"Our greatest responsibility is to be good ancestors."

-Jonas Salk

Friday, March 20, 2009

Climate Science Open Thread

An experiment:

All points of view on all climate-relevant topics welcome in comments here. No ad hominem attacks, foul language, or innocent people's credit card numbers, please. Otherwise have at it.


Michael Tobis said...

Raven writes:


Thank your for you answer and I want emphasize that I do *not* want to get into a debate about the merits of the hockey stick studies here. I brought them up because because they had particular relevance to my own path to AGW skepticism. I posted on this thread because there appeared to some discussion on why people are skeptical (I can move it to your open thread if that is where you want it).

I also agree that the issues with the hockey stick studies do not refute the larger body of climate science.

That said, they are significant to me because it makes me question whether I can trust the climate science community and that is why the reaction of the climate science community to the criticisms of the hockey stick is more significant than the criticisms themselves.

You said yourself that you do not know enough about the hockey sticks to defend them but you accept them as accurate because you trust the authors. This suggests that you recognize the importance of trust when it comes accepting scientific claims made by others so you should understand how come I lost trust in the scientists who are willing to defend what I see as deeply flawed science.

This loss of trust means it need to see concrete and *unambiguous* empirical data that demonstrates the correctness of the AGW claims. Hand waving arguments about how the recent cooling trend is too short to refute the models only work with people that already accept that the models are likely to correct. I cannot accept that the models are likely to be correct because I have reason to distrust the people that came up with the models because of the obfuscations over the hockey stick.

In other words, reaching out to people like me who know enough to recognize that scientific mistakes have been made will require that people start by admitting that mistakes were made. Unfortunately, the political climate seems to made such admissions unlikely which leaves us at an impasse.

Anonymous said...

Earlier you said that you trust Steve McI. What is his stance, about, say, climate sensitivity to doubling CO2?

Is the stance that all the studies are bogus and that we don't know; that it's lower; that IPCC figures are probably the best guess; or something else?

I do understand that not everyone can research everything themselves and hence the "no arguments from authority" is often a vacuous argument. People have finite time and effort.

David B. Benson said...

Here is a quotation which ought to do for those "middle" folk.

Jim Galasyn:
Fundamentally, climate science is based on well-understood principles of thermodynamics. Before humans burned the sequestered carbon (fossil fuels) and released CO2, Earth was in radiative near-equilibrium with space. Humans introduced a sudden, 500-gigaton excursion in the global carbon budget. Because CO2 is a “heat-trapping gas,” Earth is now in disequilibrium with space. To return to equilibrium, the atmosphere must warm.
The rest is details. Interesting details, to be sure, but the basic thermodynamics have been understood since Svante Arrhenius published in 1896.

Anonymous said...


SteveMc does not have a opinion on CO2 sensitivity other than a general complaint that there is no rigorous engineering quality derivation of the quantity.

This simply confirms what I have found from other sources that IPCC estimates of CO2 sensitivity are model outputs and depend heavily on the reliability of the models (the estimates from the paleo-data are not independent of the models since these estimates appear to rely on many of the same assumptions that are built into the models).

This means that the entire AGW argument rests on the reliability of the models. To make matters worse, there does not appear to be any objective mechanism that can be used to evaluate the performance of the models.

I realize that the modellers have published analyses that claim the models are accurate but I have seen other credible analyses that show the the opposite. The fact that some climate modellers have chosen to obfuscate on the hockey stick means I am not going put much weight on any analysis of models by the modellers or by people closely associated with them.

This means I have no choice but to conclude the models do not provide any useful information that can help us make any policy decisions.

What we need is a way to independently evaluate the models much like we depend on accountants to independently evaluate the finances of a corporation.

Anonymous said...


I do not dispute the heat enhancing effects of CO2. The important questions are how much warming we can expect and what are the consequences of that warming.

The IPCC relies on climate models to answer those questions and, as I mentioned above, I don't have any confidence that these models provide useful information at this time.

This is directly related to the fact that I do not trust the people who are responsible for the models.

It is worth keeping in mind that saying that the models provide no useful information does not mean that the CO2 issue should be ignored. It just means discussion of possible policy options will be quite different if one includes the possibility that the models are completely wrong and CO2 sensitivity is in the range suggested by Spencer or Lindzen.

Anonymous said...

I haven't followed the "hockey stick saga" that closely (maybe someone else who knows about that really can chime in, I have heard claims for example that the Wegman thing was erroneous), but it seems to me the recent other reconstructions mostly agree with those ten year old ones.


Of course truth matters. Is your claim that Mann et al cheated and it's just by chance that the other reconstructions are mostly the same? Or do they have common errors? Or do they differ decisively?

It's interesting that Steve McIntyre's posturing around that single issue has caused so many to disregard the whole of climate science.

It's probably not even the same people who do the reconstructions and who do the climate models.

I guess if you are concerned about the fraudulency or errors of the modelers, you should aim at open source climate models.

I actually do think that is a good idea (and open sourcing generally with important science codes - more and more of it is getting closed source). Maybe you should just ensure enough compensation for the people who "divulge their code".

Isn't there some "open source" project for a temperature measurement network, weeding out bad stations?

That kind of work. You could make new community built models.

Then if the results seem to somewhat agree with the earlier papers with closed source models, the older versions are "validated" in that sense. Of course there are quite many sources for differing results, so one can not immediately cry fraud if the results are different.
AFAIK there are currently some Japanese climate models that differ from the others quite a lot.

In the mean time, there are some simpler models that can give some rough estimates, models that people can create in one evening.

David B. Benson said...

Raven --- The warming from just CO2 can be determined to be 1.2 K for 2xCO2. The remainder is then determined in a variety of ways, not just from GCMs. Annan & Hargreaves have a very good paper combining the various estimates; there is a thread about it on RealClimate and also on James' Empty Blog.

The short answer is that the most likely value is close to 3 K, but there remains a 5% chance that it is less than 2 K.

Want to bet the future on those odds?

David B. Benson said...

Here is one of the shorter, simpler approaches.

Barton Paul Levenson:
1. CO2 is a greenhouse gas (Tyndall 1859).
2. CO2 is rising (Keeling et al. 1958).
3. The new CO2 is mainly from burning fossil fuels (Suess 1955).
4. Temperature is rising (NASA GISS, Hadley CRU, UAH, RSS, etc.).
5. The increase in temperature correlates with the increase in CO2 (60--76% for temp. anomaly and ln CO2 for 1880-2007). See

Anonymous said...


RE: Sensitivity-

Agree or disagree with this statement (and why/why not)?

There exist large, naturally occurring, negative feedbacks in the climate system sufficient to prevent temperature change on the magnitude of ~5-6C.

Anonymous said...

Er- on the order of/of a magnitude of

Anonymous said...


I don't want to get into a back and forth on the merits of the hockey stick. If you want to learn more about why I think the way I do then I suggest to start with this:


The argument that "other" reconstructions produce the same results is not credible to me because all of the them rely on the same problematic datasets.

I also have been careful to say nothing about motivation because motivation does not affect my conclusion that I cannot trust the scientific judgment of the people who did the studies or the people who insist that the studies are correct.

You are also correct to say the people doing the modelling are not the same as the people doing these paleo studies but their credibility has been damaged in my view because the climate science community/IPCC refused to address the legimate concerns with the proxy studies when they were made public. That refusal means that I cannot trust the IPCC to find problems with the science it endorses on its own.

As I said before, it all down to trust and once trust is lost it is tough to get it back.

That said, I do believe that in science data trumps everything and if the world was warming at a rate predicted by the models then I would not be a AGW skeptic despite by misgivings about the paleo studies.

Michael Tobis said...

A couple quotes from an article on RealClimate (one of some considerable interest to me):


Rod B Says (18 March 2009 at 12:39 AM):

Well, you’re getting warmer As stated, the temperature data has nothing to do with the formation of the theory. It’s one of many lines of evidence which can be used to show that predictions from the theory are correct.

So we have the theory, and we have many lines of evidence - glacier retreat, arctic ice melting, earlier bloom times, poleward range extensions, and more - that match the predictions from theory. The temperature data matches the predictions as well, but some people want to “audit” the data, adjust it so that they can claim it doesn’t show a warming trend, and then claim that the adjusted data invalidates the theory.

I admit to not having more than basic college statistics, so I can’t offer an informed opinion on those audits. However, I can do a bit of trimming with Occam’s Razor. Start with N lines of evidence that match predictions from theory. After the auditors do their work, there are N-1. Even presuming the auditors have no particular axe to grind, what’s more likely: that the N-1 are wrong, or that the auditors are?

Of course if any one of the many lines of evidence differers markedly from others, it needs to be looked at. Professional statisticians here & elsewhere have done just that, and have pointed out (as far as I understand them) exactly where and why
the auditors’ methods are in error.


James wrote in 274:

I admit to not having more than basic college statistics, so I can’t offer an informed opinion on those audits. However, I can do a bit of trimming with Occam’s Razor. Start with N lines of evidence that match predictions from theory. After the auditors do their work, there are N-1. Even presuming the auditors have no particular axe to grind, what’s more likely: that the N-1 are wrong, or that the auditors are?

I suppose you think that the likelihood that a given scientific conclusion is wrong decreases as an exponential function of the number of lines of evidence.

That only works for members of the reality-based community. For denialists its the other way around. Knock out any one line of evidence and you’ve knocked out the conclusion — at least until someone else brings up the other lines of evidence. But then you can ignore them, go home, come back tomorrow and start afresh.

Anonymous said...


None of the data (GMST, OHC, Troposphere Temps) I have looked at gives me any reason to believe that CO2 sensitivity is much more that 1 degC/doubling. That said, even if the data allows me to reject higher sensitivities with 95% certainty there is still a 5% chance that the sensitivity is somewhere close to what the IPCC says.

Given that uncertainty I think we should hedge our bets in terms of policies by investing heavily in R&D without committing to a radical restructuring of the global economy. My feeling is this issue will be resolved one way or another in the next 5 years (i.e. if the temps do not start rising rapidly soon then we can safely assume the models are wrong).

Your correlation argument does not mean much because other things correlate better with temperatures(e.g. PDO). The most plausible alternate explaination for the observed rise is internal variability in of a chaotic system which manifests itself via ENSO events. I realize that the climate modellers try to use the hockey stick studies to claim that the climate does not have that much internal variability, however, I do not accept that explaination for reasons I have already stated.

Anonymous said...


I don't know what the real CO2 sensitivity is and I am not convinced anyone else does either.

All I know is I have seen credible scientific arguments for CO2 sensitivity from 0.5 degC to 3 degC and the data seems to suggest it is around 1-1.5.

That said, things could change as we get more data but we also run into that trust issue whenever I hear about existing datasets being 'adjusted' to better match the models. How can I know whether these adjustments are reasonable when I already have reason to suspect that some scientists have a big problem with confirmation bias?

Anonymous said...


Things like glacier retreat, arctic ice melting, earlier bloom times, poleward range extensions only provide evidence for warming. They provide no evidence for the cause of the warming.

As far as I can tell, attribution studies with climate models are the only line of evidence that specifically shows that CO2 is cause of the warming yet those same attribution studies presume that the climate would be completely stable unless there was some external forcing.

I have yet to see any evidence other than the hockey stick studies which show that the assumption of a stable climate is valid. OTOH, I have see other studies which show that the assumption of stable climate is false.

IOW - there are not N lines of evidence. There is only 1 line and that line depends on assumption which is not necessarily valid.

Anonymous said...

We can go back an forth on the different arguments but at the end of the day comes down to a question of who do you trust. I realize that everyone here uses the 'safety in numbers' argument (i.e. when in doubt trust the majority opinion).

Unfortunately, for people like myself, the hockey stick episode means that I am not willing to automatically trust those who align themselves with the majority IPCC position even if I do not absolutely reject the claims.

Yet when I express my lack of trust I get accused of denying the scientific evidence. This is a ridiculous accusation from my perspective because I am very interested in finding out what the scientific evidence really says but I am forced to rely on intermediaries who tell me what that evidence says. Why should I accept the interpretation of evidence from intermediaries who I do not trust? (this is not a rhetorical question).

I have already mentioned that real data which unambiguously supports the model claims would restore my trust. The depoliticization of the discussion would also help but that is never going to happen.

guthrie said...

Just once, I'll join in:

Raven, the cause of the warming is well understood- as previously commented CO2 is a greenhouse gas, as you admitted upthread.

Attribution studies show cooling stratosphere, patterns of warming which roughly match those observed, etc.
Add them together with the fact that CO2 (and of course other gases) causes warming, and what more evidence do you need?

I've never seen anyone claim that the climate is or should be stable, except those commonly labelled deniers, therefore your straw man is irrelevant.
Whilst 30 years ago people might have thought that the climate was fairly stable, the intervening research has shed light on how unstable the climate can be, what with El Nino/ La nina, the shifts in rainfall in AFrica, ocean currents etc. And that is before we add in changes in solar output, volcanos and human interference.

Ironically I recall seeing many people, sceptics and deniers, using that graph from the early IPCC report that shows a very smooth last 1000 years of global temperatures, with a huge but quite flat peak around 1000 to 1400 AD, the "medieval warm period".
Fast forwards 10 years, with many studies showing that different bits of the world were warm at the same time as others were cold, and the science has advanced, showing just how variable local weather and climate can be, and people accuse climatologists of assuming the climate is stable, when they are the very ones collecting the data showing it isn't.

Anonymous said...


I have looked at the stratospheric cooling claim and have few issues:

1) Everything comes down to the effect of water vapour and clouds in the troposphere. The temperature profile in the dry stratosphere could be exactly as predicted by the models yet the models could still be completely wrong on CO2 sensitivity.

2) The stratospheric data is not unambiguous. At lower elevations there is no continuous trend that tracks CO2 levels. Instead we see series of steps triggered by volcanic eruptions followed by a period of stable temps as CO2 rises. This could mean nothing or it could be a sign that models don't have the mechanism correct even if the long term trends match.

3) The temperature profiles in the troposphere do not match the model predictions well at all. It is not reasonable to point to strospheric temperatures as evidence while ignoring the problems with the troposphere.

I am also aware that there are different opinions on how well the models actually match the troposphere. Some opinions claim that the range is models is large enough that any possible tropospheric trend is 'consistent' with the models. Others using the same statistical techniques but with data up until 2008 instead of 1998 show that the models are not consistent with the data even with the large range.

Who should I trust when it comes to these kinds of statistical analyses? Scientists that tell people to FOAD when asked for the data supporting their work? Or people who show every step of their work online and appear to be open to criticism?

It all comes down to trust. When I see arrogant prima-donna like behavoir as documented here: http://www.climateaudit.org/?p=4314

The first thing that comes to mind is I cannot trust the person that wrote that email response to SteveMc. Can anyone give me a reason why I should trust that scientist?

Anonymous said...


If you look at the attribution studies in AR4 you will see a plot of temperatures with human forcings. This plot suggests that the climate would have cooled over the period from 1970 too 2000.

Here is a link to the figure:

The model runs used in this analysis have a very narrow range which indicates that the models do *not* allow for climate internal variability over decadal timescales.

This is what I mean when I say that the models do not take into account the internal variability of the climate.

Anonymous said...


I don't think that my question is terribly complicated. Is there a reason why you are refusing to answer? Regardless of your opinion/understanding of sensitivity in terms of doubling preindustrial CO2- agree or disagree, and why:

There exist large, naturally occurring, negative feedbacks in the climate system sufficient to prevent temperature change on the order of ~5-6C.

Michael Tobis said...

I agree with Thing's thrust here. Raven is all worked up around trying to argue that the sensitivity *might* be small, but that is not especially policy-relevant.

Costs are dominated by the other tail, amd noting McIntyre says or Raven says as far as I can tell are reassuring about how *bad* it might plausibly get.

It is as if Raven is saying don't put lifeboats on the Titanic, because it probably won't sink. Don't buy fire insurance, since your house might not burn down. Quit your job, you might win the lottery.

The weaker the science, the less we can constrain the dangerous side. The less we can constrain the dangerous side, the more vigorously we should avoid it.

Raven is a typical case of the person getting it backwards. The more seriously I take Raven's arguments about the science being untrustworthy, the more scared about the future I get.

Anonymous said...

"There exist large, naturally occurring, negative feedbacks in the climate system sufficient to prevent temperature change on the magnitude of ~5-6C."
Utterly meaningless question.

Anonymous said...

The best template for the Earth's response to external forcings and possible socio-economic-ecological effects is in the Earth's geologic record. Researchers looking at paleoclimate events such as the PETM, Cretaceous and Eocene hothouses, Last Glacial Maximum, the last millennia, response to Pinatubo, response to the solar or seasonal cycles, etc as well as a wide variety of modeling studies have pinned down sensitivity to be between 2 and 4.5 C per doubling. It is almost impossible to reconcile this paleoclimate record with a very low or very high sensitivity.

The argument for a low sensitivity is the latest tool in the denial toolbox because it's easier to attack than the basic fact that it has been warming and that CO2 absorbs infrared radiation. At least we've seen "skeptics" getting smarter. What I mean is that there is no simple quantitative derivation of climate sensitivity, and there are actually real question marks (unlike the outdated questions like "is AGW happening??").

However, the data is there to constrain sensitivity, but if we want to argue about "tail ends" then Roe and Baker among others have shown that the intrinsic nature of feedbacks make it more difficult to rule out a very high sensitivity than a very low one.

I'm not sure I agree that policy makers have to worry about a >6 C sensitivity, but there's also more to this topic than the "sensitivity" which is more or less a number. It says little about ecological effects, precipitation changes, possibility of "abrupt climate change," shifting agriculture zones, and other things. I would not mind seeing the wider community move beyond these academic exercises which may not be practical like "what is the best number for Earth's sensitivity."

The world also does not end at a doubling of CO2, or "by 2100" so regardless of the true sensitivity, bad things are going to show up eventually.

Politicians must rely on the mainstream view for scientific-based action, and the mainstream view says that if we continue to systematically increase CO2 concentrations in the atmosphere then we will witness consequences.

Anonymous said...


You have to remember that there is a huge cost to being wrong no matter what we do. This is not a situation where we can reasonably say "better safe than sorry" because jacking up energy prices to levels required to force a switch to non-CO2 emitting sources will result increased human misery. They only way to justify causing such deliberate harm to people is to be absolutely certain that the alternative will be worse.

If you prefer analogies we are in the situation where a doctor (who has been caught giving bad advice in the past) says that a lump in our leg could be cancer but he can't be sure but is still recommending amputation. Most people would rather wait for more certainty in the diagnosis before committing to amputation even if it increases the risk to the life.

It also important to remember that different people have different risk tolerances. Some people have no fear of jumping out of planes for enjoyment, others would never consider it. These same differences in risk tolerances affect the policy choices that people are willing to embrace even if they agree on the risk probabilities. There is no *denial* of scientific evidence in this case. This is nothing more than risk assessment in the face of uncertain data - something we have to do all of the time and it is unreasonable to expect everyone to agree with you on the best course of action given the same set of data.

Anonymous said...


We have next to no data about the earth's past climate. You like to focus on the PETM because it re-enforces the high CO2 sensitivity arguement. But when I look at the complete history I see little evidence that CO2 is a major driver of climate:


Now I realize that the past CO2/temperature data is very uncertain but that cuts both ways.

If the uncertainty means that I can't use data from jurrasic period to show that CO2 cannot possibly have high sensitivity then you cannot use data from the PETM to claim that it does.

From my perpective the only useful data on sensitivity comes from the last 50 years but that is confounded by so many other factors because we cannot seperate the effect of CO2 from things like land use change or natural internal variability.

I personally think that planning for extreme climate change scenarios is like planning like planning for a major asteroid strike. i.e. everyone can agree that it is a possible outcome but the cost of dealing with it is too high to justify any effort given the unknowns. We would be better off setting much more modest goals that are politically achievable.

Michael Tobis said...

Not a bad answer, that. Better than I expected.

At that level, it becomes a matter of what is at risk.

The leg will have to come off anyway (eventually we will run out of fossil fuels, and by most accounts not too far into the future) so if it enhances survivability so much by doing it a bit early, let's go for it. It's not as if fossil fuels were a sustainable pattern anyway.

Anonymous said...


I agree that fossil fuel use is not a viable long term strategy but I object to setting policy based on the secondary benefits. If the secondary benefits are what is important then they should be primary objective.

For example, we can agree that oil is the fossil fuel with the biggest supply risk in the short term. Therefore, a policy that encourages a switch to electric transport by ensuring a cheap reliable source of electricity while increasing taxes on gasoline would make sense.

However, a policy that simulatenously jacks up the cost of electricity by outlawing coal would be less able to achieve the objective of reducing oil consumption because it cost differential between oil and electric power would not go down.

As with all things in life it is best to succeed with a narrow focus on what is achievable than to fail by trying to do everything at once.

Michael Tobis said...

"As with all things in life it is best to succeed with a narrow focus on what is achievable than to fail by trying to do everything at once."

In normal times I would agree.

This would make sense had we not spent the last 28 years (!) with US administrations that had as their primary impulse deferring long range difficulties to the wisdom of the short-view processes of Wall Street. (Mixed in with more than a little bit of geopolitically ambitious ignorance in the last bunch, which iced the cake...)

So many long term problems have built up as a result that it no longer makes any sense to think of them as separate. We have One Big Mess of stuff that the marketplace can't handle, and we had better start rolling up our sleeves.

Interestingly, the conventional maximum planning horizon in business is 30 years. Coincidence?

Anonymous said...


In the last 40 years we have seen multi-national anti-acid rain and CFC policies successfully enacted by Republican and Democratic administrations. So I think it is wrong to say that the political/economic system cannot act on long term environmental risks.

What is missing in the case of CO2 are cost effective alternatives. This is the real reason for inaction and it will be the main reason for inaction as we move forward.

I realize that starry-eyed types like Romm have put together "plans" that sound achievable but most pragmatic people recognize that the problem cannot be solved with the technology we have now and no amount of government funding/market manipulation can ensure that cost effective technologies will appear within the timeframes mandated.

This recent newsweek artical illustrates how intractable the problem is:

"To get 10 terawatts, less than half of what we'll need in 2050, Lewis calculates, we'd have to build 10,000 reactors, or one every other day starting now. Do you like wind? If you use every single breeze that blows on land, you'll get 10 or 15 terawatts. Since it's impossible to capture all the wind, a more realistic number is 3 terawatts, or 1 million state-of-the art turbines, and even that requires storing the energy—something we don't know how to do—for when the wind doesn't blow. Solar? To get 10 terawatts by 2050, Lewis calculates, we'd need to cover 1 million roofs with panels every day from now until then. "It would take an army," he says. Obama promised green jobs, but still."

For me this is the key element. I am not willing to support policies that will most likely fail in their stated objective even if I agreed that the policies make sense given scientific evidence. I would rather see the money invested in something that will produce guaranteed returns such as adaptation even if it costs more in the long run.

David B. Benson said...

We have quite good measurements of trace gases and temperatures from Vostok, extended now by Dome C. For this length of deep time, and far longer, we have the LR04 stack. So there is plenty of data to uyse to check that, indeed, sensitivity is most likely 3 K.

Sorry Raven, but it does seem to me your are being deliberately obscuranist.

Anonymous said...


Vostok is one point on the globe. If we only needed one point to completely understand what the climate is doing then why are we wasting time with a network of thousands of ground stations, tidal guages and weather ballons?

Wouldn't be cheaper to set up a single station in Vostok that would tell us everything we need to know?

I think you need to think more carefully about how little data we have from the past.

BTW the PETM was 55 million years ago which is way older than any ice record from Volstok.

David B. Benson said...

Raven --- I said also the LR04 stack. I'll admit that the more recent, the more data. From the Eemian interglacial on there is plenty to use to aid in determining a value of climate sensitivity.

To correct a misconception you have

"For a gray body earth without any feedbacks, this temperature rise is ΔT×2 = 1.2°C." from

Using the oceans as a calorimeter to quantify the solar radiative forcing

To correct another, note that the PDO is an oscillation without secular trend. Global surface temperatures since (at least) 1850 CE demonstarate a clear secular trend: up. This paper

On the Pacific Decadal Oscillation and the Atlantic Multidecadal Oscillation: Might they be related?

I found particularly good.

Anonymous said...


Small changes in clouds have a huge effect on the earth's climate yet we don't have any proxy data that would tell us what was happening to the clouds. If we don't know what was happening to cloud cover we cannot know what the CO2 sensitivity is.

You are also assuming that CO2 sensitivity is a constant. Since we have no mathematical derivation for the number we have no way to know if the sensitivity when the earth is coming out of an ice age is going to be the same as the sensitivity today.

As far as the PDO goes: if 70 year ocean cycles exist then where is the evidence that 1000 year ocean cycles do not exist? How can we know we are not in the middle of century long super El Nino like the one that probably occurred during the MWP or the Roman Warm Period?

These are unanswerable questions. We can guess. We can assume. But we cannot know. For that reason we cannot pretend that we *know* that the majority of warming over the last 30 years was due to CO2. That is simply hypothesis produced by models that *assume* that long term ocean cycles do not exist.

Anonymous said...

Raven wrote:

"We have next to no data about the earth's past climate. You like to focus on the PETM because it re-enforces the high CO2 sensitivity arguement. But when I look at the complete history I see little evidence that CO2 is a major driver of climate:


"Now I realize that the past CO2/temperature data is very uncertain but that cuts both ways."

That's not a credible source, Raven. You need something both up-to-date and direct (i.e. not subject to fiddling by someone like Nasif Nahle). Also, bear in mind that it's not at all the case that CO2 is the only factor driving global temps over such time scales, so even if correct that graph by itself doesn't support your argument.

Here's the most recent paper that I know of on this broad subject. More recent specific work has added detail, but is basically consistent.

Noting that the more recent source for Nahle's graph is Mark Pagani, you might want to read this related recent paper on which he was a co-author.

You might also want to look at the recent ANDRILL results, in particular the paper that came out a few days ago (press release).

Anonymous said...

Just to note the things are even worse with Raven's linked graph than it first seemed:

It turns out the graph has nothing to do with the 2005 Pagani paper, but appears to be taken entirely from Scotese 2002 (noting that I haven't checked to see whether Nahle changed it in any way).

In the deep paleo field, 2002 is pretty much ancient history, but more interestingly the Scotese graph (if that's what it really is) is actually contradicted by Pagani. First of all Nahle gets the title of the paper wrong (it covers the Oligocene, not the Paleocene as Nahle has it), but on the substance the alleged Scotose graph shows a sharp CO2 drop during the Eocene and steady levels through the Oligocene, whereas Pagani finds that the drops happened during the Oligocene.

This stuff is complicated, but it's not that complicated.

Anonymous said...


Why should I assume that Nahle is a less credible source than Royer or Pagani? Because you say so?

I looked at the Royer paper and it does not provide any comparison between independent temperature records and CO2 records. It simply assumes that CO2 is a major climate driver and tried to put bounds on the sensitivity based on that assumption.

I find the wording interesting:

"Phanerozoic records *generally* show a positive coupling between CO2 and temperature but determining DT(23) quantitatively has proved difficult because there are no convenient proxies for global-mean surface temperature."

In other words, the records do not *always* show a coupling between CO2 and temperature which is the same observation I made from Nalhe's graph.

Here is another reference to Scotese that shows the same anti-correlation in the Jurrasic: http://www.geocraft.com/WVFossils/Carboniferous_climate.html

Pagani may have a different opinion but since none of this can be proved one or another so it is not clear to me why I should trust Pagani's opinion. Especially when I know that biases of the scientist strongly affect interpretation of the proxy records.

Note that I am not saying that I trust Nahle and Scotese more than Pagani. I am just saying I have no reason to trust any one of them more than the other.

David B. Benson said...

Raven --- The ocean oscillations have to do with Rossby and Kelvin waves. While these have modes, a roughly 1000 year mode would be too small to have any noticable affect.

Read Annan & Hargreaves to see that sensitivity has not changed appreciably since at least LGM. Since it is mostly water vapor feedback, one can calculate that part quite readily.

Could cover data comes these days from satellites. For older data people have gone to the trouble of using weather station reports of sunrise and sunset times to estimate cloud cover. That abstract indicating a slight positive correlation between cloudiness and the solar cycle.

But that is such a small effect that it doesn't much matter.

With regared to RWP, MWP, LIA and so on, do read W.F. Ruddiman's popular "Plows, Plagues and Petroleum" to discover that humans have had an influence on climate for much longer than many are aware.

David B. Benson said...

Raven --- Modern climate began with the closure of the Isthmus of Panama, about 4 +- 1 million years ago:


Since that is about when lines leading to humans diverged from lines leading to the cetral chimpanzees, it might be best to restrict attention to this period.

Especially as we have better data for it.

David B. Benson said...

Here is another person's take on this.


The science that remains settled is:
1. The role of atmospheric CO2 as a greenhouse gas.
2. Thermodynamics
3. The fact of anthropogenic emissions of CO2 as a primary forcing function of increasing global warmth.
4. The negligible impact (on global climate change) of solar cycles, orbital changes, cosmic rays, etc.

The science being aggressively altered includes:
1. The roles played by various ocean circulations and oscillations.
2. The effect of the interactions among the complex (in its technical sense) feedback mechanisms.
3. The timing and intensity of changes implied by anthropogenic atmospheric CO2.

I never claimed that ALL the science was settled -- if that were true, no research would be needed. My point (which, by the way, Dr. Lindzen often agrees on) is that there is more than enough new and unsettled science at the boundaries of our understanding -- so much so that we don't have time to waste in fruitless rehashing of the aspects that are and remain settled.

Anonymous said...

Ah, the old "prove to me in terms that I will decide to agree with that a high school biology teacher isn't just as qualified on the science as some of the recognized leaders in the field, 'especially when I know that biases of the scientist strongly affect interpretation of the proxy records'" gambit.

Then you misread that paper and link to an even more crap graph, this one by West Virginia's chief mine safety engineer (of all people).

Gosh, that leaves me with... absolutely nothing polite to say.

Michael Tobis said...

I think Raven might be a pro. I am flattered.

tidal said...

Michael Tobis said...
I think Raven might be a pro. I am flattered.

I was wondering same... peut-etre a "Carleton" raven?

Else, rather sad...

Anonymous said...


The Rossby Wave-PDO link is new information for me. Any papers not behind a pay wall on it?


Whether you want to admit it or not there are a lot of uncertainies. I know these things can't be proved and I don't really expected them to be but I take exception to the >90% certainty claimed in the IPCC reports. I do not feel it is justified given the number of unknowns.

You also have taken exception to the links I provided because you do not consider the sources to be authorative. Which is fine. I don't expect to convince anyone here. However, you have missed by original point about the problems I have with the "authorities" who defended the hockey stick papers.

The failure of the "authorities" to acknowledge the problems makes me very suspicious of claims which cannot be independently verified. This means that almost any paleo study suspect (including those that may support a skeptical view).

That is why your reference to Pagani studies which happened to reinterpret proxies in ways that conveniently supported the alarmist paradigm set off all sorts of warning bells.

That is why I have been extremely dismissive of evidence from the paleo record and I am more interested in how the models succeed at predicting the future climate. I do not feel the current crop of models has done a good job but I know they are being improved and perhaps the next batch will do a better job.

Whatever happens I will re-evalute my position as I get new information.

Anonymous said...

In the thread that gave birth to this one, M. T. said "disagreement is an opportunity for learning, not an opportunity for viciousness and mutual contempt."

And what do I see here? Paranoid assumptions on both sides.
I don't want to play the false equivalency game though: tarring all scientists and all the data they collected with a broad brush because a scientist wrote a taunting email and another might have misused stats tells me Raven's got an axe to grind.


I still don't get what's your issue with "AGW". You think sensitivity is a valid concept and you apparently think its value is likely not negligible. Have you figured there are fast CO2 sinks somewhere? Because this spells trouble otherwise... large random variability may mask the signal for a while but you're still going to get hotter peaks and hotter troughs.
That said, I think you make some good points.
You seem to have invested much time in this topic and you seem to have a handle on the physical principles so I keep thinking that you could do better than blindly trust or distrust climate scientists as far as the big picture is concerned. But far from me to distract you from your passion for the finer points...

Anonymous said...


You right in the sense that I am angry at the people who have let politics interfer with science by being afraid to admit that scientists do make mistakes. I am also angry at scientists who think that they have a monopoly on the discussion of the science and that outsiders should be denigrated and marginalized. I am angry at people who use terms like "denialist" or "flat earther" to describe people who taken the time to learn about the science yet come to different conclusions on the best policy choices.

That said, there are lots of skeptic echo chambers for me to hang out in if grinding an ax was my objective. I post here because facing people I disagree with is the only way for me to learn.

I appreciate the fact that Michael is letting posts show up without advance moderation. I hate spending time on a post that never sees the light of day because I said something that the forum owner did not like.

As for the trust issue that I keep harping on: I am trying to make the point that the climate scientists out there need to understand that trust must be earned and can be lost. It is not enough to point to a stack of peer reviewed papers and demand unquestioning trust from people who are not intimidated by advanced degrees because they have their own.

Anonymous said...

I think trust can be earned and lost as well. But *must* trust be earned? I trust people by default. Sure, I only trust them so far by default... but tnough to assume most scientists aren't scammers.
You seem to distrust climate scientsts by default because of a few anecdotes in which scientists have displayed human failings, if somewhat ugly ones. Is this really reasonable and constructive?

Agreed, peer review and degrees aren't enough.
But this isn't the only thing climate scientists have: they have a fairly comprehensive and coherent explanation that makes sense and produces numbers that pass the smell test.
Do you have an alternative theory compatible with most current knowledge? You haven't given us much in the way of numbers so far.
You've given a sensitivity number (which isn't denial territory in my humble opinion) but I've seen no numbers for your variability, your clouds and so on.
You say you're interested in models. Do you have your own? Even a crude one, just to see how the numbers could fit together as well as with the knowledge we have?

It's also my impression that uncertainties tend to be understated but I'm loathe to make forceful pronouncements becuase I haven't done my own crude quantitative homework to see how well numbers outside of the ranges would fit in the big picture.

Anonymous said...


I do not distrust the climate science community by default. I distrust them *because* the hockey stick issue was handled in such an unprofessional way. I would be a lot less distrustful if the IPCC had:

1) Acknowledged that the original hockey stick was fundamentally flawed and removed all references to it (the fact that MBH98 is still included as a reconstruction in AR4 is inexcusable)

2) Put procedures in place to ensure that similar failures do not happen again (e.g. by being very skeptical of any study that makes it case by inventing new statistical methods).

3) Refusing to accept any peer reviewed work unless all data/procedures were made available to others for replication (forcing SteveMc to use FOI requests to access data is inexcusable).

To me the steps above would been done without prompting by any scientific organization that was primarily interested finding out the truth. The fact that those steps were not taken despite years of very public criticism suggests to me that the IPCC is an unaccountable and untrustworthy organization.

I realize that many, if not most, climate scientists are not responsible for the failures of the IPCC. However, even good people can get trapped in toxic organizational cultures and that can lead to serious mistakes (think about the FBI operative that tried to warn her superiors about suspicious characters in flights schools).

Anonymous said...


I take also issue with the claim that the current CO2 hypothesis is a comprehensive and coherent explanation. There are simply too many questions that do not have good answers. For example, all of historical recontructions depend on unverifiable estimates of aerosol forcings. Even though these estimates may be constrained by what is considered to be realistic, my understanding is there is still considerable latitude in setting these values. More importantly, my understanding is the models with high sensitivity tend to have the largest aerosol forcings which suggests that a certain amount of tuning is going on.

There are also the matter of things which simply cannot be modelled. For example, Pielke Sr has numerous papers how land use forcings can have an impact on GMST. The models don't have the resolution to take these things into account.

The same goes with cloud cover - Spencer's hypothesis that cloud cover is chaotic and can vary significantly without any external forcing is consistent with my understanding of chaotic systems. He has produced simple models that show this effect but he has not been able to get the results into the peer reviewed literature. This bothers me because Spencer's theories may not be right but they are certainly plausible enough to merit publication.

My gut feeling on is that there is no single explanation (be it CO2, cosmic rays, clouds, land use, whatever) and we are wasting our time looking for one. I suspect all of these factors play a significant role and even interact with each other in unexpected ways. These interactions are also probably beyond our ability to model at this time which means insisting that alternate explanations must be modelled before they can be accepted creates an artificial barrier to progress.

David B. Benson said...

Raven --- That is the only paper I have found so far relating the PDO to Rossby waves. I opine that the explanation of the approximately 20 year quasi-periodic component as a Rossby wave is quite a good hypothesis. But my current interest is in the amplifier for the solar cycle, since from solar minimum to solar maximum increasing global surface temperatures by (an astonding) 0.17 K or thereabouts. I suppose it is possible that the solar cycle tends to pump the PDO, but not the other way around.

Chaotic cloudiness? Maybe so. Enough to matter much? No. Reasoning is simple: clouds tend to precipitate out, providing a negative feedback; less clouds, more evaporation; more clouds, more precipitation. Looks to be stable around some accumulation point to me.

Anyway, here is another simple model to accounts for Earth's temperature using just CO2 and water vapor:


Anonymous said...


I don't know what you're talking about with the "CO2 hypothesis". Is this a strawman?
Nobody is saying that there's a single explanation. I though you knew better.

Sure they are unknowns. Explanations only have to explain what's known.

Land use, clouds... sure. But how much of an impact, quantitatively?

I'm not insisting that anything be modelled. I mentionned that because you seemed to have an interest in models. It would make sense though: models are useful.
They don't need to be as complex as the real system by the way.
I am insisting that alternative theories be quantitative however. This isn't lit crit.

As to the hockey stick business and the IPCC, I don't know... do you have a better link with more meaty stuff about the egregious torturing of data that's being alledged?
In any case, if your issue is only with the IPCC and its influence, does that mean that you have no issue with the body of work that predates it? Or do you distrust the milieu out of which the IPCC emerged as well?
I distrust academia in general and the "social sciences" district in particular so I think I sort of understand where you're coming from but you seem to have a marked "baby with the bathwater" way of thinking about it.

Anonymous said...

For clarity, make "single explanation" above "single-factor explanation".
I may have used the word "explanation" in to many different capacities.

Anonymous said...


I try to hard to avoid the "baby with the bathwater" problem by looking at each claim as separate issue.

Those claims which are supported by some skeptic sources I don't dispute. For example, Lief Svaalgard has spent a fair amount of time on skeptic forums explaining why he thinks the solar TSI changes have no measurable effect on the climate. He makes a good scientific case and does not take sides when it comes to the broader question on the effect of CO2. His second opinion was enough to convince me that the IPCC is probably right on the TSI issue.

When I first read the summary of Mann 2008 which pegs the MWP at around 1980 levels my first assumption was that Mann had learned from his mistakes and that his analysis would stand up to scrutiny (I posted that opinion on CA when the paper first appeared). I was extremely dismayed to see SteveMc demonstrate that the 2008 paper was yet another meanless exercise in data mining with 'novel' statistical methods.

For that reason I will dismiss papers like Pagani until I can get that indepedent verification. But even then, I try to be careful not to insist that they are wrong - just that they do not provide supporting evidence for things like CO2 sensitivity.

When it comes to my claim of low CO2 sensitivity there are 3 lines of evidence that I see which supports this:

1) The missing tropospherical hot spot.

2) The fact that the actual GMST are below the 95% limit for the AR4 model predictions.

3) The flat/slightly declining ocean heat content.

I am very aware of the counter arguments w.r.t short time periods. However, the data does mean my claim of low sensitivity is plausible even if not proven.

Anonymous said...


Here is a link to a more detailed overview of the hockey stick issues and Steve's reply to the claims that the issues have been addressed:


The issue with the IPCC is not simply the recent papers. There is also the issue of things which are apparently ignored by IPCC. For example, this 1978 paper on internal climate variability:


I know the author of the above paper no longer seems to believe his own paper but the example demonstrates why simply going back an looking for older papers that support the IPCC view does not actually address my concerns about the influence of the IPCC.

Anonymous said...

I haven't read through all the thread, but saying putting CO2 taxes or cap and trade to reduce CO2 emissions is like amputating a leg is a bit weird.

If you end up doing some things differently when you take CO2 into account, and that gives less money people, it probably ends up being similar to just getting ever richer, but at a slower speed.

Probably, for example, and as a thought model, this is important in the context, amputated people on average are probably suffering from problems and are generally less happier than people with all limbs. On the other hand, general happiness in the western world doesn't seem to have increased much since the seventies when reasonable wealth levels for most of the people were achieved.

So I'd say cap and trade or CO2 taxes can not be reasonably compared with amputation.

Also, the emissions control affect the rich countries more who suffer less from slight income reductions. Those who are starving are not producing much CO2 anyway. They don't need to do cuts.

Also, the "more R&D, but no new laws" seems like a very left wing naive way of thinking things - that the government should just give us all a solution.

Firstly, the R&D won't be free anyway. So it's more taxes.

Secondly, the industry and people will not move to less CO2 producing ways, unless there is a cost for producing CO2. Voluntary actions are just pure fantasy.

If there is a cost, then the *industry* can do much more efficient R&D closer to their relevant areas. Their R&D could be much more efficient and application specific than general government programs.

If CO2 emissions have no cost at all, if you save one cent by emitting ten thousand times more CO2, then it has to be done, according to the rules of economics (investors only invest in profit maximizing companies). If that CO2 cost just 2 cents, then the decision would go the other way. The current economics doesn't care about the cost of one's actions to others. That is why there is the legal system, to force a company or a person to take others into account.

I do agree that some bigger more basic science programs (advanced nuclear that was stopped in the seventies) could be good from the government. But doing *only them* would be stupid.

Both laws and R&D should be done.

For the pricing affecting CO2 versus accelerating peak oil (coal electricity vs oil), the cost of electricity affects electric transportation very little - actually the cost of electricity is currently so small that it probably doesn't feature at all when looking at the cost of driving an electric road vehicle. The vehicle, and especially the batteries are the limiting system. Also, the automobile engines are less efficient than centralized stationary power plants, so I think CO2 emissions could go either way, if you compare a gasoline car and a coal-electric car.

And this is a red herring anyway - coal for pure electricity (industry, homes) is one of the first things that need to get rid of. They are the first few thousand low hanging ripe fruits. (Ok, even lower are the super GHG CFC factories, but I think they have been mostly taken care of).

Transport fuels come later. They are much harder to replace. But there is so much work that is so much easier that we don't have to worry that much about transport fuels yet. Saying that transport fuels are hard to replace is no logical motivation to avoid replacing coal electricity.

Anonymous said...

Also, the asteroid threat was a point. I am actually very enthusiastic about space and spacefaring (which the humanity does not currently do).

Asteroid deflection is one important reason of doing it.

There's the slightly humorous saying, "The dinosaurs didn't have a space program" - though I think the Program mentality might have done space progress actually more harm than good.

I also think that if a large asteroid was seen to collide with Earth and, say, with a 70% probability cause the destruction of cities home to tens to hundreds of millions of people (tsunami via ocean impact is probably the most likely), then, it would probably be reasonable and logical to try to do something about it - though it depends a lot on details.

One can't really make blanket statements that are often thrown around either way.

Anonymous said...

Hmm, I read a little about criticism of Mann, and have by no means really digged into the statistics myself.

You say, like Wegman, that the statistical methods were erroneous (decentering instead of centering). Then some climatologists show how different methods (when using centering, or other things) yield pretty much the same results from the data.

And also others produce independent results that are somewhat more different but still similar.

Then you switch to saying that that's because the proxies are wrong fundamentally.


I guess then use different data than trees. Which have been used otherwise anyway and have pointed in the same direction.

I don't really know how a logical person can get evidence from here that a low climate sensitivity is very likely.

What I don't really get is why on Earth do overly complex and error prone statistical methods have to be used so much?

I know some myself about statistics and especially data algorithms and must say that it's a careful business, classification especially...

Probably rigorous statisticians should help scientist in all branches to improve their work a lot!

And don't get me started on using P values wrongly and the wonderfulness of Bayesian inference which actually makes sense and is mathematically correct.

Anonymous said...


I don't think you really appreciate how much energy we need to maintain the people on this planet and how much the cost of energy affects our standard of living. I also think you do not appreciate the scale of the problem and the technical limitations of alternate technologies that will make it impossible to eliminate CO2 emissions no matter what laws the government passes.

Did you read the quote from the newsweek article I posted above? It provided some hard numbers that illustrate why talking about limiting CO2 to 350ppm or 450ppm is a waste of time.

I also think there are lessons to be learned from the CFC ban. Companies initially opposed the ban because viable alternate technology simply did not exist. However, after some years of R&D they were able to find something suitable and were more than willing to support government regulation. That is why I recommend R&D only. If we are lucky we will find something that can be rolled out in the scale required. Any attempt to legislate CO2 reductions before cost effective alternatives are proven will simply kill the economy.

This artical from Germany is quite telling:


"Germany’s environment minister Sigmar Gabriel (Social Democratic Party) is pushing for the construction of new coal-fired power plants in Germany. “We need eight to twelve new coal plants if we want to get out of nuclear energy,”

The "Social Democratic Party" in Germany is a left wing party that has long advocated strong action on the environment. The fact that they are now backing new coal plants illustrates how intractable the problem is.

In short, trying to legislation strict caps on emissions with the technology we have today *IS* like amputating a leg.

If I believed the problem could be solved with a few minor sacrifices I would not quibble about the uncertainty in the science.

Anonymous said...


If you asked a student to solve:
2+2 = ?

How many marks would you give for this answer:
2+2 = 11+11 = (22*2)/11 = 4

For me, the fact that the answer happens to be correct is irrelevant.

Yet the IPCC seems to believe that MBH98 deserved full marks because it included MBH98 as one of the reconstructions. Does that make any sense to you?

The fact that other proxies studies have their own problems is another issue. Even if they were reliable it would still be wrong to continue to cite MBH98.

Anonymous said...

re: sensitivity

I don't see how your lines of evidence support low sensitivity. They're arguments against some models and therefore against sensitivity values given by these models. But how do they constrain sensitivity?

We don't know the oceanic heat content (not with such precision anyway)... and that's the crux of the issue.
Time is needed to average variability and to allow for mixing. How can you derive sensitivity from today's measurements?
Paleo data can give us information from past time and models can give us information from virtual time. If you discard both, what are you left with?
Once you've thrown out the paleo baby and balked at modelling, how can you support a low sensitivity? Uncertainties make it plausible... the same way they make much higher sensitivities plausible.

re: politics

There's a lot of waste that could be trimmed, stuff that could be done differently and so on. People could live more fulfilling and healthier lives with much smaller footprints.
But you're right: jacking up prices would be painful and isn't likely to bring about a better world.

I found your characterisation of the SDP quite amusing by the way. Politicians are so depressing...

Unknown said...

Agree with ANONYMOUS on Raven's sensitivity arguments - the short-time responses can tell you nothing at all about sensitivity because the effect is only barely noticeable (now) over decades.

But on the tropical troposphere "hot spot" - gaaah! Didn't we have an unending fight on this at Lucia's that ended up with her retreating to the semantics of "spot-ness" and the psychology of the IPCC in showing something "red"?

The presence or absence of the "hot spot", if by that we mean amplification of surface warming with height in the tropics, tells you nothing at all about sensitivity, it is not related to sensitivity in any way. Its one and only cause is the change in lapse rate expected with surface warming over water.

But not only that, the amplification is not "absent", it is *observed* on most time-scales, as even Willis Eschenbach's analysis showed, and as was shown earlier by Santer and company. There are some timescales where the data is more ambiguous and you can pick and choose data sets to make it go away, but by and large, what is expected is definitely there when you look at the 1-year to 10-year sort of time-period where we have the most reliable comparable measurements.

Raven, why do you cling to such dubious or ambiguous sources of evidence on sensitivity to argue your case, and reject the vast quantity of paleo and model-based information that shows otherwise? Even the historical CO2 vs. temperature curve for the past 150 years supports a sensitivity well above 1 C/doubling; to doubt that you have to cast doubt on historical temperature measurements too.

To believe what Raven believes you have to
(1) doubt the historical temperature record
(2) doubt any inference from paleo-climate reconstructions
(3) disbelieve essentially every model calculation
(4) put your trust in a handful of measurements that are barely at the threshold of separating signal from noise, or not even actually relevant to the question at hand.

That's conspiracy-theory thinking. Science relies on Occam's razor: the conclusion from many lines of separate evidence on sensitivity is simply unambiguous at this point.

Anonymous said...


You could get a low sensitivity from the temperature record by assuming different values for aerosols, solar and such, isnt'it? Not to mention the lack of measurements of ocean temperatures for most of the period.
It's a bit of a stretch but I don't believe that overturning the temperature record is actually required.

I don't know how you'd rule out higher, long lag sensitivities without paleo or models though.

And Raven alledged institutional bias, not conspiracy.

Anonymous said...


You have no business demanding that everyone trust the IPCC blindly. The onus is on the IPCC to demonstrate that they have examined the science impartially.

The IPCC has not done that. In fact the IPCC has acted in ways that suggest that it is very biased and it does not examine the science impartially. The fact that the MBH98 reconstruction was included in AR4 despite the well documented problems with its methodology is hard evidence of this bais.

I realize that evidence of bias does not automatically mean that the IPCC is wrong. But it does mean that reasonable people will look at what the IPCC says and try to factor out that bias.

In this situation, Occam's razor works in reverse. e.g. if the IPCC is biased but not dishonest it will do whatever it can to make the problem sound as bad as possible. Logically, this means true situation is most likely less serious than what is claimed by the IPCC.

I realize that you do not believe that the IPCC is biased. You are entitled to that opinion but that is your opinion and not a fact and many reasonable people have looked at the evidence and concluded otherwise.

You also need to remember that few of these peer reviewed paleo studies which are used as evidence are actually replicated by anyone. So it is completely unreasonable to insist that these studies be accepted as truth given the large number of serious problems that SteveMc and others have found in the studies that they do choose to analyze. The same goes with model verification studies done by people with a vested interest in the models.

To put it another way: if I am thinking of investing money in a company I do not automatically assume that CEO is a crook. On the other hand, I would never rely soley on the word of the CEO or the word of the CEO's buddies. I would insist on independent verification of the claims. I am applying the same standard to scientific studies.

Most importantly: when I get what I consider to be independent verification of a claim I adjust my views accordingly. This is what I did with the TSI issue and will likely do with others as I get more information.

But it is unlikely that I will be able to accept any results from paleo studies given the current attitude of many scientists.

Anonymous said...


My understanding is the process of calculating CO2 sensitivity goes something like this:

1) Choose suitable estimates for all of the quantifiable forcings.

2) Run the model and compare to the historical record.

3) Repeat 1) and 2) as many times as necessary to get a good match to the historical record.

4) Run model with different CO2 forcings to calculate the sensitivity.

What this process does is attibute all unknown or unquantifiable forcings to CO2. This implies that finding a preiously unknown source of warming must result in a lower CO2 sensitivity if all else is left the same.

The trouble is all else is not necessarily left the same because the estimates for the other quantifiable forcing could be adjusted in a way that would preserve the original CO2 sensitivity while incorporating the new sources of wamring.

I realize that the modellers insist that they do not tune their models in the way I described but I don't believe them because I have some experience with computer models and know that with these kinds of problems there are not a lot of other options if one wants to produce something reasonable.

This is why I put a lot of emphasis on comparing models predictions for the future to real data. That is the only way to determine if the tuning produced a reasonable model instead of simply covering up flaws in the models.

Michael Tobis said...

Raven says:

When it comes to my claim of low CO2 sensitivity there are 3 lines of evidence that I see which supports this:

1) The missing tropospherical hot spot.

2) The fact that the actual GMST are below the 95% limit for the AR4 model predictions.

3) The flat/slightly declining ocean heat content.

This "hot spot" business was new to me. RC seems to have responded to it pretty convincingly at


Suppose it were missing. How does this imply a lower sensitivity?

In the end, I think you are bluffing. The "hot spot" argument the way you make it simply isn't how real dynamics is done. But even if it were, it's not clear how this supports your low sensitivity argument.

As for the ocean heat content, its decline is so implausible that one would look for a measurement error, since ocean skin temperature is obviously increasing and sea level is obviously incerasing. Sure enough, there was one:


By GMST, I suppose you mean global mean surface temperature.

The smoothed temperature is not remotely outside the AR4 predictions, which don't represent events of high frequency. That is not a reading of AR4 that can claim both competence and fairmindedness. Come back in two years with that one and if it's still below the trend we may have something to talk about. Until then this is not a claim that should be treated as even slightly supported.

Anyway, although you seem to be grasping at straws, let's postulate that you aren't. Suppose low sensitivities are plausible. You aren't done. You have to show high sensitivities are implausible to make a sensible case for delaying carbon policy.

You sounded serious at first blush but your points are looking awfully tired already.

David B. Benson said...

Especially since it is clear that Raveen has not bothered to read Annan & Hargreaves.

Anonymous said...


Where is your real, verifiable evidence that CO2 sensitivity is high?

Unverifiable paleo studies which rely on a lot of guessing and data from a few points on the globe are not the same as a repeatable experiment that produces real data each time it is run. It is a mistake to treat these paleo studies as the last word on sensitivity.

IOW, if the real data that we collect today suggests that CO2 sensitivity is lower than the paleo studies say then the paleo studies are most likely wrong or the paleo conditions are not representative of today. No futher explaination is required.

If you disagree with my assertion then please explain how science could move forward if those paleo studies are, in fact, wrong. If you think about it the only way they could be shown to be wrong is if the real data shows that CO2 sensitivity is on the low side. For that reason, it is unreasonable to insist on refuting the paleo studies before concluding that the real data suggests low sensitivity.


Deductions with baysian satistics are completely dependent on the choice of priors. For that reason I do not consider Annan&Hargreaves to be proof of anything. I also suspect that such an analysis would have been roundly rejected as junk if it did not happen to confirm what people already believed to be true.

The last point is a good way to evaluate the worth of a scientific argument. i.e. when you see a paper that draws conclusions by analyzing data you should ask yourself if you would accept those conclusions if the answer did not confirm what you already believed. If the answer is 'no' then the methods used in the paper are not particularily interesting.

I have used that approach to reject skeptical arguments that intially looked plausible.

Anonymous said...


To clarify my issue with A&H:
They make the explicit assumption that all estimates of climate sensitivity are independent measurements. This is not strictly true because CO2 sensitivity is an inferred quantity that depends on a common set assumptions about how the climate works. If these assumptions are wrong then that would introduce a bias into the all of the estimates.

A non-negliable GCR effect over long time scales is an example of a potential source for systematic bias.

David B. Benson said...

Raven --- Look at the priors used to see if those used concern you. I've done enough Bayesian analyses to know that the contribution of the priors rapidly vanishes. Learning about this point might help you to see that Annan & Hargreaves actually do arri8ve at the distribution that almost everybody uses; there is a reason for this concensus.

Anonymous said...

And yes - systematic errors could, in theory, mean that CO2 sensitivity is underestimated....

Anonymous said...

Begin quote, Raven:
I also think there are lessons to be learned from the CFC ban. Companies initially opposed the ban because viable alternate technology simply did not exist. However, after some years of R&D they were able to find something suitable and were more than willing to support government regulation. That is why I recommend R&D only.
End quote

And do you believe the companies had done the R&D if there had been no imminent laws and push for changes? Do you think they did it voluntarily? Would the change have happened if there had been no laws at all?

That is just not logical. Companies are created and exist to make money for the shareholders. Not do environmental benefits to others.

The changes happened because of pressure on the companies, not voluntarily. *That* is the lesson to be learned. Once they realized that business as usual was not an option, they stopped complaining and lobbying and funding the fake experts, but started acting and solved the problems.

As for "giving marks" to MBH, well, that's a false thought model.
Correctness is important, but absolute correctness everywhere is an unrealistic assumption. People flew to the moon and had lots of hardware break on the way, but there was enough redundancy in the system to cope with it. (Unless one is part of those folks who insist that the moon landing was a hoax since any single failure would have doomed them, so the probability of all sequential successes is very small.)

I haven't studied the tree proxy and principal component issue very closely myself. I see it is possible that it has been defended too enthusiastically. If that has happened, too bad. The results seem to be all the same anyway, so it doesn't change the fundamentals much. Wegman never claimed the error they postulated to have any effect on the results. They specifically said they did not research that.
And finally, like I said, there are other completely independent past climate reconstruction methods, like boreholes.
They differ somewhat and have different error sources then again, I'm sure. I haven't looked that much into it.

As a somewhat related note, I've been thinking lately, and I see that the existence of multiple open source climate models could be very beneficial for development and correctness of climate models in general.

On the other hand, there are good reasons to think that it's not useful to use too much resources on the basic science of the existence of global warming anymore (it's good enough already, and improvements will not convince those who have chosen otherwise anyway), and instead focus should now be on solutions.

Anonymous said...

Quote Raven:
Deductions with baysian satistics are completely dependent on the choice of priors.
End Quote

This shows the very common and fundamental misunderstanding of statistics.

In reality, one can not make inferences from data without making assumptions.

This can not be stressed enough.
Bayesian methods are mathematically correct ways of getting knowledge out of data.

If you claim to make inferences from data without making assumptions, you most probably don't know what you're doing and your assumptions are just hidden. Sometimes the hidden assumptions are correct enough that the results are useful.

But there always are assumptions.

Bayesians just make them explicitly visible in the form of the prior.

Anonymous said...


In retrospect my criticisms of A&H were too harsh and I agree that these kinds of analyses often do provide useful insights provided one always remembers the caveats that go with the analysis.

In the case of A&H the caveat is all of the sensitivity estimates presume that all of the significant climate forcings at work have been correctly quantifed. If we discover new forcings (e.g. cosmic rays) then the results of the baysian analysis are meaningless.

Based on what I have read I do not believe we know what all of the significant climate forcings are. In particular, I don't think the effect of internal variability has been properly taken into account. For that reason I do not think the probabilities produced by A&H are reliable enough to use for policy decisions.

Anonymous said...


You are entitled to dismiss the significance of the problems with the hockey stick. I do not share that opinion. I feel the problems with the various reconstructions are well documented and serious. The fact that so many people in the climate science community choose to minimize these issues is why I do not feel I can trust their scientific judgement.

IOW, I am forced to reject claims made by the IPCC unless I can find unambiguous data and/or independent opinions that support the view.

My original intent in participating in this discussion was to make it clear that the IPCC's actions have damaged its credibility and the failure to come to terms with this is the reason why many reasonable people are AGW sceptics. If you really believe that immediate action is required then you will likely find that acknowledging the problems and baises with the IPCC is a more effective approach than pretending they don't exist.

Anonymous said...

The IPCC could use a more level-headed auditor than this McIntyre fellow. The link you provided was light on technical details and heavy on political grandstanding.
Regardless, I'd say there is indeed a credibility issue. I'm not coming at his with predetermined conclusions though so I'm more worried about the upper part than the lower part of their error bars myself. That's where the danger lies.

Raven, I'm sorry to insist but you haven't explained how you figure sensitivity is low based on what you call "real data".
Making hay of shortcomings on the part of the scientific establishment is fine but what alternative have you got?

David B. Benson said...

No change in GCRs in past 50+ years.

Climate warmed significantly in past 50+ years.

Conclusion: its not GCRs.

Anonymous said...


I suggested that cosmic rays have an effect over longer time scales which would result in much lower estimates for CO2 sensitivity.

You can find more information here:

Note that Shaviv claims:
"A more recent analysis, which includes: (a) Corrections to the temperature reconstruction due to ocean pH variations, and (b) more empirical comparisons between actual temperature variations and changes in the radiative budget further constrain the global sensitivity to about 1-1.5°C change for CO2 doubling.

A 1-1.5C sensitivity also happens to be consistent with the data being collected today.

Here are Shaviv's responses to his critics.


I suggest you look at those responses before responding with links from RC that "debunk" Shaviv's claims.

I have noticed a tendency in this debate where some people only look at sources that confirm their prejudices and then assume that is the last word on the topic.

Anonymous said...


I have found the childish back and forth between between SteveMC and the team at RC (mostly Gavin) to be quite tiresome. I have given up trying to figure out who started it and try to focus on the substance behind the mutual jabs.

Here is another link to his key arguments:

My argument for low sensitivity from the current data comes from the fact that there is no warming in the oceans or in the air. This means there is no "warming in the pipe" and any additional energy from GHG forcings must have already been radiated into space. The mechanism is likely related to clouds. The lack of a hot spot in the troposphere is evidence that supports the idea that the models have got the basic feedback mechanisms wrong.

Pielke Sr has a number of posts and papers which explain why a short number of years with flat or declining OHC is significant:


Lucia has used the statistical techniques from Santer 2008 to compare the AR4 model runs to actual data and found that the models consistently predict more warming than actually occurred over the last 30 years:


These kinds of analysis are not the final word on this issue but they are substantial and do demonstrate that low sensitivity is a plausible conclusion given the data available.

David B. Benson said...


Using the oceans as a calorimeter to quantify the solar radiative forcing
Using the oceans as a calorimeter to quantify the solar radiative forcing

Nir J. Shaviv

"For a gray body earth without any feedbacks, this temperature rise is ΔT×2 = 1.2°C."

once again. How, Raven, can it be less than this?

And I posted for you above Greenhouse 101 which shows how to approximate the temperature of the Earth. I guess you didn't bother with it.

Anonymous said...


What point are you trying to make with the Shaviv article? That the current cooling is a result of a large solar effect? If the sun is responsible for the current cooling then it must also be responsible for the some of the previous warming. This means CO2 sensitivity must be less than the models claim because the models attribute all of the previous warming to GHGs.

I looked at the GH101 link again and I have seen the calculations before. I don't see why you think they are so important. I am not claiming the Greenhouse effect does not exist. Just that the CO2 sensitivity is not as high as the models claim.

The models predict a sensitivity higher than the theoretical sensitivity because of a net positive feedback. If the real net feedbacks are negative then the actual sensitivity will be less than the theoretical sensitivity.

It is worth noting that most stable natural systems have negative feedback mechanisms that keep them stable. So the idea that the climate would automatically compensate for extra GHGs is more plausible than the claim that extra GHGs could lead to runaway warming.

Anonymous said...

And if the world is flat, it won't be spherical. Sounds familiar reasoning, right?

Water vapor feedbacks are positive, it's physics again.

This all played out after Arrhenius.

Search for vapor here:


You can see many other common viewpoints and hypotheses of the current do nothing establishment that actually came and went in the scientific world decades ago, like that CO2 can't have an effect because water vapor already absorbs all the IR etc etc.

Anonymous said...


The theoretical water feedback may be positive but clouds are a chaotic wildcard that can theoretically provide a large negative feedback. The current crop of models rely on non physical parameterizations to model the behavoir of clouds - parameterizations that may be good enough to provide plausible hindcasts but give us no confidence that the models can predict future trends with any degree of reliability.

This paper provides real data that supports that claim:


"Earthshine and FD analyses show contemporaneous and climatologically significant increases in the Earth’s reflectance from the outset of our earthshine measurements beginning in late 1998 roughly until mid- 2000. After that and to date, all three show a roughly constant terrestrial albedo, except for the FD data in the most recent years.
This implies a reduction in the net sunlight reaching Earth. In the context of the recent climate change, it is important to point out that the physical causes behind these large decadal variations in albedo are still unknown, and that we just don’t know yet whether we should expect the albedo changes observed during the modern period to persist into the future"

At the end of the day experimental data ALWAYS trumps theory and the data that has been coming in over the last decade casts considerable doubt on the theory and suggests that important elements are missing.

The response from the pro-IPPC crowd to this contradictory data is basically: "trust us, our theory is right, the warming is just hiding and will reappear soon".

The theory could be right but believing in the theory given the data available requires a leap of faith which I am not prepared to take at this time. IMHO, the flat earthers are the ones demanding unquestioning faith in the theory even when the data appears to contradict it.

Anonymous said...

BTW - I know that the Palle paper conclusions are trying to say "don't treat this increase in albedo as evidence that AGW is not happening because we don't know why it occurred and it could reverse at any time".

That statement may be true but this statement is also true: "sudden unexplained changes albedo mean that we have no idea whether a significant portion of the warming in the lst 50 years was cause by CO2 or a similar sudden shift in albedo".

Anonymous said...

You're slipping into denial Raven.
Are you saying we don't know how much forcing is attributable to CO2? The concentration and the effect on radiative transfer have been measured.

The negative feedback that prevents a runaway is well known: OLR is a function of temperature.
I don't know that anyone is saying that CO2, methane or any such gas is going to cause runaway warming. That's another strawman of yours.

You say observations are paramount but you discard some (CERES looks at incoming and outgoing radiation) and misrepresent others (again, we don't know exactly how much heat there is in the system).

You have still not explained how you've come up with a low sensitivity. Your links above contain arguments (and I'm being generous with the Piekle post here) against some models. At, best, they undermine other people's sensitivity but they do not support yours.
I'm going to try one last time by using a different angle: how have you determined that high sensitivities aren't plausible?

You're apparently suggesting that sensitivity is a function of warming over a few years. It would follow that sensitivity is variable over a time period as small as a decade and was much higher ten years ago. This is not what sensitivity is about. If what you call sensitivity is simply the first derivative of temperature measurements, then you've got a circular argument.
But if you use the established definition of sensitivity, the variability you have talked about in several posts above means that observations over a short period are basically useless to determine sensitivity. So which is it? Is there strong variability or not?

Anonymous said...


My understanding is CO2 sensitivity is the number derivived by running the models with different CO2 forcings after the models have been tuned to provide a reasonable match to past temperatures. e.g. run the model with CO2 forcing set to X. Run the model with CO2 forcing set to 2X and calculate the ratio.

Is this what you think CO2 sensitivity is?

If it is then the sensitivity calculation depends the assumption that the models have correctly accounted for all forcings that affected past temperatures. Changing these forcings will result in different values for CO2 sensitivity since the model parameterizations will have to be re-tuned to match the historical record.

Anonymous said...

How one would derive sensitivity from a model is one thing. The definition of sensitivity is something else. Please refer to some introductory material if you're unsure about the definition.
You can derive sensitivity however you want as long as you're consistent and you make sense.

If there was indeed a mechanism linking increased temperatures to a significant negative net cloud forcing, you could indeed get a low sensitivity. But that's speculative. Chaotic variations in cloud cover don't come into it.
The direct forcing from CO2 and the effect of temperature on vapor pressure is hopefully not contested.

Anonymous said...


The issue is not the theoretical definition of sensitivity but how it is calculated by people who claim it is high. My understanding from RC and other sources is it is calculated from the climate model outputs. This means the reliability of the number depends entirely on the reliability of the climate model used. If the climate model does not accurately produce hindcasts then the sensitivity number from that particular model is a meanlingless quantity.

That is where cloud cover variations come in. If there is a decadal trend over the last 30 years which the models did not take into account then all of the sensitivity numbers from those models must be discarded.

At that point it becomes a question of how the modellers incorporate the trend into their models. They could increase aerosols forcings or make similar adjustments that would offset the cloud trend while ensuring that the model continued to produce the same CO2 sensitivity number. However, there is a limit what tuning can be done.

I am making the general assumption that finding new sources of warming will result in models with lower sensitivity because that would be the only way to produce a credible hindcast after incorporating the new forcings.

That said, I will concede that reduced sensitivity is not necessarily going to be the outcome since matching the hindcasts is largely a curve fitting excercise and the modellers have many knobs that they can play with.

Here is a post that uses a simple that illustrates how the CO2 sensitivity is entirely dependent on the forcings incorporated into a model:


Anonymous said...

I know one could come up with lower sensitivity by ditching paleo and using lower aerosol values for the last century. In fact, I suggested as much a few posts up.

But you have to understand that sensitivity is not only a model output. It's not arbitrary.
Unless it is quite small and random temperature variations are equally short and small, it's not going to be driven by the last decade's temperature.
You don't even need your own model to figure this out. You can derive sensitivities for yourself from your own static values.

Your example with the trend in cloud cover illustrates how it's more complicated than you make it to be: the hypothetical trend could be a recurring random variation, an exceptional event or a result of the warming.
If it's a random variation, you need to include it in your calculations or you need data for a period long enough that the variation averages out. But you don't need both so you could come up with the right sensitivity even if you didn't know anything about the variation.
If it's an exceptional event, there's no way around taking it into account.
But if the trend is a result of the warming, it's a feedback and it's part of the sensitivity. The higher sensitivity value you come up with by ignoring cloud cover would be the right one.
Again, with data from a long enough period, it should be quite plain what's behind trends in cloud cover. Jumping to conclusions based on a few decades (or worse: a few years) of sparse data is fraught with risk.

You apparently belive that pointing out uncertainties enables you to decree that sensitivity is low. But uncertainties cut both ways! If we don't know, we really don't know...

Anonymous said...


"The issue is not the theoretical definition of sensitivity but how it is calculated by people who claim it is high."

Actually, if you want to determine mitigation policy, the issue is not the definition, nor how it is calculated. The issue is what is its actual value on today's planet Earth.

Climate sensitivity to CO2 is defined as the number of degrees of warming the earth will experience at equilibrium given a doubled level of CO2 in the atmosphere. It includes short term feedbacks such as water vapor, clouds and albedo from sea ice. It does not include long term feedbacks such as albedo from ice sheets and carbon cycle changes.

It can be, and has been, calculated in many ways and from all available evidence appears to be around 3oC. What people are trying to tell you is that even if valid, your arguments only undermine our certainty in that number but do not in any way imply it must be lower. You rightly point out that cloud cover changes are highly uncertain and poorly understood, but you have not explained why this makes you confident that the model numbers are too high and not too low.

If cloud cover feedbacks were in fact a natural negative feedback on warming then a globally warm MWP is even less likely and you have a real challenge in explaining the glacial-interglacial cycles.

I would greatly appreciate that you address this concern first and foremost, but while I am commenting I would like to make a couple of other points.

<"If the sun is responsible for the current cooling then it must also be responsible for the some of the previous warming."

This does not follow at all.

"This means CO2 sensitivity must be less than the models claim because the models attribute all of the previous warming to GHGs."

This is not true. There are about a dozen major forcing factors whose measurement are fed into these models, the resulting hindcast is a result of all their interactions. For example, about 30% of the early 20th century warming is attributable to solar. This is a common problem on both sides, forgetting that we are not searching for the One True Climate Driver and it must be either CO2 or the sun or GCR's or Magic Pixie Dust (aka unforced natural cycles). So this is a straw man argument.

But again, please do not focus on this, rather the initial point I made, if you have time for only one.

Anonymous said...


I think I can agree with everything you said in your last post including the bit about including chaotic changes to cloud cover in CO2 sensitivity calculations if the CO2 induced warming was causing them.

However, my feeling is the internal variability of the system is being underestimated and that this variability cannot be attributed to a forcing like CO2. My evidence for this variability is the temperature record which clearly shows distinct regime changes that last decades and don't appear to be connected with anything. There is support in the peer reviewed literature for the view (e.g. the Robuck paper I referenced above).

My feeling is that if one took this variability into account and accepted that sensitivity from the paleo-records is not necessarily correct then one could come up a model that matched the temperature record but had a much lower sensitivity.

That said, I realize that I cannot specifically exclude higher sensitivities, all I can claim that the probability of higher sensitivities is lower given the data over the last 50 years and the unreliability of the paleo-data.

What this means for policy choices is the interesting question. We cannot exclude the possibility of a rogue asteroid strike in the next 50 years and the risk is probably large enough to justify developing an early warning system. On the hand, spending trillions on space technology required to deflect the asteroid is premature. We face a similar problem when it comes to climate change. No one can reasonably argue against R&D and the promotion alternate technologies but it is hard to justify inflicting significant harm on society in order protect against a risk that may be real but cannot be quantified.

In the end, adaptation is most likely the safest bet because any money invested will produce results and the technology already exists (e.g. dikes and dams).

Anonymous said...


I think I answered your main question in the post above.

I see what you are saying on cloud negative feedback and the MWP but I am not specifically arguing that cloud feedback is negative. I am saying that it is chaotic and can induce changes in temperatures that last decades for no particular reason and that it is a mistake to assume that all temperature changes have to have a cause.

As for the ice ages: Shaviv has already put forward an alternate hypothesis that includes a much lower CO2 sensitivity. The trouble is we don't have the cloud cover data that would allow us to falsify one hypothesis or the other. It is also possible that they are both wrong and the cloud behavoir on a planet covered in mile high ice sheets is sufficient to provide a significant additional feedback as those ice sheets melt or form.

The claim that 30% of the early 20th century rise is attributed to solar is no longer supported by the data (at least according to Leif Svalgaard). This means that any climate model that assumes this to be true is wrong.

That said, my understanding is the models are fairly pliable and some of them were able to compensate for the solar component by tuning the aerosols. This pliability is the reasons why I don't consider models to be a reliable source of information on climate.

David B. Benson said...

Raven --- Shaviv is wrong about low sensitivity and I understand his mistake in the paper I referenced.

My point was that a gray body earth has a no-feedback-sensitivity of 1.2 K. You had earlier posted an even smaller value.

As Ray Peierrehumbert shows admirably in

Climate Book

there are no significant negative feedbacks. There are limit states, at one end the world entirely iced over. This state, so-called Snowball Earth happened (or nearly so) at least once in Earth's long history. The other, a pseudo-limit, is no ice at all. That has occurred for most of the past 65 million years.

So cycling between glacial and interglacial via orbital forcing requires exquisite balance in the modern (last 4 million years) climate system. A charney sensitivity of close to 3 K provides that. Nothing else can explain the pseudoperiodic variations seen in LR04 stack and Antarctic ice cores. A low climat senstivity would not produce such large excursions.

I think you are being tendentious. There are many small matters to be sorted out in climatology; the approximate value for climate sensitivity of around 3 K is not one of them.