor – are physically realistic. But they could be – probably are – way off. And even if they aren’t the model still doesn’t explain what happened to the 1.3 trillion tons of anthropogenic CO emitted since 1900, which had it all stayed in the atmosphere would have raised CO2 concentrations by about 160 ppm.
I get at least as good a fit when I use fossil fuel emissions to model atmospheric CO2, and here I know that 1 gigaton of emitted CO2 adds 0.12 ppm of CO2 to the atmosphere, all other things being equal. I can’t be sure about the CO2 time constant, but the 50-year number that gives the best fit matches the IPCC’s low estimate, which I assume has some backup somewhere. I also don’t have to worry about explaining where the 1.3 trillion tons of anthropogenic CO2 went because the model explains it.
TB:
As far as I know the only conclusively documented impact of increased CO2 is that it makes plants, crops and trees grow faster. So why don’t we humans just admit that we dunnit and take the credit?
Roger A:
The only important thing at this stage is to be scientifically correct in assessing the uncertainty concerning the relative scales of human and natural contribution to the increase in the airborne fraction.
If it’s mostly natural, then mitigation is expensive and pointless.
If sensitivity is low then mitigation is pointless even if the human contribution to the increase is large. (though that’s an argument for a different day).
You said:
the 50-year number that gives the best fit matches the IPCC’s low estimate, which I assume has some backup somewhere.
I thought your best fit was achieved by assuming a 5 year residence time?
Never assume the IPCC has backup for it’s statements in the literature. The brief synopsis of the greenhouse effect itself given in the AR4 report said it was properly explained later on, but isn’t.
Roger Andrews:
You say;
“And even if they aren’t the model still doesn’t explain what happened to the 1.3 trillion tons of anthropogenic CO2 emitted since 1900, which had it all stayed in the atmosphere would have raised CO2 concentrations by about 160 ppm.”
(n.b. I have altered “CO” to “CO2” in the quotation because I interpret it to be a misprint: RSC).
I respond that – in common with all CO2 emissions of all kinds – the anthropogenic CO2 entered the carbon cycle. This is not an evasion of your point because that is the only possible answer which can be given in response to a question about what happens to any CO2 emission from any source.
For example, natural coalfield fires vary their emission of CO2 in unknown ways but they are estimated to emit a very similar amount of CO2 as the anthropogenic CO2 emission (i.e. they each emit ~3% of the total CO2 emission).
So, your question could equally be posed as
“And even if they aren’t the model still doesn’t explain what happened to the 1.3 trillion tons of CO2 emitted by natural coalfield fires since 1900, which had it all stayed in the atmosphere would have raised CO2 concentrations by about 160 ppm.”
Therefore, I respectfully submit that your question has no value unless and until a quantitative understanding of the entire carbon cycle is obtained.
Richard
When I look at the atmospheric 14-C decline after the atomic weapons tests and the partial test ban treaty, the aggregate of N and S hemispheres [New Zealand and Austria] declines with an approximate ‘half-life’ of 5 to 12 years.
What am I missing that the IPCC do not?
Another point from the data as can be seen in the diagram
http://en.wikipedia.org/wiki/File:Radiocarbon_bomb_spike.svg
is that the time to equalization of the two hemispheres is of the order of roughly 5 years. I don’t know where this number would fit in the definition of a “well mixed gas”. Is it a high or low figure?
[I presume mixing within a hemisphere is taken as being far more rapid than this.]
Richard:
My SST “model”, if such it can be called, assumes that changes in SST were the only control on CO2 generation and that any CO2 generated by other sources effectively disappeared into space, and the same of course goes for my emissions model. It’s the same approach as you used for the models you discuss in your paper, each of which “assumed that a single mechanism is responsible for the rise in atmospheric CO2 concentration ..”. The only significant difference is that I used observational data and I don’t think you did.
Roger Andrews:
At September 14, 2012 at 4:26 pm you say to me
“My SST “model”, if such it can be called, assumes that changes in SST were the only control on CO2 generation and that any CO2 generated by other sources effectively disappeared into space, and the same of course goes for my emissions model. It’s the same approach as you used for the models you discuss in your paper, each of which “assumed that a single mechanism is responsible for the rise in atmospheric CO2 concentration ..”. The only significant difference is that I used observational data and I don’t think you did.”
Pardon!?
We fitted Mauna Loa data to HadCRUT data. Both are “observational”
Richard
PS We also input estimates of annual anthropogenic CO2 emissions. These are estimates provided by the IEA so perhaps that is what you meant by our not using “observational” data.
Michael Hart:
I think your post at September 14, 2012 at 4:12 pm may be confusing the decay rate for a pulse of CO2 into the atmosphere and the e-folding time of that pulse.
As you say, the 14C data indicates the molecules of a pulse of CO2 are removed from the atmosphere in ~5 years.
However, the pulse increases the concentration of CO2 in the atmosphere. That concentration takes time to reduce and the e-folding time is the time required for the concentration to return to half the level it had prior to the pulse. The value of e-folding time is determined by the sequestration model used. The IPCC uses the (ridiculous) Bern Model.
Richard
Richard:
Sorry, missed that. So I effectively re-invented the wheel again. 🙂
Roger Andrews:
You say to me
“Sorry, missed that. So I effectively re-invented the wheel again”
Probably not.
Please note that our purpose was to disprove ‘natural’ or ‘anthropogenic’ cause of the observed change to atmospheric CO2 emission. So, we used attribution studies to determine if we could eliminate causes, and we failed to determine that.
We produced 6 models: 3 were altered by temperature and 3 were altered by the anthropogenic emission. Every one of them provided a perfect fit (within measurement errors) to the Mauna Loa data for annual atmospheric CO2 levels.
And we point out that other models are probably possible. I suspect you have produced one of a large number of other possible models.
Perhaps now you can understand why
(a) I refuse to accept that the cause is known to be anthropogenic, or natural or a mixture of anthropogenic
and
(b) why I point out that data indicates nothing when it can be shown to represent anything.
Richard
“data indicates nothing when it can be shown to represent anything.”
Absolutely.
But now we have data from AIRS that shows plumes of CO2 from sun warmed ocean surfaces and
nothing at all downwind of human population centres.
Explain how humans can be responsible for atmospheric CO2 increases in light of that.
Stephen Wilde:
You ask,
“But now we have data from AIRS that shows plumes of CO2 from sun warmed ocean surfaces and
nothing at all downwind of human population centres.
Explain how humans can be responsible for atmospheric CO2 increases in light of that.”
I cannot, but that – of itself – only proves my ignorance. I explain as follows.
Firstly, as I have repeatedly explained including above in this thread where I wrote,
“Also, our referenced paper analyses the dynamics of seasonal sequestration at each measurement site. These dynamics demonstrate that local sequestration processes can easily sequester all the emitted CO2 (both natural and anthropogenic) of each year. This provides a paradox: the atmospheric CO2 is rising but the sequestration processes have dynamics which indicate all the emitted CO2 could be expected to be sequestered.”
This paradox is explicable as being the carbon cycle adjusting towards an altered equilibrium.
The rate constants of some processes in the system are years and decades so the system takes decades to attain an altered equilibrium.
If the system is altering towards an altered equilibrium then that changes the question of causation to being, ‘What caused the system to change?’
One of the possible answers to that question is the anthropogenic emission. Now assume that is the correct answer, then how the system adjusts may have no obvious spatial relation to the sources of the anthropogenic emission (although the anthropogenic emission is the cause of the change that has happened to the system).
As you know, I am not willing to position myself on the ‘anthropogenic’ or ‘natural’ sides of the causation argument. However, the AIRS data is the best evidence for a natural (i.e. non-anthropogenic) cause. And it is yet another ‘nail in the coffin’ of the false claim that the anthropogenic emission is “accumulating” in the atmosphere.
Richard
Stephen says:
But now we have data from AIRS that shows plumes of CO2 from sun warmed ocean surfaces and
nothing at all downwind of human population centres.
Hi Stephen, Are there similar maps for the rest of the year that you know of? I’d like to take a look.
Whoops.
Sorry Richard. When I posed that question I was meaning to address it to RA rather than you since I am aware of your cautious position.
TB.
Haven’t seen any other versions of that particular AIRS chart. A year’s sequence would be nice but would probably only show seasonal changes.
It is really multidecadal changes we need to see and that will be some time coming.
Stephen Wilde:
I apologise for mistakenly thinking your question was addressed to me. I did not intend to disrupt your conversation and I hope my answer has not.
Richard
TB:
Might this might be what you’re looking for?
Richard,
By “e-folding” do you mean ln(2) divided by tau, the time constant? I put the single quotation marks around the term ‘half-life’ out of a learned habit of using “half-life” and not “tau” or “time-constant” from my grad student days in Chemistry. And, of course, it is not related to the radioactive half-life of nuclear-decay by 14-C. Terminology varies across the sciences, but my point still stands as far as I can see.
The pulse of 14-C is vanishingly small when compared to the concentrations of “cold” isotopes of Carbon, so it has effectively zero affect on the bulk concentrations and thermodynamics of the whole system [unless you make the case that the individual physical/chemical responses to the three carbon isotopes varies by a lot, which I would dispute]. That is why radio-chemical tracers have contributed so much to Chemistry/Biochemistry and remain the de-facto gold standard.
My second point, about exchange between N and S hemispheres was a slightly different line of thought prompted by considerations of the classical means by which perturbation-kinetics have been used to study fluxes in complex systems while causing minimal disruption to pre-existing steady-states or equilibriums. I simply don’t know what authors of IPCC reports would consider to be a fast/slow exchange of atmosphere between the hemispheres.
My suspicions are that some have given the matter little thought with respect to this particular data set.
Michael Hart:
Please forgive my lack of clarity. I will try to correct that with this reply.
You ask me,
“By “e-folding” do you mean ln(2) divided by tau, the time constant? I put the single quotation marks around the term ‘half-life’ out of a learned habit of using “half-life” and not “tau” or “time-constant” from my grad student days in Chemistry. And, of course, it is not related to the radioactive half-life of nuclear-decay by 14-C. Terminology varies across the sciences, but my point still stands as far as I can see.”
I answer:
By definition, the e-folding time is the time required for the amplitude of an oscillation to increase or decrease by a factor of e (2.718…).
And
By definition, the half-life is the time required for a quantity to fall to half its value as measured at the beginning of the time period.
And
By definition, the residence time of a CO2 molecule in the atmosphere is the period from the time of its injection to the atmosphere to the time of its sequestration from the atmosphere.
With respect to a pulse of CO2 into the atmosphere, the estimates are so coarse that e-folding and half-life are often used interchangeably. A good discussion of the issues is on Judith Curry’s blog at
Various estimates range from 14 years to 250 years.
But please note that with respect to the decay of a pulse of CO2 injected into the atmosphere its e-folding time and its half-life are NOT the same as its residence time (i.e. the residence time of all the injected molecules). The residence time is ~5 years as indicated by 14C put into the atmosphere by A-bomb tests.
This short residence time results from the high rate of cycling of CO2 in and out of the atmosphere.
The wide range of e-folding times results from the different models (i.e. the different assumptions) of carbon cycle behaviour which are used to determine it. I tend to think it is at the lower end of the estimates (i.e. I think it is less than 20 years) but this is a minority opinion. The IPCC is not clear but seems to think the e-folding time is ~240 years (derived by back calculation by me: RSC).
I hope this explanation is better.
As to your other point:
As a result of the Hadley Cells there is little atmosphere exchange between the hemisphere and it is mostly a result of storms. The IPCC does not discuss the matter. Little empirical data exists.
I hope this additional answer is more helpful than I think it is.
Richard
RA:
AS TB said to EM previously
“No, what we’re saying is that the tropical oceans is where most of it gets released, not where it seasonally accumulates in the following months.”
Although I would go further and narrow it down to the sunny regions beneath the subtropical high pressure cells.
That lengthy display shows where it accumulates and not where it originates.
From the AIRS chart that I use one can see where it originates.
As you can see, the accumulation is indeed increasing over the years in the northern hemisphere.
What must be happening, therefore, is that the oceanic sources under the subtropical high pressure cells are currently running ahead of the sinks especially in the northern hemisphere with a consequent rise globally.
The flow rate of CO2 from sources to sinks is ‘backing up’ in the northern hemisphere but not so much in the southern hemisphere.
The southern hemisphere is mostly ocean with large areas of cooler waters towards the Antarctic so the sinks are much more effective as one moves towards the south pole.
The northern hemisphere is mostly land masses and so has much reduced cooling ability for oceanic water moving from equator to poles. That greatly reduces the efficiency of the northern oceanic sinks.
Furthermore, the land mass configuration directs warm water into the Arctic Ocean via Spitzbergen thus further reducing the ability of the northern hemisphere to operate as an oceanic sink for CO2..
So at any time when oceanic sources are running ahead of sinks you will see just that pattern of accumulating CO2 in the northern hemisphere.
When oceanic sources become less productive then the sinks in the northern hemisphere can catch up again and CO2 concentrations globally but especially in the northern hemisphere will decline once more.
That pattern is exactly what I would expect to see if the natural sources are running ahead of the sinks given the Earth’s landmass distribution.
Richard Courtney:
You say: “The wide range of e-folding times results from the different models (i.e. the different assumptions) of carbon cycle behaviour which are used to determine it. I tend to think it is at the lower end of the estimates (i.e. I think it is less than 20 years) but this is a minority opinion.”
I’m not sure it is a minority opinion. There are a lot of published “residence time” estimates (which I assume are equivalent to time constant or e-folding estimates) in the 5 to 15 years range, including the 36 separate estimates summarized by Segalstad (1998) and the later estimates of Schwarz (2007) and Essenhigh (2009). However, your opinion certainly doesn’t dovetail with the IPCC AR4 estimates, which range from 50 to 200 years.
I’m not qualified to comment on who’s right and who’s wrong here, or even on the question of whether there really is such a thing as a CO2 time constant. But if we assume that anthropogenic CO2 emitted to the atmosphere does in fact get reabsorbed in accordance with the expression e^-(t/T) we can use CO2 time constant values to estimate how much of the recent rise in CO2 was anthropogenic. And here’s what we get when we do this (Note: I believe these numbers are correct but can’t guarantee it. An independent check would be helpful):
Time constant 5 years: 20% anthropogenic
Time constant 10 years: 35% anthropogenic
Time constant 20 years: 60% anthropogenic
Time constant 50 years: 100% anthropogenic
I draw two preliminary conclusions from these results:
1. If it can be shown that CO2 time constants are indeed in the 5 to 10-year range (the average of the 38 estimates cited above is 7.7 years) then we can make a case that most of the recent CO2 increase was natural and not anthropogenic and I would have to revise my opinions accordingly.
2. The IPCC’s low-end 50-year time constant fits observations but values significantly higher than 50 years don’t, and a case can be made for falsifying them on these grounds. Certainly if the 250-450 year estimates of Archer et al. (2009) were anywhere close to reality – and so long as the anthropogenic emissions numbers aren’t gross overestimates – present-day atmospheric CO2 concentrations would be far higher than they are.
I don’t claim that this analysis “proves” anything, but maybe it could serve as a point of departure for further discussion.
Roger A Says,
“There are a lot of published “residence time” estimates (which I assume are equivalent to time constant or e-folding estimates) in the 5 to 15 years range”
see the bolded part of Richard’s earlier post. I’m not sure what difference this makes to your calcs.
By definition, the e-folding time is the time required for the amplitude of an oscillation to increase or decrease by a factor of e (2.718…).
And
By definition, the half-life is the time required for a quantity to fall to half its value as measured at the beginning of the time period.
And
By definition, the residence time of a CO2 molecule in the atmosphere is the period from the time of its injection to the atmosphere to the time of its sequestration from the atmosphere.
With respect to a pulse of CO2 into the atmosphere, the estimates are so coarse that e-folding and half-life are often used interchangeably. A good discussion of the issues is on Judith Curry’s blog at
Various estimates range from 14 years to 250 years.
But please note that with respect to the decay of a pulse of CO2 injected into the atmosphere its e-folding time and its half-life are NOT the same as its residence time (i.e. the residence time of all the injected molecules). The residence time is ~5 years as indicated by 14C put into the atmosphere by A-bomb tests.
This short residence time results from the high rate of cycling of CO2 in and out of the atmosphere.
The wide range of e-folding times results from the different models (i.e. the different assumptions) of carbon cycle behaviour which are used to determine it. I tend to think it is at the lower end of the estimates (i.e. I think it is less than 20 years) but this is a minority opinion. The IPCC is not clear but seems to think the e-folding time is ~240 years (derived by back calculation by me: RSC).
Roger Andrews:
Firstly, I am grateful to Tallbloke for his earlier reply to your question because I also suspect you are using residence time as though it were e-folding time.
Please provide your calculations which provide your listed results. I am willing to comment on your results if I know how they are derived.
Richard
Richard, I’m really grateful for your willingness to spend time here and offer your expertise. We will resolve these confusions and I think we are making real progress in this topic area.
Tallbloke:
No problem but it is now Saturday evening so if serious input is needed I will not be able to contribute to it tomorrow.
Richard
Richard:
I’ve sent you a spreadsheet showing my calculations for a 50-year anthropogenic CO2 time constant (the best-fit anthropogenic case).
TB:
I admit to being confused about the different metrics used to quantify how long CO2 emissions stay in the atmosphere, and incorrect definitional assumptions on my part could indeed make a difference to my calculations. Here’s an example, I’ve assumed that a five-year residence time means a five-year time constant, but if it means that ALL of the CO2 emitted in a particular year is gone five years later then it’s about equivalent to a one-year time constant, in which case I get:
Time constant 1 year: 5% anthropogenic, 95% natural.
But they don’t affect my basic argument, which is that the length of time man-made CO2 stays in the atmosphere, regardless of how we choose to define it, is the metric we should be looking at if we want to determine how much of recent increase in CO2 was man-made and how much natural.
SW:
“From the AIRS chart that I use one can see where it originates. As you can see, the accumulation is indeed increasing over the years in the northern hemisphere.”
Well, I might be able to if you posted a link to the AIRS chart you use. 🙂
RA:
The increase in the northern hemisphere over the years comes from your graphic.
I should have made that clearer.
“the length of time man-made CO2 stays in the atmosphere, regardless of how we choose to define it, is the metric we should be looking at if we want to determine how much of recent increase in CO2 was man-made and how much natural”
That must be right given that demolition of the mass balance approach depends on a rapid local sequestration of human emissions such that the AIRS data fails to reveal any additional CO2 over or downwind of major population centres.
SW:
Ah. You’re referring to the carbon tracker simulation, apparently.
You say: “the AIRS data fails to reveal any additional CO2 over or downwind of major population centres.” I wouldn’t expect it to. Over the 2000-2008 interval covered by the simulation some 1,300 gigatons of carbon was exchanged naturally between the atmosphere and oceanic and terrestrial sinks but only 67 gigatons of anthropogenic carbon was emitted. It would be extremely difficult if not impossible to pick this small increment visually out of all the fast-moving technicolor swirls we see on the animation.
Although you might go back and take another look at China.
You also say: “demolition of the mass balance approach depends on a rapid local sequestration of human emissions”. If it could be proved that anthropogenic CO2 has a very short residence time then the anthropogenic origin theory would indeed be demolished, which is why I’m proposing that the residence time approach is the best way of addressing the issue. However, the fact that I’m presenting data that doesn’t support my own conclusions in an effort to get my point across doesn’t mean that I now agree that the CO2 increase was natural. It means only that I’m more concerned about finding out what’s really going on than I am in proving myself right.
Roger Andrews:
This post may be abrupt or curtailed because I must leave for somewhere else in an hour. My reply (even if incomplete) is provided now to show I have treated your work with respect and, therefore, as a priority.
I have downloaded your spreadsheet and studied it. I am now commenting as I said I would.
Firstly, I state what I understand to be your model. This enables you to correct any misunderstanding I have.
Your model seems to utilise the ‘mass balance’ argument.
It starts from the atmospheric CO2 burden in 1959.
Then for each subsequent year to 2010 it
(a) adds the anthropogenic emission
and
(b) determines the removal of CO2 from the atmosphere according to the calculation of e^-(t/T) where T is a time constant and t is a value of something not stated.
If my understanding is correct then I don’t need to know what t represents because I can comment on the model as it exists.
This is a very simple curve fitting model. If correct it explains nothing about any mechanism because it does not represent any mechanism. Several other equations can also be fitted to the curve.
The model implicitly assumes the anthropogenic emission is added to the atmosphere and the change in the atmospheric CO2 burden is because not all the emission is sequestered in any year. The assumption cannot prove or – in this case – disprove itself.
Hence, I fail to see the usefulness of the model, and I would welcome an explanation of this.
SOME BASIC MODELING PRINCIPLES
I make some basic observations on modelling. This is not intended as an insult to anybody: my purpose is to ensure all who follow our conversation can understand the meaning and importance of my statement that says,
“I fail to see the usefulness of the model”.
A model is a simplified representation of reality.
Being simplified, no model is an exact emulation of reality; i.e. no model is a perfect and no model is intended to be perfect.
A model is constructed for a purpose.
For example, a model of heat loss from a cow may assume that a cow is shaped as a sphere with the surface area of a real cow. And this simple model may provide an adequate quantitative indication of how heat loss from a cow varies with the cow’s metabolic rate. Thus, this hypothetical model may be very useful.
But that model of a cow cannot be used to indicate the movements of a cow. A model of a cow which includes legs is needed for that.
Another model of a cow may be constructed purely for the pleasure of the modeller. In this case it may be carved from wood and painted.
Possible purposes for models are infinite.
A model may have many forms.
It may be physical, abstract, algebraic, numeric, pictorial or an idea. If its form fulfils the desired usefulness then it is an appropriate model; i.e. it can fulfil its purpose.
Models of atmospheric CO2 concentration are usually constructed for one of two reasons; i.e.
1.
To represent variation of atmospheric CO2 concentration with time according to a specific idea (e.g. the idea that the variation is accumulation of anthropogenic CO2 in the atmosphere in a manner defined by a specific theory).
2.
To demonstrate that a particular suggestion for the cause of a change in atmospheric CO2 is not correct.
In this thread the intention is to disprove the cause of the recent rise in atmospheric CO2 is (a) natural or (b) anthropogenic. Your model assumes the rise has an anthropogenic cause and uses a curve fit to represent the change in atmospheric CO2. Other curves could also fulfil the fitting requirement and your curve does not represent a stated mechanism. Similarly, a variety of curves can be fitted by use of an assumption that the atmospheric CO2 is a smoothed response to temperature change. Hence, the purpose of your model is not clear and needs to be stated.
I hope the above considerations are helpful to discussion in the thread.
Richard
Richard:
Your prompt response is very much appreciated.
First on the specific question of “t”. It’s the elapsed time in years between the year in which the carbon was emitted and the year in which I calculate how much of it is left. As an example, taking 1970 as the year of emission and 2000 as the year of calculation gives t=30, which with a time constant of T = 50 gives us e^-(30/50) = 0.54 of the carbon emitted in 1970 remaining in the atmosphere in 2000.
On more general issues:
The model is a simple carbon-in-carbon-out model that uses only two assumptions. First, that anthropogenic emissions add carbon to the atmosphere, and second that much of this added carbon gets reabsorbed over time. (If it didn’t atmospheric CO2 concentrations would be about 70ppm higher than they are.) It makes no assumptions about reabsorption mechanisms and no attempt to identify what these mechanisms might be. All it’s designed to do is investigate whether anthropogenic carbon remains in the atmosphere long enough to explain the increase in ppm CO2 since 1959.
If CO2 remains in the atmosphere for decades or centuries, as the IPCC et al. claim, then the model shows that man-made carbon emissions can explain all of the observed increase in atmospheric CO2 (although I agree that this doesn’t prove that they did because other models can explain the increase too, as we’ve discussed.)
But if it remains there for only +/-5 years, as others claim, then the model shows that man-made emissions can’t possibly explain more than a small fraction of the post-1959 increase, and if it could be demonstrated that residence times are indeed this short then this in my view would constitute proof that most of the increase was natural.
So what I’ve been doing is culling residence time, e-folding time and half-life estimates from the literature, inputting these estimates (probably incorrectly in many cases) as time constants in the model and recording one single number – the ppm increase in CO2 since 1959. So far this has allowed me to establish that the +/-75 ppm post-1959 CO2 increase is somewhere between 95% natural and 100% anthropogenic, which doesn’t move the science forward very far. But at the same time I’ve been able to demonstrate that the IPCC’s claims of CO2 residence times of hundreds or thousands of years aren’t realistic because they don’t match observations, and I guess that’s progress of some kind.
I’m not using the model for curve-fitting. I’ve already done this, and the spreadsheet I sent you is the result. It shows that anthropogenic CO2 emissions explain the CO2 record since 1959 to within a ppm or two in all years assuming T=50, but again this doesn’t prove that they caused it.
I’m not sure whether this will be sufficient to satisfy you that the model has a valid application, but whether it does or not I would be grateful for your assistance on the question of how long man-made CO2 remains in the atmosphere because it may help me narrow the range of possibilities down a bit. If possible what I would like to get from you would be your high-low-best estimates expressed as CO2 time constants or half-lives. (If we use the standard expression tau = V/q it seems to me that CO2 “residence time” is fixed at at around 5 years, but feel free to correct me if I’m wrong.)
Thank you again.
Roger
Roger Andrews:
Thankyou for your understanding that a genuine haste to reply was the reason for the curt statements in my comments on your model, and my bluntness was not – and was not intended to be – offensive.
I address your responses to my comments on your model.
Your explanation of t agrees with what I thought from study of your spreadsheet. And this is a good illustration of need for proper awareness of assumptions when considering models.
I could have assumed my understanding of t was correct, but it was not stated by the modeller so I stated that I did not know what it is. You tell me my understanding of t was correct, but it may not have been and – in that case – my adoption of an erroneous assumption (about t) would have resulted in our talking at ‘cross purposes’.
Clearly, it is important to identify and to explicitly state all assumptions when modelling.
You say of your model,
“The model is a simple carbon-in-carbon-out model that uses only two assumptions. First, that anthropogenic emissions add carbon to the atmosphere, and second that much of this added carbon gets reabsorbed over time.”
Sorry, but NO!
Those are NOT the only assumptions used in the model. The most important assumption of your model is that the modeled system is invariate over time. And there are several ways in which the system may be varying (to alter atmospheric CO2 concentration) that have no relationship to the anthropogenic emission.
For example,
a new natural coal field fire may have started,
or
a new natural leak of gas or oil may have initiated from an oil field,
or
sub-sea volcanism may have emitted more sulphur and so has made an imperceptible change of 0.1 to ocean surface layer pH
or
etc..
Please remember that the volcanic sulphur possibility could alone be the cause of the observed increase to atmospheric CO2 concentration, and – if real – would be completely undetectable, but would not require nature to have provided any additional CO2 emission. And there are countless other possibilities – both unknown and unknown – which would cause the system to vary in ways which your model assumes cannot happen.
In other words, the greatest limitation of your model is its assumption of an invariate climate system, and you say you have not recognised that your model uses that assumption.
You say,
“I’m not using the model for curve-fitting.”
Agreed, your model IS a curve fit. And, as I said, that is why its usefulness is limited: several curves can be fitted to the data.
And you say to me,
“I’m not sure whether this will be sufficient to satisfy you that the model has a valid application”.
It is your model and not my model. Hence, it is for you to state the “valid application” which you think it has. In science all others – including me – have a duty to dispute your claim that your model can be validly applied for the purpose you state.
I explained (and in this post I have added to the explanation) that,
“I fail to see the usefulness of the model, and I would welcome an explanation of this”.
The scientific method says you need to tell me how and why your model is useful and those who consider your model – including me – need to challenge what you tell me.
I still fail to see the usefulness of the model, and I hope this answer has clarified my problem in understanding what usefulness it has.
In conclusion, you ask me
“I would be grateful for your assistance on the question of how long man-made CO2 remains in the atmosphere because it may help me narrow the range of possibilities down a bit. If possible what I would like to get from you would be your high-low-best estimates expressed as CO2 time constants or half-lives. (If we use the standard expression tau = V/q it seems to me that CO2 “residence time” is fixed at at around 5 years, but feel free to correct me if I’m wrong.)”
I thought I had answered all that in my post in this thread at September 15, 2012 at 12:16 am where I provided standard definitions of terms then I wrote,
“With respect to a pulse of CO2 into the atmosphere, the estimates are so coarse that e-folding and half-life are often used interchangeably. A good discussion of the issues is on Judith Curry’s blog at
Various estimates range from 14 years to 250 years.
But please note that with respect to the decay of a pulse of CO2 injected into the atmosphere its e-folding time and its half-life are NOT the same as its residence time (i.e. the residence time of all the injected molecules). The residence time is ~5 years as indicated by 14C put into the atmosphere by A-bomb tests.
This short residence time results from the high rate of cycling of CO2 in and out of the atmosphere.
The wide range of e-folding times results from the different models (i.e. the different assumptions) of carbon cycle behaviour which are used to determine it. I tend to think it is at the lower end of the estimates (i.e. I think it is less than 20 years) but this is a minority opinion. The IPCC is not clear but seems to think the e-folding time is ~240 years (derived by back calculation by me: RSC).”
Obviously, your question proves my explanation of my views was not adequate and I will try to do better if made aware of the failing in that explanation.
Please be assured that I am trying to be helpful. I have been attempting to determine the cause of the recent CO2 rise for decades and the more people involved the better: your interest may discover what I want to know so I wish to encourage your involvement.
Richard
Richard
No, none of the things you said offended me in the least. And I hope none of the things I say offend you.
And if I take my time replying it’s because I’m having to think about what I say to make sure that I’m not going off on tangents and that I understand what I’m saying before I say it.
Having said that, I now propose to forge ahead with my model (which isn’t really a “model” at all, but simply a mathematical formulation that calculates how much of the anthropogenic carbon emitted since 1760 remains in the atmosphere between 1959 and 2010) and see where it takes me.
But before I start I would appreciate some extra clarification on the question of why there are such enormous differences between a) the CO2 time constant and half-life estimates, according to which CO2 hangs around in the atmosphere for decades or centuries, and b) the residence time estimates, which as I understand them imply that it all goes away in 5-15 years. They can’t possibly both be right, so I have to assume that there’s something I’m not grasping here.
That’s all for now, but I may be back with more later.
Thanks for your help
Roger
Roger A: I’m interested in Richards response too, so I’ll say how I currently understand it so he can correct me too wherever I’m wrong.
An individual average co2 molecule is likely to be in the atmosphere for five years or so before it is re-absorbed. But in the meantime, more co2 is emitted, and because of the feedbacks the IPCC believes to be in the system (water vapour feedback etc) the increase in temperature they fondly believe to be caused also generates more co2 release. So the effect is predicated into the future, and the additional release of more co2 is thought to be inevitably ‘in the pipeline’ because of the ‘committed warming’. So the e-folding time is greatly extended by these theoretical considerations as the projected attainment of a new equilibrium where sinks once again overcome sources and start reducing the airborne fraction is pushed further into the future.
TB:
As an average molecule myself I know that there will be a small number of very smart molecules who will get there much faster than I do, a fair number of smart molecules who will get there faster than I do, a large number of middling molecules who will get there about when I do, a fair number of dumb molecules who will get there after me, a small number of really dumb molecules who will get there long after me and a few idiot molecules who may never get there at all. I think this works out to exponential decay, but I too will await Richard’s comments.
Roger Andrews and Tallbloke:
You both say you want an explanation of why ‘residence time’ is not an indication of ‘half life’ (or e-folding time).
The short answer is that residence time is a function of the mixing rate in and out of the air but half-life (and e-folding time) is almost independent of mixing rate. I explain this as follows.
(a) CO2 residence time is an indication of how long an average CO2 molecule put into the atmosphere stays in the atmosphere
and
(b) half-life (and e-folding time) is an indication of how the carbon cycle responds to an addition of CO2 into the atmosphere.
CO2 is cycled in and out of the atmosphere from several sources notably the oceans. Each year the oceans ‘breath’ CO2 out and in: this is part of what is called the ‘seasonal variation’ in atmospheric CO2 concentration. This ‘breathing’ is very large: each year the oceans emit an order of magnitude more CO2 than all human activity. And each year the oceans sequester a very similar amount of CO2 from the air. But they don’t emit and sequester the same molecules (they emit and sequester a similar number of molecules) each year.
Therefore, a CO2 molecule has a probability of being sequestered as part of that cycling. The residence time of a molecule in the atmosphere is ~5 years. This indicates that it has a 1:5 chance of being sequestered each year. Put another way, 1/5 of the CO2 in the air is cycled in and out of the atmosphere as the seasonal variation. Importantly,
RESIDENCE TIME IS A FUNCTION OF THE RATE OF MIXING OF CO2 IN AND OUT OF THE AIR.
So,
~20% of the CO2 in the atmosphere enters the atmosphere each year
and
~20% of the CO2 in the atmosphere leaves the atmosphere each year.
and
these percentages would not change if there were no annual increase to atmospheric CO2 concentration.
But there is an observed annual increase to atmospheric CO2 concentration. In a typical recent year, about 2% of the CO2 emitted to the air (from oceans, biosphere, human activities, etc.) does not return to e.g. the oceans. Put another way the annual CO2 rise is the residual of an inequality in the seasonal variation of the year.
Any annual CO2 rise will affect the annual emission and sequestration of the following year.
For example, if there is more CO2 in the air then that will increase the sequestration of CO2 into the oceans in the following year. The total effect will be a combination of the activities of several interactive mechanisms (this is discussed in the Section titled ‘Mechanisms of the Carbon Cycle’ in the paper I sent to you).
Therefore, the carbon cycle adjusts in response to any change (e.g. temperature rise, the anthropogenic emission, etc.).
The initial effect is an increase (or a decrease) to the CO2 in the air. And this will not make much difference to the proportion of CO2 cycled in and out of the air as the seasonal variation. Therefore, an annual increase (or decrease) to the CO2 in the air makes negligible difference to the CO2 residence time. The residence time is a function of mixing rate so an increase to atmospheric CO2 concentration of 2% only affects mixing rate by ~2%.
But the increase (or decrease) to CO2 in the air alters all the interactive mechanisms which emit CO2 to the air and sequester CO2 from the air.
For purpose of illustration and purely hypothetically, assume the annual rise of one year changed the carbon cycle such that the CO2 in the air was fixed at the new CO2 concentration. Then residence time would still be 5 years (because mixing rate does not change) but half-life would be infinite (because the new atmospheric CO2 level never falls).
HALF-LIFE (AND e-FOLDING TIME) IS ALMOST INDEPENDENT OF MIXING RATE.
IT IS A FUNCTION OF HOW THE CARBON CYCLE RESPONDS TO A CHANGE.
The half-life is a function of how the carbon cycle responds to a change. And nobody knows how the carbon cycle actually responds to a change so there are many possible responses that can be imagined and modeled.
I hope this explanation is adequate and sufficiently clear.
Richard
Brilliantly clear, thank you Richard.
When you say nobody knows how the carbon cycle responds to a change, I’m sure this is true so far as the long term big picture is concerned. But we do have some clues from examining modern accurate data which may enable us to make some reasonable extrapolations regarding a limited ‘all other things being equal’ typa analysis.
For example Bart’s points about the difference in lag time of co2 behind temperature for different sized swings in temperature. Clearly it would be perilous to take such extrapolations too far or claim strong significance for such an extrapolation, but might we reasonably epect such a method to put us in the position where we might legitimately pass judgement on the IPCC figure of ~240 years you back calculated as against the 15 years Bart calculated? Or at least be in a position to say the truth is very likely closer to one rather than the other?
Tallbloke:
Thankyou for confirming that my explanation was adequately clear, at least for you.
Of course, I was trying to explain why residence time is almost independent of half-life. It was my third attempt to explain it so I was trying to make succinct statements that I repeated in different ways.
The succinctness of my summarising conclusion seems to have been misleading. It said
“The half-life is a function of how the carbon cycle responds to a change. And nobody knows how the carbon cycle actually responds to a change so there are many possible responses that can be imagined and modeled.”
That was NOT intended to suggest that modelling has no value. On the contrary, models can explore what is possible and indicate what is not possible according to the available empirical data. Indeed, the paper I sent to you explains, demonstrates and illustrates this.
Importantly, extrapolations from models are predictions based on the understanding (or hypothesis) which is modelled. Comparison of real outcomes with the extrapolations either refutes the understanding (or hypothesis) or provides confidence to it.
Models are of fundamental importance for use In an infant science such as study of the carbon cycle. And I apologise if my summarising conclusion implied other than that fundamental importance.
Please note that this importance of models is exactly why I have a problem with the model of Roger Andrews. In an infant science a model needs to test a clearly defined understanding or hypothesis. But his model makes so many assumptions that if its predictions were to agree with subsequent reality then there would be no way to determine what that indicates. This is not to denigrate his model which may be very useful: but – as yet – he has not explained its usefulness.
Richard
Tallbloke:
In retrospect, I think my recent reply to you could be misinterpreted as being evasive in that it did not provide a specific reply to questions you asked me. Those questions were
“For example Bart’s points about the difference in lag time of co2 behind temperature for different sized swings in temperature. Clearly it would be perilous to take such extrapolations too far or claim strong significance for such an extrapolation, but might we reasonably epect such a method to put us in the position where we might legitimately pass judgement on the IPCC figure of ~240 years you back calculated as against the 15 years Bart calculated? Or at least be in a position to say the truth is very likely closer to one rather than the other?”
I answer: in principle, yes. But his model needs to demonstrate some predictive skill first. At present his model has no confidence except that he has presented it.
Richard
Richard:
Thank you very much for your very clear and detailed explanations, which help a lot but don’t tell me everything I would like to know. But I would like to know everything, so this is no fault of yours. 🙂
What I am most interested in getting now is a mathematical expression that allows me to convert residence time estimates to time constants. I can’t find a suitable expression in the literature so I’ve had a go at developing one myself.
I start by making two assumptions. First that CO2 in the atmosphere decays exponentially, and second that the “average molecule” is the 50th in a chain of 100 emitted CO2 molecules (or the 5 quintillionth in a chain of 10 quintillion) and that the residence time is when this molecule gets reabsorbed. This occurs when:
e^-(t/T) = 0.5
Where t is the elapsed time after emission (i.e. the residence time) and T (tau) is the time constant.
Solving for t/T gives:
ln(0.5) = -t/T
t/T = 0.69315
According to this the residence time is always 69% (let’s say 70%) of the time constant. So a residence time of 7 years corresponds to a time constant of 10 years and a residence time of 14 years to a time constant of 20 years etc. etc.
Is this a reasonable approximation?
Roger Andrews:
You ask me
“According to this the residence time is always 69% (let’s say 70%) of the time constant. So a residence time of 7 years corresponds to a time constant of 10 years and a residence time of 14 years to a time constant of 20 years etc. etc.
Is this a reasonable approximation?”
I answer: Yes, but so what?
Richard
TB:
You say: “When you (Richard) say nobody knows how the carbon cycle responds to a change, I’m sure this is true so far as the long term big picture is concerned. But we do have some clues from examining modern accurate data which may enable us to make some reasonable extrapolations regarding a limited ‘all other things being equal’ type analysis.” That’s what I’m trying to do.
And here are my latest results. Using my 0.7 residence time/time constant conversion factor – which Richard at least agrees is reasonable – I find that anthropogenic CO2 explains about 25% of the CO2 increase since 1959 assuming a 5 year residence time, about 60% assuming 15 year residence time and about 40% if we use the average of all the estimates (8 year residence time).
I also find that time constants much greater than 50 years aren’t plausible because they leave too much CO2 in the atmosphere. If the IPCC’s 240 year estimate was correct atmospheric CO2 concentrations would now be much higher than they are.
Are these findings valid? Well, I don’t know and maybe never will. But I like to think they represent progress of some kind.
Incidentally, could you point me to where Bart came up with his 15-year estimate? I’ve been back through the comments but can’t find it.
On the WUWT thread I think, but mentioned here too:
Your result looks like a compromise which could be lived with 🙂
By the way Richard, your history article has had over 540 views this last week. Good stuff.
Roger Andrews:
re your comment at September 18, 2012 at 5:43 pm.
You are reporting the performance of your model. However, your model is not validated; i.e. it is not yet demonstrated that your model emulates the real world. For your ‘findings’ to have any value your model needs to be validated.
Please note that this is not a ‘put down’. It is simply a statement of the way things are. (Science is a funny business).
Richard
Tallbloke:
About a dozen of those views of the history article will be me checking to see if there are comments I should answer.
Richard
[…] between 1910 and 1950 so it may be a lot less. Humans are only responsible for at most around 50% of the airborne increase in co2, so we are responsible for, at most, around 15% of 0.4C = 0.06C. So crippling our economies […]
[…] Tallbloke found a temperature CO2 link that is a very tight match. We know that the physics of CO2 says it dissolves well in cold water and comes out of dry hot materials. The images show CO2 higher in dry hot places and times. We can see cold water moving into warming conditions and CO2 higher down wind of them. Is it really all that much of a leap to think that maybe, just maybe, the atmospheric CO2 cycle is driven by temperature and water? […]