07 January 2010

CO2 and temperature for 800,000 years

First, the figure gives the answer on CO2 and temperature over the last 800,000 years:



Here we have each value of CO2 plotted against the temperature deviation from reference values -- with the temperature being for 1000 years before the corresponding CO2 value. The correlation for this is about 0.89 (R), meaning that you could explain 79 percent (0.89*0.89) of the variation (R^2) in CO2 by looking at the temperature. As is mentioned in my post Does CO2 correlate with temperatures, where we saw equally high correlation between temperature and CO2 (even higher if you give CO2 a 20-30 year lead on temperature), and quite a few times in the comments, correlation is not causation. Could be that both temperature and CO2 are pushed around by something else. More about that in a moment.

It's common to see the claim that temperature 'leads' CO2. Loosely speaking, this means that temperatures generally change before CO2 does, and that there is a consistent pattern to the connection -- if temperature rises, CO2 does as well. This is true; it leads by about 800 years [Caillon and others, 2003]. The amount of lead also seems to depend on whether you're in a glacial period or (as we are now) an interglacial. The correlation between temperature and CO2 is lower (by a very small amount) if you take their values for the same time (drops to 0.88). And it drops by more if we take the temperatures for 2000 years before the CO2 value, to 0.87, indicating from this very simple approach that the lead is between 0 and 1000 years, probably closer to 1000 -- as is found from the more serious approach in the reference.

So there's some interesting science to do, to understand why that lead exists, why it is many hundred years (rather than a few dozen years, or a few thousand), and why the lead time depends on whether you're in a glacial period or an interglacial period. Oddly, to me, most of the time that people mention this lead relationship, they are not referring to any of this.



Rather, they want to conclude, or for you to conclude, that because in the ice age record temperature changes before CO2, that a) the current rise in CO2 is also caused by temperatures (sometimes citing the medieval warm period as the warm time that is causing the current rise in temperature) and b) that CO2 doesn't have any affect on temperatures.

Have another look at the figure.  See that point sitting way the heck away from all others?  That's the most recent value.

You don't need a lot of scientific background to see that whatever it is that produced that CO2 value, it certainly was not the relationship that held for the other 798,000 years of the temperature - CO2 record.  We can be more precise about it.  The best fit line (the one that gives that nice high correlation) through the data points is CO2 = 266 + 8*T.  So, for temperatures at the reference value, we expect a CO2 level of 266 parts per million.  For temperatures 10 C below reference, we expect CO2 levels of 186 parts per million.  Both of these accord fairly well with what we do see in the record -- except for the most recent CO2 value.  As you can see from the plot as well, there is indeed scatter around the best fit line.  The standard deviation is about 11 ppm.  That means we're not surprised to see values 22 ppm away from the line (2 standard deviations), and, given 799 data points, we expect a few to be 33 ppm away.  On the other hand, a value 9 standard deviations away -- the case for the current CO2 levels --  is ludicrous.  Something must have changed.

We can turn this around, and ask: "If the relationship that held for the previous 798,000 years still did, what would the temperature need to have been to give the observed CO2 level -- maybe CO2 is so sensitive to temperature that we simply have a one-time temperature observing problem?"  Certainly the ice sheet temperatures do have their own observing issues.  As I've mentioned, all data have problems.  Still, same as we know that, we also know something about the size of the problem.  So, turn the equation around and solve for the temperature that would correspond to the observed (modern) CO2 level.  That's a temperature anomaly of about 13 C (20 F) -- equal to the entire range from the very warmest interglacial to the very coldest glacial!  And that had to have occurred 600-1000 years ago, by the lead relationship.  Meaning there had to have been an absolutely enormous warming (many times larger than even the highest medieval warm period values) and nobody noticed it ... or else the previous relationship between temperature and CO2 broke down.

Since we know what did cause the CO2 rise -- human activity -- we're not really surprised to see this answer. The old relationship did get broken.  Rather than CO2 rising because the oceans released CO2 to the atmosphere, it is because humans have been burning fossil fuels and making cement.  But we do like to be able to arrive at our conclusions from different directions.  Here, we need only to look at the temperature and CO2 values themselves to see that the relationship that used to hold has broken down in the modern day.  Don't need to know the first thing about isotope geochemistry (which is explained nicely in that faq by Jan Schloerer) to see this.

Let's return, though, to the fact that correlation is not causation.  There are generally 4 options when a correlation is seen: it's chance, A causes B, B causes A, both A and B are caused by something else.  We use statistical tests to decide whether chance is plausible (in this case, as with Does CO2 correlate with temperature?, the answer is a clear no).  So one of the other 3 is involved.  We can reject CO2 being the (sole) cause of temperature changes (it can certainly feed back on temperatures) because it is the temperatures that changes first.  So it's one of two now -- either temperatures drive CO2 (with feedback from CO2 back to temperatures), or something else drives changes in both temperature and CO2 (and perhaps manages to change temperature sooner than CO2).

Oddly, most of the comments about temperature leading CO2 ignore the last option -- in other words, they assume that correlation is indeed causation.  Now ... is there anything in the universe that could affect temperatures?  Could any of those things also affect CO2 levels?  That will be a separate post.  But start thinking about it yourself. Richard Alley gives away the answer in his talk, to which I referred you Tuesday.

Going farther:

I've labelled this a 'project folder' post in part for the questions I mentioned inside it -- why the lead is what it is (or at least was what it was), and so forth.  But also because the things I did you can do yourself.  The data I used were from a BBC news article on ice and CO2.  You can get it yourself and experiment with the lead relationships, find the CO2 response (slope of CO2 versus temperature in the regressions), and the like.

You can go farther as well.  For instance, are the correlation and slope different in different subsections of the record?  Is it different for different temperature ranges? (compute the slope for only temperatures from -10 to -5, -5 to 0, and so forth)  Does the correlation improve if you use the logarithm of CO2 rather than CO2 levels themselves? (if the only thing involved were the radiative properties of CO2 and its connection to temperatures, we would expect a 'yes' here)

37 comments:

gary thompson said...

i enjoyed the column but i was waiting for you to explain why the current co2 level outlyer (that is around 365 ppm) didn't correspond to a temperature rise of around 12.5C (from solving the equation for T given CO2 level). either the CO2 measurement is wrong (doubtful - i think everyone agrees with these levels) or there are other forces at play. don't you think we've reached a saturation of CO2 and that the temperature changes caused by that have stopped?

Robert Grumbine said...

The paragraph you want is:

Since we know what did cause the CO2 rise -- human activity -- we're not really surprised to see this answer. The old relationship did get broken. Rather than CO2 rising because the oceans released CO2 to the atmosphere, it is because humans have been burning fossil fuels and making cement. But we do like to be able to arrive at our conclusions from different directions. Here, we need only to look at the temperature and CO2 values themselves to see that the relationship that used to hold has broken down in the modern day. Don't need to know the first thing about isotope geochemistry (which is explained nicely in that faq by Jan Schloerer) to see this.


Follow up with Jan's FAQ for more detail as to why we know the 365 ppm is due to human activity.

carrot eater said...

Last time I looked, there was not a strongly supported mechanism for exactly how CO2 comes into and out of the atmosphere during the ice age cycles. The simple difference in ocean solubility is maybe insufficient.

Will you be addressing the current hypotheses on that matter? If it is addressed in the control knob lecture, then that will suffice, and I'll have to set some time aside for it.

John Mashey said...

Nice discussion, and I really like the chart.

Lou Grinzo said...

I'm not a climate scientist, so by all means feel free to point and laugh at me for what follows--provided you also explain why the notion I'm about to explain is wrong. (We're all here to learn from each other, after all. I've tried to research this on my own and found nothing published. Perhaps my Google skills aren't what I thought they were.)

I think one of the factors we're just starting to appreciate is the way the land configuration in the northern hemisphere might magnify a warming trend. There's a large amount of land in Russia, Canada, and Alaska that's not at the north pole but close enough to support permafrost, and just far enough away to start melting pretty quickly.

The latest data I've seen says there is about 1.6 trillion tons of C in northern permafrost, in addition to methane hydrate deposits. Given the amount and location, it seems primed to trigger a very significant positive feedback once warming starts for some unrelated reason.

And there was the BBC story just yesterday about the surprisingly large observed levels of methane emissions in Siberia. So it could be a multiple-stage process: Warming (for reason X) frees up some hydrates which feeds warming, and that in turn causes more hydrate release plus more permafrost melting that results in a mix of CO2 and methane emissions.

OK, this is the part where you laugh and point.

jg said...

Lou,
I'm no smarter than you, but I'm going to attempt to address your comment. This way, anyone laughing will laugh louder at me.

I posted a diagram here that I redrew from Nature 19 Oct. 2000:

carbon reservoirs

It shows the sizes of the types of carbon reservoirs: deep sea, land, surface ocean, atmosphere. This diagram suggests the carbon content of the terrestrial biosphere is 2,100 Pg while the deep ocean is 38,000 Pg, making the ocean a much bigger player than the northern lands.

I've read papers in Nature that describe what you cite: that the northern hemisphere tundra, peatlands, and permafrost will add a lot of carbon as the north warms. I think this scenerio gets less attention in Alley's talk and in this post because of the greater reservoir of the ocean and perhaps because the ocean is a better candidate for the initial increase of CO2 needed destabilize northern hemisphere icesheets: e.g., as orbital changes increase the warmth the planet, the north stays frozen till some other reservoir releases the CO2 which then gets the ice sheets melting.

thanks,
jg

carrot eater said...

Lou, I don't know why you think people should laugh at you. The article you read was pretty much in the context you imagine - that warming can activate a permafrost feedback. There is surely some discussion if you go looking in the IPCC report WG1. Water vapor and albedo get more attention, so they're probably assigned more importance in the near term, but I don't have the numbers/probabilities in front of me.

Robert Grumbine said...

Lou:
Your question is closely related to carrot eaters observation that we don't have a well-defined mechanism for CO2's flows. (It's more a problem for how to get the CO2 to rise fast enough at the end of ice ages, than for the slower decline as you enter an ice age.)

The permafrost carbon stores may well be part of the story for how CO2 can rise so fast at the end of ice ages. It is certainly a significant concern as a feedback on present atmospheric carbon dioxide levels. Bad enough for human CO2 to go up there, but if it starts extracting CO2 from buried reservoirs like the permafrost, things get even uglier.

jg's (as usual) nice graphic on the reservoirs is good as far as it goes. One thing missing for our current discussion is the permafrost reservoir broken out from the 'terrestrial'. More important is the size of fluxes between the different reservoirs. While the deep ocean dwarfs terrestrial, the terrestrial has a much faster exchange rate with the atmosphere (I assume that's what jg means by the 18 year for terrestrial vs. the 1250 year for deep ocean, but fluxes in Pg, which I think are available ... somewhere, perhaps Jan Schloerer's FAQ or cites therein, would be more useful here). For the situation at hand, terrestrial -- permafrost in particular -- is a more immediate concern.

As a rule of thumb, 2 Pg Carbon = 1 ppm CO2. Humans have added about 200 Pg, then, carbon to the atmosphere in the last two centuries, most of it in the last 50 years. The 1.6 trillion tons of carbon you mention (could you provide a source?) means 1.6 trillion tons * 1000 kg per ton * 1000 grams per kilogram, so 1.6*1e18 g, or 1600 Petagrams (Pg; aside: 1 Peta = 1000 Tera; 1 Terabyte computer disk is now fairly routine, 1 Petabyte disk is 1000 terabytes. Next up is exabytes, or, in our case, exagrams. The permafrost is 1.6 exagrams from this reference). This is large compared to jg's figure for total terrestrial carbon, so a source will be particularly useful. It may well be that the permafrost carbon was not considered in that older figure for terrestrial carbon.

At any rate, from that rule of thumb, the permafrost contains enough carbon to raise atmospheric CO2 by 800 ppm. That's 8 times what humans have already done, if the whole reservoir were to be emptied into the atmosphere. Not a small number.

So, Lou, your interest is tied to 2 different open questions:
Why does CO2 rise so fast at the end of ice ages?
What will happen with permafrost carbon in the next century.

Aside: An 'open question' in science is one that we don't have strong answers for. In some cases, we might not even have a good start on an answer. (At least not what scientists think of as a good start.)

Given that they're open questions, I don't have firm answers. But, since they are open, I can certainly say that there's nothing to laugh at in your writing, even if I were inclined to laugh and point at questions. (Think I'll take this up as a separate note, in fact.)

Steve Bloom said...

Bob, isn't it clear that the CO2 surges at the start of interglacials are from a reservoir in the deep Southern Ocean? I seem to recall a couple of sediment studies that found the fingerprint of such a thing. That having been established (or so I thought), I had the further impression that the mechanism for the release is down to wind shift (per Toggweiler and Russell), productivity changes or a combination of the two. Then there's the further interesting idea from Huybers of an Antarctic-region rebound-actuated surge in vulcanism to explain the double pulse of CO2.

gary thompson said...

thanks for your reply penguindreams. i perhaps was not clear in my question so let me try again. if the hypothesis that the major driver of increase global temperatures is CO2 then why hasn't the data point associated with 365ppm related to higher temps (conversely it is associated with near zero temperature increase). i get the same thing when i plot the HADCRUT3 data vs. the CO2 at mauna loa - the CO2 keeps going up but temperatures have stabilized and dropped in the past 10 years. this doesn't support the hypothesis.

Lou Grinzo said...

Thanks for the clarifications, everyone. It's sincerely appreciated.

The source for the 1.6 trillion tons of carbon thing is "Vulnerability of Permafrost Carbon to Climate Change: Implications for the Global Carbon Cycle" by Schuur et al., available here:

http://dl.dropbox.com/u/3476601/tcoe%20cc%20history/B580807.pdf

I was aware that the C in the oceans is immense compared to other deposits, but it seems that the C in permafrost and hydrates is certainly large enough tobe "interesting", plus it's in a very precarious position and could be, in effect, two or more very easily tipped over dominoes.

Honestly, I sometimes wonder if I would prefer to know a lot more about climate science or a lot less...

EliRabett said...

Eli believes that the answer to the question of why the current CO2 mixing ratio is above the line is simple:

Wait

carrot eater said...

gary thompson:

Look carefully at what is plotted. First, our host tells us that he plotted temperature with a 1000 year lag. Then, if you go to his data source, you'll also see that each temperature point is averaged over a 1000 year span.

So the temperature associated with 365 ppm in the plot is the average temperature (someplace in Antarctica, I presume?) between 0 AD and 1000 AD.

So the 365 ppm represents the year 2000 AD, but the corresponding temperature does not.

Anonymous said...

Gary Thompson,

There are two additional reasons that the current datapoint is at a lower temperature than what you would expect from the glacial period relationship:
1) Other climate forcings at work
2) Time lags in the climate response

ad 1) During the ice ages, there wer other forcings and feedbacks besides CO2: notably orbital forcing and albedo feedback. CO2 is responsible for about a third to half of the forcing (from memory; someone correct me if I'm wrong with this number). Currently there are also other forcings at work, notably other GHG and aerosols (who nearly cancel each other though). This means that the expected temp at current CO2 levels is about half to a third of what you would expect based on the historical relationship (this ties in to Hansen’s system sensitivity of 6 deg for the ice age cycles, whereas the Charney sensitivity is thought to be closer to 3 deg per doubling of CO2. Bob’s plot actually has a delta T of around 18 deg for a CO2 doubling, which would be consistent with CO2 being a third of the forcing and a long term earth system sensitivity of 6; merely thinking out loud here. Distinguishing between forcings and feedbacks in the definition of sensitivity gets confusing)

ad 2) The current climate is not in equilibrium with the current forcing, which means we expect the temperature to be lower than the forcing would eventually cause. In the absence of significant ice sheets response to the warming, we’d expect an eventual warming of around 2 degrees above pre-industrial with just the current CO2 forcing ignoring the fact that there are other forcing s at work, but as I mentioned above, they close to cancel each other). This is about twice as low as you’d expect from the ice ages relationship, because the ice sheet response was a significant positive feedback at those times, and we think/hope that they won’t be in the current epoch.

Bart

carrot eater said...

I might be reading something wrongly, but I suppose there might then be a slight issue with the phrasing here.

If the CO2 level at 2000 AD is being paired with the average temperature between 0 and 1000 AD, then the lag being used is somewhat more than 1000 years. It's tempting to call that an average lag of 1500 years, but I don't know if that's a good way to phrase it.

None of which really matters. Just nitpicking.

carrot eater said...

If you plot the data against time, it becomes quite striking how warm the previous interglacial was. I take it that the orbital forcing was much stronger back then. In some crude sense, you can say that we're trying to use extra CO2 to catch up to where the Earth was back then.

There is maybe another interesting way to show the relationships here. If you plot the CO2 level against the orbital forcing, I think you should see about same level of correlation. But the modern point will again be a lonely outlier. This would underscore that CO2 used to be a feedback to the Earth's orbit, but is now a forcing of its own.

So, er, does anybody have a data source for the orbital forcing over these time scales?

jg said...

CarrotEater, I got this link from OpenMind a few months ago:

http://www.imcce.fr/Equipes/ASD/insola/earth/earth.html

The program on this page will calculate orbital parameters and insolation values going back millions of years.
I've pulled out the eccentricity, climatic precession, and tilt values over the past million years and have been working them into an animated illustration. I'm in the process of adding the insolation values and welcome any recommendation as to which latitudes insolation is most important (e.g., is it still the 60-65 degree north zone that is thought to tip the balance between ice age and interglacial?)

jg

Robert Grumbine said...

jg, carrot eater:
You have, of course, named the source I was asking people to think about. The orbit affects both temperature and CO2 (independently). Thanks, jg, for the link to another source for orbital parameters. I have a different one, not as usable by everybody.

It is indeed still 65 N that is taken as the latitude to work with for ice ages.

Regarding the present CO2 or temperature:
The data set (link provided in the article) is not giving averaged values. The 'reference' was the one time an average is used. Deviations from the reference, though are for short term temperature (low decades) on the ice sheet, not 1000 year averages. Regarding Gary's possible question, it is indeed correct that the temperature which is being plotted against the CO2 is the temperature for 1000 years ago (from the ice core) -- because that's the best approximation to the lag.

One thing looking at the data says, now that we've looked at all of it, is that you cannot consider the current CO2 to be part of the usual relationship that used to exist. Human sources have swamped the previous connection, placing us in an entirely new regime.

But you have to look at the data before you decide this, so I kept it on the plot.

Next up will be to look at the connections between orbital variations (Milankovitch variations) and the temperature and CO2. Even with only 3 'things', there are 125 possible models that could be considered.

Robert Grumbine said...

Lou:
It can be more than a little unsettling to find out more. I have to lean on 'Forewarned is forearmed' (even when some folks seem to want no warning). Or I remind my wife of the fuller quote from Pope about 'a little learning is a dangerous thing' --

A little Learning is a dang'rous Thing;
Drink deep, or taste not the Pierian Spring:
There shallow Draughts intoxicate the Brain,
And drinking largely sobers us again.

carrot eater said...

Ah, apologies to all. I now see how I misread the description of the temperature data.

Before moving on to orbital forcing, a couple loose ends from our suggested homework:

Using the natural log of CO2: the R^2 actually gets a bit worse from 0.79 to 0.77. Not surprising, given that other forcings are doing the driving here.

I also wanted to see if this plot could be used to detect two different regimes, one for the interglacial and one for the rest of the time. We may not know exactly where the carbon comes from at the end of the ice age, but my gut tells me that it's a mechanism that is somehow less accessible under current conditions. Why? Because if it were active in the same way, we should be able to physically observe it in real time today.

So, I arbitrarily set "interglacial" to be anything above -1 C. During the interglacial, the R^2 is horrible, 0.28. The trendline slope is also weaker. I think that is consistent with my guess?

Robert Grumbine said...

carrot:
glad to see you take up some of the projects I mentioned. Compare the R^2 you had for interglacials to what I found for the pre-1959 temperature-CO2 observations. It'll look familiar.

Also have a look at the correlations for temperatures less than -1. You might be surprised here, too. Especially for what happens in the coldest, say, 100 observations (the most thoroughly glacial temperatures).

By the way -- for 800 data points, explaining 28% of the variation is usually doing quite well. You're a little spoiled with seeing the 79% that we have for the full range of data.

carrot eater said...

Actually I'm a little spoiled by what I normally do, where the R^2 values are rather higher. But there the data are different, the analysis is different, the objectives are different. That's just a reminder to not let instincts carry over from topic to topic, if there are such differences.

Horatio Algeranon said...

As Penguin points out, given the strong correlation and the fact that the current CO2 level is so far above the least squares line on the graph (5-10 std deviations is my "eyeball" guess), it is pretty clear that we are in a completely different regime.

In fact, assuming the relationship on the graph and assuming that temp change causes CO2 change would lead one to expect only about a 6ppm increase in CO2 level for the 0.75 C rise in mean global temp that we have seen over the past century.

Of course, in actuality, we have seen about 16 times that CO2 rise (100ppm).

But despite this disagreement between expectation (based on the graph) and observation (and despite the isotope evidence which shows that most of the rise was most likely due to fossil fuel burning!!), there are still some (including at least one prominent climate scientist) who propose that the increased mean global temperature in recent history might (somehow) account for most of the increase in atmospheric CO2.

That's just illogical.

carrot eater said...

Horatio, regarding people who think the current atmospheric accumulation of CO2 is mainly caused by an increase in temperature (itself caused by.. who knows what): And you didn't even give the most simple line of evidence, based on the conservation of mass.

We know how much carbon is being added to the system through combustion and cement. We also have some idea of deforestation. The accumulation in the atmospheric amount is well measured and is about half the input amount, with a good portion of the rest accumulating in the ocean (and lowering the pH). Given all that, how would these people redraw the carbon cycle by adding a input source that's larger than combustion/cement, and still both obey the conservation of mass and be anywhere near consistent with observations?

When there are so many consistent lines of evidence against an idea, you'd think people would drop it.

stephen said...

I'm a little late coming to this party, so please pardon me.

As the article suggested, I downloaded the dataset from the link given in the BBC article (800000.xls).

Using Excel CORREL() function,
I was able to calculate the correlation coefficient (R) value of .88 for correlating the CO2 concentration with the "same-year" temperature figure, and an R = .89 for correlating CO2 concentration with the "1000-yr-ago" temperature. So far those match the article.

However, the article says the standard deviation is ~11ppm. Using the Excel STDEVP() function, I'm getting 26.03 over the range D14:D813 (the CO2 concentration data column). I get the same using STDEV(), but differing in some smaller-order places. (I also make it to be 800 data points/pairs where the article says 799 points/pairs; but I don't see how that would make a difference in the standard dev. value I'm getting.)

My question is why is my standard deviation calculation different from that presented in the article?

Thanks,
Stephen

Robert Grumbine said...

Stephen:
We differ because in one place, I make an arithmetic error, and in the other, you and I are computing different things. My simple arithmetic error was that I subtracted 1 from the wrong number. We have to subtract 1 from our count of data points when we work with the lagged data. We start with 1 data point per 1000 years, for 800,000 years. My error was thinking this was 800 points. We actually start with 801 -- a data point for time 0 ('today').

The standard deviation issue is not quite as obvious, but still simple. When you computed the standard deviation, you did it for standard deviation of all CO2 observations. This is a correct computation for assuming that we know nothing about the CO2 concentrations. In mine, after verifying that the correlation between temperature and CO2 was exceedingly high, I decided that we did know something about CO2 variation -- that it correlates to temperature variation. The standard deviation I compute is for the difference between the best fit line (that knows high temperatures correlate to high CO2) and the observed CO2. Since CO2 correlates strongly with temperature, the standard deviation I get for this 'residual' (difference between expected and observed CO2) is very much lower than you get.

It's something like what happens in murder mysteries. Detective finds mud at the crime scene. It's muddy outside, so no surprise (your original large standard deviation). But the detective then looks at the mud outside, which is black, and the mud by the body, which is reddish, and there's a lot of surprise (my much smaller standard deviation -- much greater suprise to see the 365 ppm CO2). We have a clue. I'll take this up in its own post later this week. This sort of issue is very important to doing science.

Horatio Algeranon said...

carrot eater said you didn't even give the most simple line of evidence, based on the conservation of mass.

Quite right, carrot eater.

The fact that the ocean surface waters and green land plants are net sinks of CO2 effectively sinks the idea that the atmospheric increase over the past century is "natural" (eg, due to ocean out-gassing due to increased temp)

I commented on that in a post (Of Upward Slopes and Isotopes (2)) a while back (under "Key point")

But I think the illogical nature of the claim that " temperature increase over the past century was responsible for the CO2 increase" is a bit more intuitive/obvious from the above graph than it is from the carbon flux argument not least of all because one can see it with one's own eyes.

The fact that "You can't get here from there" -- ie, to current CO2 level of about 385ppm from level of 280ppm about a century ago with the small temperature change that has actually occurred -- is literally staring one in the face when one looks at that graph.

Thomas Palm said...

The temperature data you use are from EPICA, but are those really representative of global temperature so that you can put the current temperature anomaly in the same diagram the way you do?

Robert Grumbine said...

Thomas:
You raise a good point -- the temperature that is given, at all points, is the ice core's isotopic temperature. The recent isotopic temperature was taken as the zero point. (The most recent plotted point is for -0.57 C because that's a lagged temperature -- for 1000 years ago, not for 'today'.)

Going between 'global mean temperature' and 'ice core isotopic temperature' is not a trivial task, and one I've argued with ice core people about.

It doesn't, however, affect the conclusions here. There was a relationship between temperature (ice core isotopic temperature) and CO2 (ice core air bubble CO2) that held for 799,000 years. And the most recent observations are drastically far away from that relationship.

Anonymous said...

Why focusing on CO2 and AGW is silly.

- The greatest contributors of warming/cooling are ocean surface flux, heat flux, humidity and clouds. Also consider light input and planet wide albedo as well as magnetosphere strength. CO2 has a very low overall impact. Insulators (CO2) have a non-linear diminishing return. The more CO2 that is put up, the less insulating effect it will have per unit.

- Modern CO2 level is as a generous 400ppm.

- Precambrian CO2 was 4500 ppm, Oxygen @ 12.5%, temp +7C modern.
- Ordovician CO2 was 4200 ppm, Oxygen @ 12.5%, temp +2C modern level.
- Carboniferous CO2 was 800ppm (over double today), yet it has a nice Oxygen level at 32.5%. Also, temperatures were the same or slightly lower than today's temp.
- Jurassic CO2 1950ppm, 26% O2 and ~ 3C above modern. (Thats about 5x todays CO2, more oxygen, similar temps).
- Cretaceous CO2, 1700ppm, O2 @ 30%, temps about 4C higher. This is the realm of the greatest biodiversity the world has ever seen. Lots of oxygen, lots of CO2 but no man-made AGW. Imagine living in an atmosphere with this much O2! Lots of CO2 around to feed plants.
- Neogene to modern. Where we are today. Most of the atmospheric oxygen loss (about 8% of the total) occurs before industrialization.
- At no point, even with many times the current CO2, did the greenhouse effect run away.

Seems to me that people should consider planting trees that produce a lot of O2 for every CO2 processed. That would help albedo and do more for the world than any attempt to prevent CO2 changes. Build nuclear power plants (lots of them) and try to get to fusion. Then pollution (everything BUT CO2, which is plant food) can finally slow down. Things like MTBE, metals, etc. CO2 is the last thing to worry about - really.

Robert Grumbine said...

Anon:
Simply asserting things is not interesting. Asserting things that are false, or are irrelevant, is worse.

Nobody here has said that CO2 will cause a runaway greenhouse. Your attack on that is, therefore, dishonest. I've actually never seen anybody, certainly nobody in a relevant science, say that the earth could have a runaway greenhouse. If you have seen a climate scientist say so, please post the link.

You talk of 'magnetosphere strength' or the like as a concern for climate. But, again, without any scientific support.

This is a science blog, and you seem unfamiliar with what that means. One thing it means is that we're concerned with what is in the science. It also means that if you make a claim about science, it should be supported. To that end, read the link policy.

A bit lost said...

more than a comment, i had a question. I was wondering where one can obtain the raw data on climatic data, like for example what you used for your Figure. I understand that these data are published, but many of the sources I see do not provide any table of data, but just a figure. Given that someone painstakingly collected so much valuable data for over 800k years, I was wondering if the measured numbers are cataloged in some database .

Robert Grumbine said...

The numbers I used were from the spreadsheet at the BBC article I linked to -- http://news.bbc.co.uk/2/hi/in_depth/sci_tech/2009/copenhagen/8393855.stm

These are not the most raw data you could get, however. They've already been analyzed to give you information at a regular spacing -- every 1000 years. To get closer to the raw data on ice cores, you want to visit the National Geophysical Data Center -- Ice Core, for instance.

A bit lost said...

i just wanted to play with the data myself .. many many thanks .. your reference was very helpful .. !!

Robert Grumbine said...

Absolutely. I encourage that. That's the reason for the 'project folder' tag -- all of those posts have ideas and/or data for you to try working with.

Tedsaid said...

Hi, Robert. I hope your wrist is healing nicely.

I previously posted this comment at Watts Up With That. You asked me to re-post it here, so here it is! (I changed a few words to try and improve clarity.)

Hi, Robert. A few comments on your post.

You state that your analysis of the ice-core record disproves two conjectures: 1) that the current CO2 rise is due to past temperature increases; and 2) that CO2 in the atmosphere can’t be affecting current temperatures. Your analysis is predicated on the fact that the modern rise of CO2 does not have an equivalent in the ice-core data and is much higher than anything seen in that record.

There are other reasons to believe the current rise in CO2 is due to human influence, or that CO2 affects temperature to some (disputed) degree. However, CO2 levels from the ice-core data CANNOT be used in comparison with the CO2 levels in the modern era (i.e., the last half of the 20th century up until now). The reason for this is the “yardstick” for measuring temperature and CO2 levels from the ice-core data is VERY crude - much less precise than the annual Mauna Loa information.

I’m assuming you used data from the Vostok ice core? As I recall, the precision for this data is 700 yrs, +/- 200. This is due to the very low rate of precipitation, and the long amount of time before the CO2 becomes “trapped” in the ice.

So essentially, your analysis is using “smoothed” data, smoothed over 700 years.

Trying to measure an annual effect using ice core data is kind of like trying to measure a person’s height using your car’s odometer. In other words, it is entirely possible that CO2 levels have had a much larger range in the past, and are simply smoothed out in the ice-core data. We would need a few hundred years' worth of data (at least!) from Mauna Loa, before we could say anything substantial in the comparison. (And, of course, that data would have to be normalized against future ice core data.)

This mistake of comparing the crude precision of ice-core data with the high precision of modern measurements is made all the time, especially in the press. Al Gore’s presentation is a particularly well-known example. I'm surprised the issue doesn't come up more often.

But if anyone else has some more information on this, I’d be glad to read it … I’m not an expert in ice-core data.

Cheers.

Tedsaid said...

Hi, Robert. I hope your wrist is healing nicely.

I previously posted this comment at Watts Up With That. You asked me to re-post it here, so here it is! (I changed a few words to try and improve clarity.)

Hi, Robert. A few comments on your post.

You state that your analysis of the ice-core record disproves two conjectures: 1) that the current CO2 rise is due to past temperature increases; and 2) that CO2 in the atmosphere can’t be affecting current temperatures. Your analysis is predicated on the fact that the modern rise of CO2 does not have an equivalent in the ice-core data and is much higher than anything seen in that record.

There are other reasons to believe the current rise in CO2 is due to human influence, or that CO2 affects temperature to some (disputed) degree. However, CO2 levels from the ice-core data CANNOT be used in comparison with the CO2 levels in the modern era (i.e., the last half of the 20th century up until now). The reason for this is the “yardstick” for measuring temperature and CO2 levels from the ice-core data is VERY crude - much less precise than the annual Mauna Loa information.

I’m assuming you used data from the Vostok ice core? As I recall, the precision for this data is 700 yrs, +/- 200. This is due to the very low rate of precipitation, and the long amount of time before the CO2 becomes “trapped” in the ice.

So essentially, your analysis is using “smoothed” data, smoothed over 700 years.

Trying to measure an annual effect using ice core data is kind of like trying to measure a person’s height using your car’s odometer. In other words, it is entirely possible that CO2 levels have had a much larger range in the past, and are simply smoothed out in the ice-core data. We would need a few hundred years' worth of data (at least!) from Mauna Loa, before we could say anything substantial in the comparison. (And, of course, that data would have to be normalized against future ice core data.)

This mistake of comparing the crude precision of ice-core data with the high precision of modern measurements is made all the time, especially in the press. Al Gore’s presentation is a particularly well-known example. I'm surprised the issue doesn't come up more often.

But if anyone else has some more information on this, I’d be glad to read it … I’m not an expert in ice-core data.

Cheers.