The Blackboard

Lucia, to sum up:

Recent (approx. 10 year) global surface temperature trends are ‘near’ the lower bound of the various 95% confidence intervals of GCM-predictions.

This probably means that either;

1) some ‘environmental noise pulse’, like some atypical (and so far undetected) transient increase in mixing of the deep ocean with the surface waters is responsible and/or

2) some (small?) ‘error’ in parameterization in most GCMs [like the fact that they all estimate about 0.5ºC higher tropical upper troposphere temperature (with respect to surface temperature) over the past century – in 'conflict with' few and far between tropical radiosonde data] is likely to be responsible.

But long term trends are probably still onward and upward (maybe closer to 1ºC/century than 2ºC/century)?

ecent (approx. 10 year) global surface temperature trends are ‘near’ the lower bound of the various 95% confidence intervals of GCM-predictions.

Yes. This is based on what my figure shows and doesn’t constitute part of the presentation by Michaels.

On your reasoning: I’m not sure if you list is exhaustive, but I agree the what the models and data show are that the long term trends are likely to persist positive. I think the models as currently constituted will likley continue to overpredict the earth’s response at specified levels of forcing. But we can’t be sure of that. Arthur …. someone suggested some ideas why they could overshoot now, but if the problem is poorly parameterized mixing in the oceans, they might still be ok or even underpredicting in the future. (In my opinion, I don’t think they’ve fleshed out that argument, but it’s been made. It’s sufficiently complicated that I can’t immediately say it’s wrong.)

Lucia,

I note from Michaels’ ‘Figure 4′, based on his somewhat odd assertion that it’s reasonable to postulate 2009 temperatures being the same as 2008, that the case for the prosecution is even stronger. To what extent are we looking at an analysis that tells us anything more than the rather obvious fact that very recent temperatures have been on the cool side of natural variation? If, for example, 2009 were to take us back to, say, 2005 levels, how would the models then, roughly, be shaping up? I realise this is a somewhat vague question – I am no statistician – and I’m not asking you to do the work on my ‘what if’, I’m just wanting to get a sense of how much this depends upon the immediate stage of development (or in other words, would we have had a ‘top of expectations’ analysis if we had cherry-picked 1998 at the time?). Cheers

Simon, the chances of 2009 being back up to 2005 levels is, I would say, very small, considering the return to La Nina conditions we are experiencing.

Thanks Lucia, that answered a lot of my questions

Andrew,

That rather misses the point of my question, I think. We are not in disagreement, I presume, that ENSO variations can be considered as a natural cycle (putting aside for now any possible long-term effect upon the cycle of any putative climate change)? So, rather regardless of when that cycle comes around, I am questioning how subject to immediate natural variation such an analysis is….? I’m asking quite naively how subject to current noise we are. We know also that we have an extended solar cycle low, for further example.

Simon-
When we statistical analysis are based on what we guess for future temperatures, we have obviously entered the realm of speculation. The post was long ,and I figured stopping at figure 3 would answer Andrews questions.

The issue with the analysis is this: has the recent cold spell been cool enough to decrease our confidence models can project of forecast.

As it happens, 1998 predates the AR4 projections. But, at the time, it could be fair to do a similar comparison to the TAR to see how they compared. We always use the most recent data.

The answers are always in probablity. And yes, if we happen to have very unusual weather, the the answer from any statistical analysis may be “If the models are correct, this weather event would have occurred in less than 1% of cases.” Reality is this should happen 1% of the time.

Lucia,

Ta – I’m interested in your expression of it. I don’t personally put too much weight on the ‘precision’ of the models at this stage – they seem too uncertain of cloud effects, too imprecise about current aerosols, too much out already with observations of ocean CO2 uptake, and so on. So, I rather think of them as a rough shot, and think that temperatures are more likely to go up than down . There’s a sort of irony niggling at me here, though, which is that Michaels is arguing that our estimates of impacts may be too high, whereas it might be said that they are just rather mistimed. If the trend turns out to be say 1.6C/century rather than 2C/century ( a relatively conservative projection, I realise), then we might be hitting the same temperature in the 2120s rather than by 2100. In terms of the current policy decisions in regard to such a future, I’m not sure that the precision of the projected rate of change makes so much difference.

Simon

If the trend turns out to be say 1.6C/century rather than 2C/century ( a relatively conservative projection, I realise)

But if the trend is 1.6C/century while modelers say 2.0C/century, then modelers are over-predicting now, and overprediction propagates into economic and environmental assessments. Not getting the trend right also cast doubt on our ability to estimate the overall climate sensitivity. It may be on the lower side of the estimated range.

So, both the timing and the ultimate destination for warming matters when making some judgments about choice and timing of policy responses.

Sure Lucia, I do realise that (and am not underestimating the importance of wanting to ‘get it right’). But if we recognise that we’re not really doing b all at present to mitigate, then the possibility of consequences being some 20% less (or later) than projected is not all that comforting . I don’t think you are pursuing this line, but I do think that the opportunity to show models to be over-egged is seen by some as being a ‘disproof’ of the problem, as if this were a binary issue.

Lucia,

Michaels’ testimony uses HadCrut data. You suggest comparing this with GISStemp. Both however, are largely land surface based metrics which tend to show high compared to satellite data.

So why not run a verification using UAH, lower troposphere and Argos ocean temp data. Based on the most recent ocean data [ref: Pielke Sr on the observed data vs. the GISS model] my hunch is that one will find that the models overstate things by a good margin compared to what is happening in the observable universe.

As you correctly note, there is a lot of political credibility at stake. The implications of allocating billions of tax payers dollars on purported mitigation policies that are based on questionable information are far reaching. My sense is that political posturing notwithstanding, governments in the EU, Canada, New Zealand and most recently Australia have become aware of sufficiently bothersome question marks in the margins about climate data so as to quietly take a time out. The next few months will tell us whether we might see something similar in Washington.

Simon

Are you saying that model predictions are not subject to “disproof?” Disproof is not – or should not – be a matter for “some” to concern themselves with. The models share a prediction of a trend. Either they are correct in that prediction or not. It is certainly too soon for what you call a “binary” decision, but the data cannot be ignored, based on your preference for outcome. We have just been told in the media once again that “things are even worse.” Yet the facts on the ground tell us the opposite for now. If you had a one in twenty chance that your medical treatment was working, you’d no doubt be looking for a new doctor. There is about one in twenty chance that global temperatures like the ones we are seeing could be within the predictions of the models. Not impossible, but not particularly likely. The more time that passes with these conditions, the less likely.

tetris (Comment#10413)

Michaels’ testimony uses HadCrut data. You suggest comparing this with GISStemp. Both however, are largely land surface based metrics

That is simply not true. Both (given that they’re the land/sea records) are largely sea-surface based.

Hi Lucia,

As the person directly responsible for the analyses in Pat’s testimony, I can answer any questions that you may have.

It is worth keeping in mind that Pat’s testimony was not written up like a scientific paper, so many of the details that would be included in a paper are absent from the testimony.

The analysis is very new and we are in the process of getting things in order for a submission to a journal somewhere, but we are still in the early stages…i.e., still hashing out what exactly we are going to do and how we are going to do it.

Assuredly, we will incorporate the GISS temps, just didn’t have them ready for the testimony (which was put together with little time to spare–and, the HadCRUT is the IPCC’s favorite dataset (Which, BTW, is what Pat’s Figure 1 was taken from).

Also, you have a few other details wrong here and there.

I know that you have been working on some model range-type stuff, but I think what we are doing is a bit different. Gavin et al. often claim that the recent years are too short a time period to accurate assess the observations against model expectations. With our analysis, we are trying to put the recent years in their proper model context.

Fire away your questions–I’d love a replication of our work and thoughts about whether or not it is a reasonable approach.

Thanks for your interest,

-Chip

Tetris–

Michaels’ testimony uses HadCrut data. You suggest comparing this with GISStemp. Both however, are largely land surface based metrics which tend to show high compared to satellite data.

I suggest using both. The surface temperature from the models is at 2m. For that reason, comparison to surface temperatures is more appropriate that other metrics. Both HadCrut and GISS are surface, including ocean and land.

That said, another blogger is working on creating a metric to match UAH and RSS. When he’s done, we’ll compare those to the specific models projections.

Simon– I’m assuming tetris means surface and accidentally said “land” for “land and sea”.

MarkB (Comment#10414)

Of course model projections can be shown to be wrong. That is entirely obvious. It’s perfectly possible of course, that they might be wrong to the upside for the next twenty years and wrong to the downside thereafter – we don’t know until we can tell.

However, if it turns out that the models are not as useful as we would hope, does that ‘disprove’ the basic physics upon which they are based? Certainly not – and it is a very familiar misapplication of a trend being falsified to suggest that AGW is therefore falsified.

You seem to have missed my point. I’m not especially worried about a given temperature being reached by 2100 rather than 2025 (say). I am concerned about temperatures rising, and that may be the fact of the matter entirely regardless of the accuracy of model trends. To turn your analogy back upon you, if a second opinion told you that you had 12 years to live rather than 10 if you didn’t stop smoking would you stop worrying and go on puffing away?

Woops, that was meant to be “reached by 2100 rather than 2125″.

Lucia
Now this is not a very useful analysis or rationale.
“But if the trend is 1.6C/century while modelers say 2.0C/century, then modelers are over-predicting now, and overprediction propagates into economic and environmental assessments. ”

Are you making a case that there would be different goals, economically and environmentally, if the trend was 1.6 and not 2? Because you are implying this in your statement.
Your figure of 1.6/decade also comes from 8 years of data… Not particularly useful, especially when we have a lot more data than that. If you were to do an analysis of any contiguous 8 year trends, I bet they’d typically be greater than 1.6/decade. Have you considered doing a statistical analysis of 8 year trends for the whole dataset (back to 1850 or whenever?)

Chip–
Thanks for stopping by.

The observation about the methodology being ambiguous was an observation, rather than criticism per se. Andrew asked, so to the extent I couldn’t figure it out I tried to guess. But, obviously, to the extent I guessed wrong, I couldn’t replicate. In a journal article, I’m sure the specifics would be provided, but I agree that if one is called up to answer questions in congress that might not be possible.

When Gavin and other claim one can’t make comparisons using short time frames, they generally a) make vague claims and b) don’t support them. Obviously, if the projection was 10 C/decade, the current would show that’s obviously wrong!

For me to say more about reasonableness, I’ll need more details. As I said in the post– the general method is one Gavin used. But Gavin got different answers, partly because, by happens stance, he computed his histogram for range of trends starting in 2000. For 8 year trends, that give almost the maximum spread of trends for any choice of start year! Averaging over many years seems better– but to really know we might want to test the method running some tests with synthetic data.

Simon–
What if the total warming at steady state is lower as opposed to reaching the same level, but later? If the trend is low now, the first is at least as likely as the second. The fact is, we don’t know.

Nathan,

Are you making a case that there would be different goals, economically and environmentally, if the trend was 1.6 and not 2? Because you are implying this in your statement.

Only implying? I said it directly. How quickly we need to respond, how much time we have for adaptation or mitigation are affected by the total rate of warming and the ultimate level at any level of CO2. Having 40 years to do implement technologies different from having 30 years. Yes.

Your figure of 1.6/decade also comes from 8 years of data…

What are you talking about? Simon suggested 1.6 C/century as a hypothetical. It’s also happens to match the trend since 1980. That’s more than 8 years. It’s also the close to he trend projected in the TAR.

Not particularly useful, especially when we have a lot more data than that. If you were to do an analysis of any contiguous 8 year trends, I bet they’d typically be greater than 1.6/decade.

Download the data. Runs the least squares. Assuming you use the term “typical” the same way most people do, you’ll discover you lost this bet before you placed it.

Have you considered doing a statistical analysis of 8 year trends for the whole dataset (back to 1850 or whenever?)

The thermometer record starts around 1880. So… not back to 1850.

I have done it since the beginning of the thermometer record. The data are availlable from HadCrut and GISSTemp. You can compute the 8 year trends. After wards, you can use other functions to count, average etc.

Lucia,

Well, of course, that’s conceivable ( I mean, I don’t freakin’ know! ). I’m just giving it my “would I get on the plane?” assessment. Risk assessments include the possibility of everything turning out fine, of course. But I wouldn’t get on that plane if you told me I had a 1 in a 100 chance of a bad outcome from doing so. I expect better chances for my future, and I’m prepared to mitigate against risks, even if I know that the calculation of such risks remains uncertain. i don’t think this is an expression of the ‘precautionary principle’, btw , since the simple notion of that doesn’t proceed from risk assessment.

I agree with you that we don’t know, and that any least worst outcome is feasible. But you won’t be surprised when I go on to say that worser outcomes are also feasible. It hardly seems a really bright idea to me to go on increasing GHG concentrations. I would rather take the risk of being wrong on the cautious side than chancing being right on the reckless side.

I don’t take comfort from uncertainty.

Lucia
So what should we differently between 1.6 and 2C/decade? It’s true that the rate of warming determines what we do, but what difference would there be between 1.6 and 2C? What would we do differently? There is no real plan for AGW anyway, so the only goal here it would seem is to encourage Governments to do less. Is this your goal?

Simon

It hardly seems a really bright idea to me to go on increasing GHG concentrations.

I’m for reducing the rate at which we use fossil fuels. I’m for promoting many technologies, and would especially like to see us bringing more nuclear plants on line to replace baseload from fossil fuel fired plants.

So, I don’t disagree with you. No one is suggesting you should take comfort from uncertainty. I just think it’s useful to compare projections to data to see how certain we can be. The current models happen to overshoot on global surface temperature.

IF the trend is truly about 1ºC/century (rather than 2ºC/century), it probably only means we have a little more time to make early mistakes in a (proper) rush to mitigate, and still avoid Jim Hansen’s dreaded tipping points. With business as usual, tipping points may merely arrive a generation later.

A +1ºC/century trend is no win for deniers!

Nathan–
I described some of my policy preferences to Simon– the answer to your question is: I want to see use reduce use of fossil fuels.

I’m not sure why you think the only goal to testing models is to do whatever you think is “less” than what has been done in the past. The most obvious goal for testing models against data is to test how well the models do what we are being told they can do: predict the evolution of warming over time.

Len–
No +1 C/century is no win for deniers! But it does give us a little more time to do things. I think that matters. Maybe Nathan doesn’t.

But then, it’s been rather clear when we quantify, Nathan’s definition of “trivial” amounts of tax paper money to waste was different from mine. So, it may be his definitions of things like ‘no difference’ or “exactly the same” differ from mine too.

I apologise as this is rather OT but …

Lucia your Figure 1a, The model outputs:

“Me thinks they squiggle too much”.

Compared with the temperature record (or records as we seem to have more than one world), the model outputs seem to have too much squiggle (or variance). If I was still an elecrical engineer I would apply a bit more capacitance (ocean) or better still a bit more conductance (lower the climatic sensitivity) as they just do not seem to be damped enough.

May be it is just my eyes or a difference in smoothing algorithms but I suspect some of the model outputs are insufficiently damped, a bit like an old jalopy with shot shocks.

Now if the ones with the most variance are the ones with the highest CS and reality has a low variance (in my eyes it does) then well … perhaps the models just need a new set of shocks.

You must pardon me, I am not usually this OT or so deep in my cups (hick, pardon me).

Alex

Alex–
Do you mean the individual model runs? A few of the modeling groups admit their “weather noise” is a bit too enthusiastic. I’m not sure if it’s their parameterization for ocean mixing or something to do with the numerics… But yes. There is one model I nicked named “planet alternating current”.

It’s up to others to decide whether the odd “weather noise” makes them more or less confident in the model’s ability to predict.

That said: It’s also possible the recent happened to be a period of unusual “smoothness” for the earth data. It is smoother than periods with volcanic eruptions. But there really isn’t enough data to say it’s smoother than other periods with no volcanic eruptions. We’ll see going forward.

Am I missing something in the back and forth between Lucia and Nathan — where the heck did 1.6C/decade and 2C/decade come from? Is the decimal misplaced, or should we be talking about a century? It’s confusing enough following the arguments without the base changing.

Lucia
“No +1 C/century is no win for deniers! But it does give us a little more time to do things. I think that matters. Maybe Nathan doesn’t. ”
Statements like this are what is used to prevent ANY action from being taken. Remember basically no action has actually been taken yet.

“I just think it’s useful to compare projections to data to see how certain we can be. The current models happen to overshoot on global surface temperature.”
Useful for whom? Do you not think the modellers are aware of the performance of their models?

“I’m not sure why you think the only goal to testing models is to do whatever you think is “less” than what has been done in the past. ”
I am suggesting that your goal is to encourage people to do less.

Darwin– sorry. Yes. It’s century. I type decade constantly and I guess Nathan does too. I just copied his quote and didn’t notice the switch when responding! (Comes from sometimes writing 0.16/decade and doing all the calcs in c/year in the first place.)

I’m sure Nathan meant /decade when he jumped in.

Nathan–
Useful for people who think it’s useful to read the IPCC reports. If the reports are useful, then comparing their projections to data is useful.

Are you suggesting that it’s better for people to be told the projections, but prevented from noticing they projections aren’t panning out? How would that advance either science of policy decisions? And do you think people aren’t going to notice if the projections are off?

Why do you think testing projections would cause people to do less that the undefined unknown amount you think they would do otherwise.

Nathan,

I think you need to retract:

“. Remember basically no action has actually been taken yet.”

The UN, US, and other PROGRESSIVE gubmints have already cost our economy Billions of dollars with this kludge with exactly 0 zero nill butkus result.

Lucia

“Are you suggesting that it’s better for people to be told the projections, but prevented from noticing they projections aren’t panning out? How would that advance either science of policy decisions? And do you think people aren’t going to notice if the projections are off? ”

No I am not suggesting people not be told.
I am suggesting that if the projections are out, by a small amount, that there is negligable difference in the path we should take.
I am suggesting that you aren’t advancing science here, that modellers most likely are aware of where their models stand (and are working to improve them).
I am suggesting you are not aiding them do that any more than I am aiding you by posting here.
I am suggesting that you are motivated to encourage people to do less, or at least promote the idea that we don’t have to do very much.

Kuhnkat

“The UN, US, and other PROGRESSIVE gubmints have already cost our economy Billions of dollars with this kludge with exactly 0 zero nill butkus result.”

That’s a joke, yes? What do you think they have spent the money on?

Lucia,

First I need to state a prejudice:

I have grave doubts about “weather noise”.

One man’s noise is another’s unattributed variance. My suspicion is that genuine weather noise (averaged globally) is rather small and rather short lived.

The contribution due to ENSO, volcanoes (which may effect the former) can be attributed. Maybe PDO can be characterised and its variance can also be attributed. If this be the case then the world’s weather noise would not, I think, reach the wild extremes of some of the model runs, and yes I do mean the individual runs.

Until we can reach out far enough into the universe to catalogue an ensemble of earths I refute the objectiveness of comparing our singular planet to an emsemble of models. Another of my prejudices.

For the record (chop my head off and hope to die) I would guess weather noise is less than 0.15K (RMS). Or rather anything remaining after all other attributable variance has been removed that is greater than this requires more scrutiny.

To put it bluntly I wonder if “weather noise” is a bit of a smoke screen. Just my prejudice you understand.

FWIW I have spent a lot of my life looking at oscillograph traces and I get a gut feel for such things. But then I get indigestion too, but rarely.

Alex

Nathan,

Stupendous amounts of [tax payers/ public] money are in fact being spent on “addressing”/ “tackling”/ “combating” something that until further tangible proof may not be a problem in any meaningful way..

Let’s recap: an increase of 0.7 C over some 150 years? A growing and increasingly incontrovertible disjunction between [presumed man-made] CO2 ppmv levels and any relevant temperature metric anywhere in the system [bar none]? 2008 global temperatures that for all intents and purposes are on par with 1988? Ocean temperatures that based on best available observable and verifiable data are lower by a factor of 2.5 than GISS models? And the list goes on.

And yet, you want me and the others participating in this thread to continue the discourse as if we are somehow facing the apocalypse and nothing is being done about it?

I think that not only our gracious host but most others participating here have their heads firmly where the sun does firmly shine. It might be useful to bear that in mind next time your denialist feelings get the better of you.

PS: As I suggested to you on a different thread here at the Blackboard: Cardinal rule: find the fatal flaw in the case and then listen very carefully how the proponents of the argument try to explain the flaws away. More often than not, they forget the basics: when you find yourself in a hole, first thing to do is to stop digging.

Nathan [10443]
You are getting rather too close to insulting your host. Get a grip!

Tetris
“Stupendous amounts of [tax payers/ public] money are in fact being spent on “addressing”/ “tackling”/ “combating” something that until further tangible proof may not be a problem in any meaningful way..”
I don’t know why you don’t think it’s a problem. Our gracious host has actually said she thinks it is.

“Let’s recap: an increase of 0.7 C over some 150 years? A growing and increasingly incontrovertible disjunction between [presumed man-made] CO2 ppmv levels and any relevant temperature metric anywhere in the system [bar none]? 2008 global temperatures that for all intents and purposes are on par with 1988? Ocean temperatures that based on best available observable and verifiable data are lower by a factor of 2.5 than GISS models? And the list goes on.”
There’s no dispute that the models could be improved. But in all seriousness, that models don’t match reality to the level you expect them to doesn’t mean there’s a problem with the theory.

“And yet, you want me and the others participating in this thread to continue the discourse as if we are somehow facing the apocalypse and nothing is being done about it?”
Never said anything apocolyptic – I have on this blog mentioned I don;t think that I personally will suffer in my lifetime (I think I have about 50 years left). And fundamentally nothing is being done. There’s been lots of talk, a lot of research, but no real action has been taken – although I think some countries may have recently passed bills to start doing something.

“I think that not only our gracious host but most others participating here have their heads firmly where the sun does firmly shine. It might be useful to bear that in mind next time your denialist feelings get the better of you.”
Don’t know what this means… Denialist feelings? Denying what?

“You are getting rather too close to insulting your host. Get a grip!”
I am confused as what she says rarely matches what she blogs about. She consistently says she acknowledges AGW and understands the need to lower CO2 emissions, but then she makes posts like this that attempt to undermine the need to lower CO2. This kind of manouever – a political manouever – is similar to what Bjorn Lomborg does. It’s not about science, it’s a political manouever.

“She consistently says she acknowledges AGW and understands the need to lower CO2 emissions, but then she makes posts like this that attempt to undermine the need to lower CO2.”

This is a common feature of the AGW social landscape. You are mistaking intellectual opposition to some AGW hypotheses – in this case the view that global temperature observations are confirming an alleged trend of 2C per century – with political opposition to taking action on emissions.

Consequently you have the typical reaction to posts which reveal some difficulty with some of the hypotheses of AGW, which is to say that these posts are political manoevres. They are not. The facts are what they are, and regardless of what you think about what we should do, we need to establish them. And yes, they will make a difference to political programs. If the warming trend were 5C per century that means something different for sensible action.

The important thing is to get the politics out of assessing what the trend actually is. Realize that it cuts both ways. If a skeptical look at the trend is political, so is an enthusiastic one. Think about the implications here.

Fred,
“Consequently you have the typical reaction to posts which reveal some difficulty with some of the hypotheses of AGW, which is to say that these posts are political manoevres. ”
No, problems with the models don’t affect the AGW hypothesis.

” You are mistaking intellectual opposition to some AGW hypotheses – in this case the view that global temperature observations are confirming an alleged trend of 2C per century ”
Fred, what were the conclusions of Lucias original post?

“If the warming trend were 5C per century that means something different for sensible action.”
Certainly but this is obviously different to 2. What difference would it make to the actions we need to take if the figure was 1.678C? Or 1.794? Or 1.463? Or 2.123C? Lucia said that if the rate was 1.6 she thought we’d have ten more years to get our new technologies sorted. What sort of difference does that make? We still have to act now. Why would we hold back because we have ten years more?

“The important thing is to get the politics out of assessing what the trend actually is. Realize that it cuts both ways. If a skeptical look at the trend is political, so is an enthusiastic one. Think about the implications here.”
Fred, Lucia consistently claims that AGW is real, and that it’s a problem. Then she makes posts like this. You can see the results in the posts above. It becomes a mechanism for people to claim that there is no need to do anything, no need to act. There’s an implicit claim we can’t trust modelling as an aid for future planning. It gets lapped up by people like Tetris and Kuhnkat who then claim that we’ve wasted taxpayer dollars etc. And that AGW is a myth. Even you have claimed that her work has shown problems with the AGW hypothesis, when it does no such thing.
Her work here looks entirely political.

Nathan,
There are transgressions of sound scientific pracices on both sides in the AGW issue. I think Lucia has a laudable concern for how scientific research is performed and presented.
As Ophelia Benson has stated for the humanities: “truth matters”.
To avoid pseudodisagreements: I am aware that most “scientific truths” are provisional, they are simply the best way of describing natural states and processes under the current level on knowledge. As our knowledge changes, so will our ways of descibing nature.

I have the impression that AGW by many today is considered a full-fledged SCIENTIFIC THEORY. Some allready consider it a theory with so much support it can be considered a “paradigm” sensu Kuhn.
Anyway, IMHO there are issues when designing good tests for AGW as a scientific theory.
Good tests for scientific theories have a number of prerequisites:
1) The predicitons should be DEDUCIBLE (from the theory)
2) The prediction should be IMPROBABLE
3) The prediciton is VERIFYABLE

The second point is occationally (often?) overlooked. It is a consequence of the philisophical problem of induction, and related, we need to avoid tests that yields positive for trivial reasons.

As far as I understand AGW has problems with both 2 and 3. In physical terms the problems are related to poor descriptions of climate sensitivity, and to factors presently not well described in the models. (spatial resolution, chaotic properties of weather, etc).

Cassanders
In Cod we trust

Cassanders,

The components of AGW are fully testable in the sense that you ask, be it the radiative properties of CO2 (and other greenhouse gases) and the source of the CO2. That should be enough, no?

To do what you demand would require us to construct a second solar system… Not likely.

“Lucia has a laudable concern for how scientific research is performed and presented.”
So is that science in general? Or just climate science?

@Nathan

—————————————————————–Beginquote

The components of AGW are fully testable in the sense that you ask, be it the radiative properties of CO2 (and other greenhouse gases) and the source of the CO2. That should be enough, no?
….
—————————————————————–Endquote

If this question really isn’t rethorical, it is quite revealing.

It obviously isn’t enough to know the radiative properties of the greenhouse gas CO2. (While the minor GG’s are fairly well accounted for, the most important, water vapor is not) http://www.realclimate.org/index.php?p=142

Just looking at graphs of temperature as function of time, emissions, or concentration of CO2 should really answer your question.

As long as the system and its components are poorly described, the ratiative imbalance isn’t well accounted for, the invocation of “chatotic weather” as an ad hoc explanation seems shoddy to me.

You have a very illuminating example of the weaknesses of the theory
(at least in it’s current state) when looking at the recent debate on the TT hotspot.
Apparently first herolded as a “fingerprint” of AGW, (i.e. following my bulletpoint 2 in my previous post) it turned out (when it failed to materialize properly) that precence or abscence of TT hotspot wasen’t an unique test after all.

Cassanders
In Cod we trust

Cassanders… Seriously this is not some big mystery. Even Lucia accepts the Greenhouse effect and AGW as given.

If you think there are weaknesses, fine.

“No +1 C/century is no win for deniers! But it does give us a little more time to do things. I think that matters. Maybe Nathan doesn’t. ”
Statements like this are what is used to prevent ANY action from being taken. Remember basically no action has actually been taken yet.
.
It is not statements like that that prevent any action from being taken .
What prevents any action from being taken is the common sense and the fact that people who decide about actions are luckily not Dr Frankensteins blindly trusting their computers .
.
Why doesn’t China care about CO2 , proves it by becoming the first world’s emitter and goes on increasing ?
Because it has to arbitrate between economical growth which brings wealth to its population and more or less 1° in hundred years .
Of course nuclear power is one leg of the solution but as they have coal , they fully intend to use it and develop it too .
You should travel more .
Last week I was in Russia and talked with some leading actors in the energy policy scene .
They don’t give a s..t about CO2 and if anything they amuse themselves by seeing (some) western governments attempting to shoot themselves in the foot by selfimposed handicaps for their industries .
One can also ask the Czech president and the majority of the population – they would certainly welcome much warmer winters coming along with a robust economical growth as a bonus .
.
The warmers fear nothing as much as having to ask the man on the street whether he prefers more taxes and an economical cisis today or 2°C more in hundred years because the answer would be clear and loud .
In Europe it becomes slowely but surely percolate in the opinion that there is a reason why the energy costs increased more than elsewhere and people begin to make the connection between Kyoto and their bills .
The laws introducing biofuels had also lead to a marked increase in fuel prices and people begin to realize that too .
When the critical mass is reached , the tar and feather businesses will see a remarkable boom
.
So I indeed agree with one thing Nathan is saying . It doesn’t make any difference if the temperatures in 100 years increase by 1 or 2°C .
In neither case there is no sane reason to spend billions for literally hot air .
While about everybody (yes even China and Russia) agrees that saving energy is intrinsically a good thing , there is only a small vocal minority , hyperconcentrated in the richest countries who considers that 2°C warmer climate in 1 century and a fortiori some CO2 concentration are a problem .

Nathan:

I am suggesting that if the projections are out, by a small amount, that there is negligable difference in the path we should take.

During the period that can be called the forecast, the projections are off by a lot. The HadCrut trend is -1.2C/cetury, the projection is over 2.0 C/century.
The hindcast period when modelers could tweak forcings used to drive their models matches.

I am suggesting that you aren’t advancing science here, that modellers most likely are aware of where their models stand (and are working to improve them).

If the modelers step up and tell us where they stand, I’d believe that. Those modelers who blog say they models are projecting just fine. The case made is implausible.

I am suggesting you are not aiding them do that any more than I am aiding you by posting here.

I’m comparing models to data. To the extent that people were interested in the projections themselves, some are equally interested in knowing how the recent cold spell relates to the model projections.

If you don’t like that people– including me– are curious, then too bad for you.

I am suggesting that you are motivated to encourage people to do less, or at least promote the idea that we don’t have to do very much.

I don’t have the slightest notion why you believe this. I think we should reduce the amount of fossil fuels we use.

In contrast, none of us have the slightests notion what you want.

And, I might ask, if you think my being honest about how what the divergence between the data and projections means suggests about wanting to do more or less, what would your appearing to want me to shut up about it suggest about your motives? Might you wish to achieve your goals by misleading people about the truth? Or letting others do the misleading while you, knowing their story is incorrect, sit around quietly saying nothing?

Because I think the fact that many deniers and skeptics believe some AGW advocates wish to mislead to get their ways that causes them to support policies to do anything. For that reason, I think honestly comparing data to projections is the way to motivate people to do more.

I recognize you may note believe what I do would motivate people to do more. But..there you go.

Now, while were are at it– Can I assume you downloaded the temperature data and checked whether the trends regularly exceed 1.6C/century? (As you suggested I do based on your singularly odd notion that they might have?) Or, having utterly failed in figuring out anyway to suggest the technical content that what I do is wrong, have you just decided to argue to motives?

Alexander

I have grave doubts about “weather noise”

Lots of us here hate the term “noise” for random seeming variations we experience as weather. It seems to have been introduced by AGW activist climate bloggers, and it’s a bad term. That said, weather does show some sort of random or chaotic state, and we need a word for that. Noise seems to have been picked.

I’ve posted a lot about some bloggers trying to explain aways the recent flat discrepancy between observations and temperature by insisting we estimate the likely amount of “weather noise” since 2001 by including the temperature excursions after a series of three very large volcanic eruptions including Pinatubo.

As for comparing to the models: I do so. But, I bear in mind that the purpose of the models is to predict the earth’s behavior. It is not the purpose of the earth to adapt it’s temperature to comform with models. Models can be wrong. Measurements can be wrong. The earth itself is not cannot be wrong.

Nathan,

Lucia’s already responded to your post and hardly needs my ’support’, but I think you’re very wrong to question either the motives or the outcomes of her interest. There are other blogs which line up on both sides of this debate where such questions might be fair, but it is my view that Lucia can deservedly be termed a sceptic, a word that is otherwise much maligned by many who have appropriated it to describe sets of attitudes that are not sceptical at all. You should note, I think, that so-called ’sceptic’ claims are treated here with scepticism as well as, for example, IPCC model projections.

I’ve argued above that I think the danger of doing too little in terms of mitigation is greater than that of doing too much, so to that extent I would agree with you. However, I think we should be hoping for increasingly robust modelling as time goes forward, so to the extent that an examination of current modelling projections contributes to that then it should be welcomed.

It seems to me that Lucia is a very careful and sincere seeker of truth. The fact that she sometimes infuriates both the proponents and the contrarians, is proof enough for me that she has no dog in the fight. This is what keeps people coming back.
Thanks Lucia,
Mike Bryant

As far as Michael’s testimony goes: it is this comment which I find most egregious: “The ensemble model failure is not a product of the selection of recent years; rather it is a systematic failure of the models as a whole to accommodate temperatures in recent decades.” His entire testimony seems to be set up to make it look like an entire two decades worth of temperatures are outside the bounds of the models, when really only 2008 itself is an anomalous year.

Also: given the range (about -0.4 to +0.8) of 95% bounds at the 5 year point, it seems strange that the 95% bounds narrow so quickly as one moves to the 15 year point. For example, just assuming that any given year is plus or minus 0.3 degrees off of a permanent 0.2 degrees/decade mean trend, I duplicate the -0.4 to +0.8 bounds at the 5 year point, but get bounds of 0 to +0.4 degrees at the 15 year point (unlike Michaels who seems to get about .15 to .25 by eyeballing). I would think that if anything, auto regressivity (not that I have figured out how to work quantitatively with AR(1) trends yet) would make the bounds decrease more slowly with longer term trends rather than more quickly. The Michaels bounds seem to be implying that near term temperatures actually have a kind of anti-correlation with each other… I guess, one could argue that a La Nina year in year 1 might decrease the chances of a La Nina year in year 2 or 3, and increase the chances of an el Nino, thereby increasing the noise at recent years but not for distant years, but at the 5+ year timespan, I’m a little dubious…

ps. For those people complaining about ENSO not being “noise”: I think for these purposes, “noise” is anything that the models aren’t expected to have line up with reality (or with each other). With the exception of the Keenlyside paper (which has its own problems) models are expected to have uncorrelated ENSO patterns. I agree that, rather than using the full “noise” estimates, a preferred methodology for trying to match model with reality would use real-world ENSO to correct the real-world trend, and model ENSO to correct all the model trends, and add volcanoes into the models so you eliminate that, and then your “noise” would be much smaller. Obviously, because we’ve been in a strong la Nina, that would actually correct the current temperature significantly upwards, so I’d think that might make 2008 look closer to model trends than it does with the current methodology…

Thanks, Lucia,

In my mind climate models provide plenty of “synthetic” data!

Here are a few more details:

Pat’s Figure 1 was to establish that the model behavior during the period 2001-2020 is pretty consistent—i.e. we shouldn’t have non-stationarity problems

For the period 2001-2020, we calculated all possible (successive) 8 years trends (to use your example) for all the individual A1B model runs I got from ClimateExplorer (there are a few more since I last checked than what I originally downloaded, so I’m working on doing it all again). For each collection of 8-yr trends for each model run, we determined the 2.5 and 97.5 percentile values. We then averaged these values together first for each model (with more than 1 run available—so as not to bias the results towards the models with the most runs) and then across all the models. That “model average” 95% percentile range is what is depicted in Figure 3, along with the current 8-yr trend from the HadCRUT3v data ending in December 2008.

I think in you Figure 4 you may have dumped all 8-yr trends from all model runs into one set and then determined the 2.5 and 97.5 percentile values from that. To me, this allows a few extreme models to dominate the tails of the distribution rather than producing more of an “ensemble range” that we attempted to produce.

You are correct in your assessment that we were trying to answer the question “Is the current observed trend as a single realization of weather outside the range of all possible weather in all possible models?” and I would add, over a variety of timescales (at least from 5 to 15 or so years).

If not, then perhaps some model tweaking is in order…

-Chip

In all these discussions there appears to be a consensus among the commentators that a 2C/century increase is something to worry about rather than to be happy about. I think more (and less biased) research is warranted to establish the pros as opposed to current focus on the – in my opinion less worrisome- cons. Surely, everybody agrees that a 2C/century drop would really be worrisome.
I am now expecting the standard alarmist remarks that rising sea levels will wipe out coastal inhabitation. Considering past climate optima I do not see Greenland nor Antarctica losing the amount of ice that would cause an alarming rise in sea level.

Simon is a typical example of someone who takes it as a given that a temperature increase is something negative when he argues: “It hardly seems a really bright idea to me to go on increasing GHG concentrations. I would rather take the risk of being wrong on the cautious side than chancing being right on the reckless side. I don’t take comfort from uncertainty.”

Chip

Pat’s Figure 1 was to establish that the model behavior during the period 2001-2020 is pretty consistent—i.e. we shouldn’t have non-stationarity problems

I pretty much agree. The trend is linear enough.

We then averaged these values together first for each model (with more than 1 run available—so as not to bias the results towards the models with the most runs) and then across all the models….

I think in you Figure 4 you may have dumped all 8-yr trends from all model runs into one set and then determined the 2.5 and 97.5 percentile values from that.

That could make the difference. You are correct: I didn’t reweight. Re-weighting is perfectly justifiable (and in fact better). I figure: When guessing a method, go for simple first. My answers aren’t that different from yours– and I think what you do is probably better than what I did.

If not, then perhaps some model tweaking is in order… Yep.

I think both your decision to a) reweight and b) average over time is a step forward compared to what Gavin did. Of course, someone might find some reason to dislike the averaging over time. I can’t think of any reason not to reweight.

Marcus
You criticize this:

As far as Michael’s testimony goes: it is this comment which I find most egregious: “The ensemble model failure is not a product of the selection of recent years; rather it is a systematic failure of the models as a whole to accommodate temperatures in recent decades.” His entire testimony seems to be set up to make it look like an entire two decades worth of temperatures are outside the bounds of the models, when really only 2008 itself is an anomalous year.

Michaels is discussing trends. His point is: it doesn’t matter what year we pick as the first year for trend analysis. The models always look bad. This is a response to people who claim that people showing models are off are relying on a) short periods of time or b) cherry picking the start year.

Right now, even if you pick a long time or try to cherry pick, the models still look bad.

However, if you want to pick a statistical method with the least possible power, it’s possible to think otherwise. When using the anomaly method, and setting the baselines to match during a recent period, it takes a very large amount of data to detect a mis-match in trends.

It is a principle in testing that one when one can chose between two statistical methods, you pick the one with greater power. This results in the minimum rate of error when doing hypothesis tests. Testing trends has greater power. Eyeballing the graphs for pop-outs has less. So, Michaels is using a suitable test and saying something true.

Lucia,

Thanks again for your comments.

As far as averaging over time goes, in my mind, it is the way to go, for you basically increase your sample size. Since any moment in time (start date) is just as good as any other moment in time as far as the models go, you may as well use *all* moments in time (start dates) that are available. Your sample size increases from something like 192 (24 model runs of 8 years (96 months) each) to something like 3456 (24 models and 144 96-month periods). I understand that with autocorrelation the degrees of freedom don’t increase from 192 all the way to 3456, but they do increase beyond 192…which makes for a more robust result.

Also, as you can see from comparing the model range in Pat’s Figure 3 (which is calculated using model projections from 2001-2020) with that in Pat’s Figure 5 (which is calculated using model projections from 2001-2030), the impact of adding the decade of the 2020s is to increase the high end of the range of trends, and has little effect on the low end. That is because the greenhouse forcing is slowly increasing in A1B over the first half of the 21st century, so slowly starts to force the rate of change upwards and lessening the chance of lower trends. So, I think our lower bounds are going to be pretty robust to choices concerning our time averaging decisions—especially since we don’t assume normality (i.e., use some multiple of the standard deviation to define our confidence range).

-Chip

Chris Schoneveld (Comment#10465)

Well, since you mention my name, Chris, I’ll give a response.

I agree with you that a 2C global drop in temperature would also have severe consequences. I ‘take it as a given’ that any such scale of temperature change over such a time period is a severe threat.

As for the ‘pros versus the cons’…..

“Globally, the potential for food production is projected to increase with increases in local average temperature over a range of 1-3°C, but above this it is projected to decrease. * D [5.4, 5.6]“ (AR4, WGII, SPM).

So there we are, the IPCC tells us that for a start. That sounds like good news for a while? The trouble is that even that apparent ‘pro’ conceals the devil in the detail. The issue is one of where food production potential will increase against where the opposite will prevail, of where regionally short-term climatic change may be deemed ‘beneficial’ against where it is damaging. It is not wise, IMV, to consider average effects, since strategic implications are not ‘averaged’. What is the advantage to, perhaps, thinly populated areas of North Central Asia and Siberia (with all their oil, gas and minerals) ‘improving’ climatically if, at the same time, some densely populated areas of China are disadvantaged? I don’t like the prospects of two destabilised nuclear powers negotiating over that situation.

As for sea level rise, I don’t know what level you would consider alarming. Many millions more people are expected to be flooded by the 2080s even on the basis of the IPCC’s cautious projection of sea level rise excluding ice sheet dynamics. These people would mostly be in the mega-deltas of Asia and Africa. They seem a long way away, perhaps, but strategic consequences have a way of coming back home to us.

So, it seems to me that regardless of the ‘net’ assessment of climatic pros and cons, and regardless of the longer-term threat to ice sheets and so on, we must anticipate a world in which human societies are radically disrupted. It’s speculation, of course, to say what that might lead to, but I don’t think it probable that it will be good for the global economy!

Simon–
With respect to Michaels testimony, remember he is appearing before congress. The climate model data is used to drive the economic impact models. Those do include projections for regional impacts, and the range of uncertainty about that.

Reading Michaels full paper, if the models are off, your economic assessments are off. If the average model projects too high and/or that the world arrives at this level more quickly that it actually will, and we use those to estimate costs, we will expect the costly impacts sooner rather than later.

This is a straightforward argument.

It doesn’t exclude your argument about uncertainty or the need to consider what is required if the higher end of projections occurs. It doesn’t exclude the need to consider that local effects may be more severe than the global average. (In fact, it’s certain that some locales will be more severely impacted, and some less.)

If our ability to predict is poor, that will argue for certain decisions as well. I like investments in nuclear because the benefit of nuclear is robust to the range of uncertainty in our ability to predict trends. If the trend ends up low: We have baseload capacity. If the trend ends up high: We have baseload capacity and possibly sufficient amount to spend some energy on recapturing CO2, desalinating water or maybe even pumping water to places that used to rely on snow melt for water. Conservation alone will not give us the buffer if the higher end of warming occurs.

Lucia, I agree with most of that, including favouring nuclear.

Returning to these pesky graphs, how is it that the AR4 models for A1B are deemed to be effectively linear in trend (+0.2/decade) over the period when the TAR stated +0.16C for 1991-2000 and +0.14C for 2001-2010? Have the models changed that much, or what?

Simon,
The AR4 models do predict higher warming than the TAR models. Differences bewteen TAR and AR4 models are discussed in various places in the AR4. For example. in page 755 of the AR4, the authors discuss the changes in treatment of aerosols, differences in outcomes for shortwave forcings are discussed in table 10.2 etc.

To find all the differences, you can search for TAR using the search tool in adobe acrobat. The A2 scenarios is evidently the TAR scenario. So, the difference in warming expected for the period I call “now” doesn’t seem to be strongly influenced by the scenario.

Hmm, I wander away for a few days and miss a neat discussion.

I do have to take an exception with some of tetris’s statements earlier in the thread:

“an increase of 0.7 C over some 150 years”

Is somewhat of a misleading statement (in part because its 0.7 C over 100 years, temps didn’t change much at all from 1850 to 1910 using HadCRU). Furthermore, 0.5 of that 0.7 degrees warming occured since 1970.

“A growing and increasingly incontrovertible disjunction between [presumed man-made] CO2 ppmv levels and any relevant temperature metric anywhere in the system [bar none]?”

Now this is just sillyness. Lets nip this “temperatures are going down while CO2 is going up” nonsense in the bud, shall we? We know that CO2 increases pretty monotonically (at least once you remove the annual cyclical variation due to seasonal effects on hemispheric plant photosynthesis). Temperature, as we are well aware, does not increase monotonically (due to ENSOs, volcanoes, potentially solar factors, and who knows what else!). So we would not expect temperature to be in lock-step with CO2, just that the trend in CO2 and the trend in temperature would be related. The easiest way to see this is via a simple linear regression of temperature and CO2. Since the only monthly CO2 data I have handy (ftp://ftp.cmdl.noaa.gov/ccg/co....._mm_gl.txt) starts in January 1980, I’ll use that as the starting point for my regression. We could start earlier if the data is available, but frankly it should not matter that much. Lets regress monthly CO2 concentrations (adjusted to remove the annual cycle) with temperature:

Given the near-monotonic increase in CO2, this ends up looking quite a bit like the temperature record. There is a noticeable dip at the end (and a spike a bit earlier representing the 1998 ENSO), but is it significant, or is it well within the bounds of the “noisy” relationship between CO2 and temps?

To figure this out, we can plot the residuals between the observed temperature and the expected temperature associated with different levels of CO2. We can see how these residuals compare to the 95% confidence interval (e.g. two standard deviations) around the trend.

Here we see that while there were a few months (Jan 2008 in particular) that fall below the 95% confidence interval, its still well within the variability of the trend. Since we are plotting monthly data, we would expect about one month to fall outside the confidence interval every year and a half or so. To make things a bit more clear, I plotted a 12-month running average of the residuals (which, I believe, we would expect to exceed the 95% confidence interval only once every 20 years or so), and it is still within two standard deviations of the trend.

As far as your argument that 1988 was as warm as 2008, well, thats only true using UAH and you and I well know that a single relatively hot year and a single relatively cool year do not make a trend. Since temperature measurement satellites were first launched in 1979, the decadal trends are as follows:

UAH: 0.126801657
RSS: 0.156521702
GISS: 0.161862669
HadCRUt: 0.158847476

Anyhow, my apologies Lucia for going a bit offtrack with this, but I enjoy fighting with strawmen at times.

Simon Evans (Comment#10471),

Both deltas (e.g. Bengal) and atolls (e.g. Maldives, Tuvalu etc) have managed to survive the thousands of years of post glacial sea level rise thanks to very process that created those topographical features.
Deltas grow as a result of sediment input from the rivers that feed them. There is apart from the seawards outbuilding of the delta also subsidence due to increased sediment load. Hence, deltas are typically governed by a balancing act between sea level fluctuations, sediment input and crustal subsidence. To just pick one spurious element (human induced sea level rise) you are simplifying the issue.

Although you didn’t mention the threat to coral islands (but MSM do all the time), coral reefs grow upwards and thus could keep pace with sea level rises of 10 mm/year with ease. If only the inhabitants would leave the corals in pristine condition and allow natural processes to prevail (storms and occasional flooding) those islands would have nothing to fear from sea level rise, on the contrary.

Lucia,

Any idea which period I should choose from which to have the anomalies calculated in reference to on the ClimateExplorer site?

Choosing different periods yields differnt annual cycles and thus different monthly anomaly values. Since I am computing (relatively) short-period trends from monthly data, they are slightly sensitive to this choice.

In the name of reproducibility, I’d like to choose an appropriate period and stick with it. Since I am only dealing with model projections (not the 20th century runs), does the period 2001-2020 seem reasonable? 2001-2030? Or perhaps the entire 2001-2100 period? I think by default, ClimateExplorer uses the whole period of record to derive the annual cycle. But, the whole period of record is different from run to run. I think I chose 1971-2000 in my previous analysis, but I notice that some A1B runs on ClimateExplorer don’t include the 20th century runs and thus the 1971-2000 period isn’t available. This is what is leading me to the 2001-2020 choice (which means I’ve got to go get everything again…bummer).

-Chip

Chip–
I’ve communicated with Geert Jan. I would advise computing the anomalies yourself. The reason is that his script uses the mean for all runs from a model. That’s fine and great for an online tool that display results, but it has a features you may not like:
The anomaly for models is defined differently than for the data. (The data has only 1 realization.) To make them the same, I rebaseline each run using data from that run only.

Otherthan that, for the purposes of evaluating projections, I use Jan 1980-Dec. 1999 because this is the baseline period discussed in Chapter 10 or the AR4. You’ll find the discussion right after figure 10.4. For models, I compute the monthly anomaly for each month and subtract.

I could do the same for the data, but I like to do the minimum fiddleing for that (it’s already anomalies after all!). So, for the data, I compute the average from Jan 1980-Dec. 1999 and just subtract. That way I know I don’t change any of the properties of it’s “weather noise”.

Other methods may be ok too. But this is the one that made sense to me.

Zeke:

You note that:

” Furthermore, 0.5 of that 0.7 degrees warming occurred since 1970.”

The CO2 concentration in 1970 had reached about 325 PPM, versus (probably) about 285 PPM in 1900, while it is currently at ~386 PPM. Assuming the response to CO2 is a log function of concentration, which I think most everyone assumes, we would expect the temperature rise by 1970 to have been about 46% (Ln(325/285)/Ln(386/285)) of the total rise to 2009, if CO2 were the principle cause for temperature rise during the last 109 years. Since only about 0.2/0.7 = 28.6% of the temperature rise actually took place by 1970, this suggests (at least to me) that there are other important factors involved in the temperature rise since 1900, and/or that there are significant long-term underlying temperature variations (multi-decade Atlantic Oscillation, PDO, etc.) which make a simple prediction of temperature rise based on CO2 concentration doubtful over periods shorter than the underlying variations. The substantial temperature rise between ~1915 and ~1945 (with little concurrent rise in CO2 concentration) further suggests other significant factors are involved.

In any case, I think that the correlation between CO2 concentration and temperature increase from 1980 to 2009 almost certainly does does not provide a complete and accurate picture of the connection between atmospheric CO2 and average global temperature. Which is the same problem that the climate models seem to be running into.

SteveF,

Well, its not like the climate system is instantaneously forced. Hence the common arguement that we have 0.6 degrees “in the pipeline” given the earth’s thermal inertia and current GHG concentrations using the function:
.
ΔF = 5.35ln(C/C0)
ΔT≈λΔF
where λ is the mean sensitivity for the doubling of CO2: 0.87 K
.
I do agree that CO2 is not the only factor involved in early 20th century warming. That said, I think you’d be hard pressed to credibly argue at this point that anything else is responsible for the bulk of warming from the middle of the century to present. My main point, however, was just that the current short-term divergence between temperature and CO2 is not particularly unusual in light of the past record.

As an addendum to the above, its important to remember that CO2 concentrations were smaller, but far from negligible, in the earlier part of the century:

Steve F [10493] and Zeke H [10480]

Steve: Thx for your commentary. Succinctly put. Any correlation is highly questionable, and arguing that there is a statistically relevant causal relationship between the two is nonsense.

Zeke: No straw man here. To suggest as you do that temperatures flat lined between the 1850 and 1970s and then somehow took off sounds like a mini version of Mann’s hockey shtick and the 1970 starting point for the 0.5C increase since then is just a tad too convenient. Records show that after the 1920-1945 warm period the 1960s through mid-1970s saw some of the lowest temperatures of the entire 20th century and led to the “Global Cooling” scare of the late 1970s. Anecdotal but real nevertheless, in the mid 1960s the English Channel froze over several times and the Dover-Calais ferries got stuck… Now that’s cold.

Are you seriously disputing that there has been no warming for the better part of the past decade? Because if you are, you are in a rapidly dwindling minority. Not even UK Met Office Hadley disputes that anymore.
And yes trends are trends: it all depends on the period you cover. Break out the 1998-2008 period from any of the ones you cite and the trend changes rather dramatically, doesn’t? And yes, single years are single years they do not trends make. That said, during the 12 month period 01/2007 – 01/2008 global temperatures dropped by a whopping 0.6C, more than the entire gain you argue since 1970.

tetris,

I’m fine explaining temperatures of the 20th century using solar, vulcanism, aerosols, and CO2 as major forcings. Though, as I mentioned, there is considerably more uncertainty re: forcings in the early part of the 20th century.

The forcings chart I was trying to link above didn’t work, so here is another go:

http://i81.photobucket.com/alb.....ture23.png

I’m not sure where you got the impression that I was disputing the fact that temperatures have stagnated over the past decade. I was simply pointing out that comparing 2008 temperature (a cold year) to 1988 temperatures (a warm year) is rather meaningless, and that a look at the long term trend is more useful.

That said, all four temperature records do show warming over the past 10 years now. This wasn’t true if you started at the height of the 1998 El Nino, but now that a decade ago is post-El Nino you have to start in 2002 to get a negative trend. Which just goes to show that the shorter the period of time in question is, the more the trend you get depends on the starting point.

From December 1998 to December 2008 the decadal trend is:

UAH: 0.1139
RSS: 0.0989
GISS: 0.1780
HadCRUt: 0.0643

I made a neat graph awhile back that shows the trend from that month to present for every month since 1979:

http://i81.photobucket.com/alb.....cture8.png

Tetris, by the same whopping logic, RSS shows 0.388C increase from Jan-2008 to Jan 2009. Why do decreases get all the amplification and not the increases? What’s left of the “more than the entire gain” point including the last 12 months?

Lucia,

Thanks, but I am not sure I fully followed your methodology.

I am only interested in model projections. Geert Jan told me that they begin in January 2001 (the data included in the A1B option at ClimateExplorer for periods prior to 2001 then must be from the 20th century runs which apparently Geert Jan appended to the A1B results).

To calculate monthly anomalies, I’ll download the monthly model temperatures for each run, create a monthly climatology over some period of time (say 2001-2020), and subtract the climatological means from the raw monthly temperatures. I’ll do this on a run-by-run basis since you point out that Geert Jan’s routine is done on a model by model basis, rather than a run-by-run basis.

That seems reasonable. Is this what you were suggesting as well?

What am I to make of the IPCC statement (Figure 10.4 caption), “Linear trends from the corresponding control runs have been removed from these timeseries”? It this something that I need to do after downloading the ClimateExplorer data, or has this already been handled?

Thanks,

-Chip

Sekerob [10501]

Your example still leaves a residual loss of 0.21C since 01/2007 unaccounted for. 12 month variations aside, overall there has been no increase in global temperatures for the better part of a decade now while CO2 levels have continued to rise. Let’s keep that in perspective: that is 1/3 of the period for which we have satellite data. Not something that we can simply walk past and pretend its not there.

Inconvenient truth for some, but a matter of record nevertheless. And let’s also keep in mind that AGW is not somehow “hiding” in the oceans as some have tried to imply [Argos satellite data shows ocean cooling at all depths] and is not somehow masked by a La Nina [as famously suggested by Phil Jones of the Hadley/East Anglia Climactic Centre] simply because La Ninas and El Ninos are integral parts of the atmospheric system. There simply is no more statistically significant warming.

Zeke:

CO2 is (of course) applying net positive radiative forcing to Earth’s climate, but I do not believe existing data support the net forcing values assumed by most climate models, nor the predicted temperature rises from those models. Both Lucia and Michaels are just comparing the models to reality to show that they are not working very well, and need to be improved.

The “0.6C in the pipeline” was first suggested by Hansen et. al., in 2003 (I think). But this is a very doubtful number, since recent Argo data (Willis et. al., 2008, one of Hansen’s 2003 co-authors, as well as others) show that the seasonally adjusted ocean heat content in the first 750 meters has been almost constant since 2003, if not falling slightly. During the same period (2003 to present) the various measures of average surface temperature have also been close to constant. Hansen et. al. used “0.6C in the pipeline” to explain why the measured surface temperature rise (as of 2003) was far lower than the expected temperature rise based on the radiative forcing of CO2 assumed in their climate model (close to or the same as the 5.35ln(C/Co) net forcing you note).

Were there really so much warming “in the pipeline”, we could reasonably expect the ocean heat content to continue rising when the average surface temperature is nearly constant, since “0.6C in the pipline” implies that the ocean is far from equilibrium with the surface temperature (that is, 0.6C cooler than equilibrium). The fact that the ocean heat content in the first 750 meters stopped rising promptly when the surface temperature did the same simultaneously casts doubt on 1) the assumed long ocean lag period, 2) the “0.6C in the pipeline” figure, and 3) the assumed forcing constant of 5.35. While there is no doubt some loss heat from the top 750 meters to the deep ocean, all published estimates of this loss are quite small compared to the “0.6C in the pipeline” value, so heat loss to the deep ocean does not explain the sudden stop in ocean heat accumulation.

I think the Argo data collected over the next several years will be very helpful in evaluating the true ocean lag, as well as the validity of the forcing constants used in the climate models. Should it turn out that the ocean lag is in fact much less than than the 0.6C assumed by Hansen et. al., as data from the last 5 years appear to suggest, then the net forcing constant for CO2 should be revised downward to the range of 3.4 – 3.5 to make the models fit the data.

Zeke Hausfather [10500]

Thx for your rejoinder. Your second graph is certainly enlightening. Would you care to comment some more on its implications?

I am puzzled by your use of the term “stagnate” as it implies that you are convinced that the warming will by definition pick up again sometime in the future. Given that I live by the sea, I guess I’m more comfortable with the notion of “rising” and “falling” [be it tides or global temperatures..]. Perish the thought of a stagnating tide..

tetris (Comment#10506) February 17th, 2009 at 3:14 pm

No, you’re brushing aside. You were creating a highly distortive picture. When looking through RSS & UAH, there’s nothing at all whopping. Was there by any change a El Nina in the neighbourhood going from 2007 to 2008?

tetris,

Given that last year was an ENSO in an unusual solar minimum, I do expect that temperatures will increase in the next few years to be consistent with the long-term trend. The main question I have is what climate sensitivity will turn out to be, whether we will end up with 1.5 or 3.5 degrees of warming for a doubling of CO2. Its still very much an open question, and an interesting scientific question moving forward. I don’t think I will ever be able to see eye to eye on this, however, so we might just have to agree to disagree :-p

As far as the above chart goes, it was more to demonstrate that the variability of the trend increases dramatically as the number of months included is reduced, and the trend obtained depends strongly on the starting point for any year after 1994 or so. I cut it off at 2004 because it goes rather crazy after that, but I’ll include the full period if you are interested. Note that the y-axis on this graph is an order of magnitude larger than on the prior version–its in degrees C per decade rather than century!

http://i81.photobucket.com/alb.....ture25.png
.
SteveF,
I agree that it will be interesting to follow the Argo results, though I think its a tad premature to draw that many conclusions based on it. Lucia and I were discussing this awhile back on another thread, and a lower ocean head update / mixing rate vis-a-vis what is currently modeled would imply either a lower climate sensitivity or a higher negative aerosol forcing (in the wide uncertainty range of possible forcings) which, incidentally, would mean a higher climate sensitivity! Regardless of which it turns out to be, that climate sensitivity is lower than expected, that aerosol forcing is strong than expected, or that the Argo data is inconclusive (and I definitely hope for the former), it will be interesting to watch and analyze the developments. Part of the fun of climate science is the enormous complexity of the Earth’s system, which is all to often something that both sides of the debate ignore or trivialize in order to score easy talking points.

Zeke [10513]
As I pointed out elsewhere above, La Ninas can not mask warming. La Ninas and El Ninos are an integral part of the system. And if the sun’s current behaviour persists and SC 24 is in fact the advent of a new minimum – and there are some pretty knowledgeable scientists in that field who are arguing that this is what is happening – we could on the contrary be in for a considerably cooler environment over the next decades than we have now.
Maybe you’re right: agree to disagree. No harm in that.

Zeke:

I have seen no published data about a substantial rise in atmospheric aerosols starting in 2003, and I think it a stretch to suggest that this might explain the sudden stop in ocean heat accumulation. If the long ocean lags consistently pointed to by climate modelers (eg. 0.6C in the pipeline) are correct, then any source of cooling would have to be large enough to cause a sudden drop of ~0.6C in average surface temperature in order establish equilibrium with the oceans and stop the accumulation of heat. Since there certainly has not been a large drop in surface temperatures, the most reasonable explanations are that the assumed long ocean lag is wrong, and that the forcing constants in the models are simply too large.

The ocean heat data appear to be in clear conflict with the models. Time will tell, but I expect the data will ultimately prevail over the models.

SteveF,

I wasn’t arguing that it required a rise in aerosols (on the contrary, aerosol emissions have declined slightly over the last decade, though the data is somewhat sketchy). Rather, there is a wide range of possible aerosol forcing (for a given quantity of aerosols), and for a given climate sensativity, a lower ocean mixing rate would imply higher aerosol forcings. Hansen discusses it in a recent lecture at the AGU here: http://www.columbia.edu/%7Ejeh.....081217.ppt

Zeke:

I had seen Hansen’ presentation before, and it contained absolutely nothing new, just the same unsupported claims of worldwide doom if we don’t return to 350 PPM CO2 immediately.
.
At no point in the presentation does Hansen address the obvious problem of a sudden stop in ocean heat accumulation versus the postulated “0.6 degrees in the pipeline” from his 2003 paper. He says only that ocean heat accumulation is a climate issue, and that “Climate is a research problem”. (An odd comment I think from someone who claims climate science is settled.) I agree that more and better quality research is needed, especially with regard to testing of climate models against reality, but pronouncements of having “nailed” the climate’s behavior with his climate model is nothing more than pure puffery (or maybe self delusion).
.
At least the photos of his grand kids were nice.

Lucia,
The reason I keep banging on about this is that what you keep saying and what you blog about are inconsistent.
You have said many times that you think AGW is real, and a problem.
Then you say things like this:
“During the period that can be called the forecast, the projections are off by a lot. The HadCrut trend is -1.2C/cetury”
Surely you are not suggesting this will continue? Surely if AGW is real, then you’d exepct this to change.

It’s fine that you don’t like the models, or think they are poor, or whatever.
“If the modelers step up and tell us where they stand, I’d believe that. Those modelers who blog say they models are projecting just fine. The case made is implausible.”
Do you seriously think that they aren’t working on improving the models? What does “tell us where they stand” mean anyway? If they are telling you they are projecting fine, isn’t that telling you where they stand? Do you know why they think their models are projecting fine?

“I’m comparing models to data. To the extent that people were interested in the projections themselves, some are equally interested in knowing how the recent cold spell relates to the model projections.
If you don’t like that people– including me– are curious, then too bad for you.”
Curiosity is a wonderful thing, and it should be supported. However, you were claiming that your analysis was “useful”. I was asking for whom it was useful.

“In contrast, none of us have the slightests notion what you want.
And, I might ask, if you think my being honest about how what the divergence between the data and projections means suggests about wanting to do more or less, what would your appearing to want me to shut up about it suggest about your motives? Might you wish to achieve your goals by misleading people about the truth? ”

Well, I want funding for alternative energies. I want the Govt to subsidise people to put solar panels on their roof. I want more bicycle paths, so people can cycle to work. I want to remove Govt subsidies to coal mining companies and petroleum companies. I want to reduce fossil fuel usage.

I don’t want you to shut up, I want you to be clear. You keep claiming that you fully acknowledge AGW, yet your posts continuously focus on the doubts. You claim that you want to reduce fossil fuel usage yet your posts present data that suggests we don’t need to. Why do you only focus on the doubts? This is where motivation is important, because you don’t provide any positive commentary on pro-AGW science. This suggests you have an agenda, especially when you claim to support the AGW hypothesis and that there is a need to act.

“Now, while were are at it– Can I assume you downloaded the temperature data and checked whether the trends regularly exceed 1.6C/century? (As you suggested I do based on your singularly odd notion that they might have?) Or, having utterly failed in figuring out anyway to suggest the technical content that what I do is wrong, have you just decided to argue to motives?”
No I haven’t done it yet. However, if your posts were clear your motive would be unimportant factor. If you really think that AGW is real and that there is a need to do something about it, why are you writing a post about the expectations that global warming is slowing and will continue to slow? Does this not strike you as being exclusive?

I retract most of my earlier critique of Michaels: I was not properly duplicated his trend analysis. Mind you, now that I have created the right Excel spreadsheet, the full line is still sensitive to the temperature in 2008, but 2008 isn’t the cherry-pick I thought it was. Of course, extending the data to 20 or 25 years makes it look better, as does using GISS instead of HadCrut. Also, the rate at which his temperature bounds drop makes more sense to me, though I think I’d like to set up a quick monte-carlo type analysis with “dumb noise” to see if I can duplicate it…

SteveF and tetris: At least on a superficial level it is possible to have heat move around within the system leading to fairly large but transient anomalies in a manner very different from a long term forcing. La Nina and El Nino are the obvious examples: for example, a drop in trade winds during an El Nino stops the cold upwelling off the South American coast, leaving increased surface ocean temperatures across the Pacific, which, unsurprisingly, lead to warmer global average surface temperatures (not taking into account weather patterns that propagate to the rest of the globe). If we could accurately measure the entire system, presumably there would not be a large change in net heat: the increase in surface temperature is compensated with a decrease in deep ocean temperatures. There might be some small net effects on total heat in the system: increased surface temperature would lead to increased radiation, for example, meaning that, counterintuitively, an El Nino probably decreases the planet’s radiative imbalance (if there is no imbalance to begin with, the El Nino could lead to cooling of the overall system). Changes in cloud patterns of course will also change radiative forcing, but I don’t know the expected direction of that.

But we can’t measure the system that well – the ocean is especially tricky due to its huge volume, high heat capacity, 3D nature, etc.. Look at the XBT/Argos data issues. Speaking of which, Tetris you may want to check your data: I believe that the Argos system is based on buoys and floats and not satellites, and additionally has been updated and now shows warming, consistent with both models and satellite sea level rise data. As long as sea level is rising faster than can be explained by glacial melt and Antarctic/Greenland mass loss (which it is), we know that heat is going into the oceans. (of course, satellite sea level rise monitoring can also be subject to correction).

And, to go along with the Michaels trend analysis: I did a quick calculation along his lines (now that I understand what he did) on the multivariate ENSO index and found that, using 2008 as and end date, the MEI trends are negative for every starting year between 1975 and 2007 except for 1999.

Now, if that trend is a _result_ of increased GHG or other external forcing, then it is a fair criticism of the models if they don’t capture that. But if the trend is mostly independent of external forces, then this trend _is_ masking warming in the same way that a volcano would mask warming (or a hot El Nino would exaggerate the underlying warming). Doesn’t it strike all the skeptics as a little odd that we have a strong la Nina and a “weak sun”, and 2008 is _still_ a pretty warm year? What does this imply for the next time we have an el Nino and an active sun?

Nathan [10528]
Using your own terminology, why don’t you please stop “banging on about this”?

Tetris–
I read Nathan’s comment. When I did, I realized that most of what Nathan considers to be my message or lack of clarity stems from his very his odd interpretations of statements.

If I correct his mistatement of facts about the recent 8 year trend, he jumps to the conclusion I am suggesting recent 8 year trend will persist. If I say the AR4 projections are inconsistent with recent observations, he jumps to the conclusion that I have said warming is slowing. (Possibly, he doesn’t know the AR4 projections represent a sustained warming trend that exceeds anything we have seen in the past. So, saying their wrong may only mean they over predict the level of acceleration– a fact I’ve stated quite often.)

He confuses my reporting the outcome of comparisons between data and observations with my wanting a certain outcome to support some agenda. He thinks my saying models over project, but the data does not overturn AGW, and that, reality is that the long term trend is consistent with AGW, means I am suggesting there is doubt about the truth of AGW!

I could understand if he said your comments suggest you believe the world doesn’t need to reduce fossil fuel use.

But I have no idea why he thinks my posts suggests we shouldn’t move to reduce fossil fuel use. And if Nathan thinks my posts instilled your belief about fossil fuel use in your head, well… I think we can both get a chuckle out of that idea!

Lucia
I am not alone in this. If you read the other commenst around here you will see exactly this:

“He thinks my saying models over project, but the data does not overturn AGW, and that, reality is that the long term trend is consistent with AGW, means I am suggesting there is doubt about the truth of AGW! ”

From people who read your posts.
If this is not the case, then that’s great.

BTW, I got -1.4C/Century from the last 9 years data, I multiplied the 9 years by 10/9ths is that wrong?

Marcus [10532]

Thx for the “peacemaking”, but my impression is that Zeke and I have more or less agreed that it is best to agree that we disagree…

For the nth time [and Pielke Sr be my witness], NEITHER El Ninos NOR La Ninas can be counted as separate elements in a climate equation.
Either the entire atmospheric system is gaining joules [from whatever source and by whatever forcing factors] or the system is losing those joules. Based on best available data on all relevant metrics the earth’s atmospheric system has been losing joules for the better part of a decade.

For clarity all around: yes, I am a dyed-in-the-wool skeptic and I argue hard, but I have no bone in this fight. That is contrary to say Al Gore [who has significant carbon trading holdings and who over the past few years has made over US$100 million on those trades { matter of verifiable SEC records, if one cares to look}] or George Soros, who has even larger carbon trading holdings. And for those who really who don’t like the “politics” of the AGW/ACC story, might I remind you that Soros through one of his “socially helpful” agencies is one of the key financial supporters of the RealClimate blog. QED.

It’s useful to remind ourselves from time to time where some of the AGW/ACC “messages” come from.

Caveat Emptor.

Nathan–

A) Yes, you are alone on in your mis-interpretations. Yes. I think it’s great my other readers don’t seem to confuse the things I say the way you do.

B) Your method of computing the trend is wrong. My regular blog visitors know that’s wrong. They know that if they suggested the idea one could get the 9 year trend that way, I would say that it’s wrong. I suspect even you know that’s wrong.

Lucia

A) I think perhaps you are being generous, if you see below there are people who are misinterpreting your work
Tetris “As you correctly note, there is a lot of political credibility at stake. The implications of allocating billions of tax payers dollars on purported mitigation policies that are based on questionable information are far reaching.”
Chris schoenveld “In all these discussions there appears to be a consensus among the commentators that a 2C/century increase is something to worry about rather than to be happy about. I think more (and less biased) research is warranted to establish the pros as opposed to current focus on the – in my opinion less worrisome- cons. Surely, everybody agrees that a 2C/century drop would really be worrisome.”

You see Tetris using your work to justify doing less, and Chris to imply that there’s nothing to worry about. Your work is consistently referred to by ’skeptics’ – On the Jennifer Marohasey Blog, for example.

The use of your work is promoted by sceptics because you only ever point out the errors.

B)
I described what I meant poorly, I meant did you multiply your 9 year trend by 10/9ths to get a decadal trend.

Marcus,

” Doesn’t it strike all the skeptics as a little odd that we have a strong la Nina and a “weak sun”, and 2008 is _still_ a pretty warm year?”

Lemme see, in 2008 several glaciers that had been retreating showed growth, studies showed most of the galloping Greenland Glaciers had slowed to normal, large snow extents in Northern Hemisphere, Arctic Ice larger than previous record low, unseasonably cool temps in Australia and other areas, major flooding and low records in midwest and eastern US and some other areas, some European Ski resorts open 11 out of 12 months, Steig showed Antarctica has been cooling for 30 years even after Mannian adjustments to the data…

If that is what you consider a “pretty Warm Year” I hate to see what you consider cold.

By the way, we are experiencing another El Nino. With a negative PDO and signs of ADO possibly going negative, get used to it.

Here is an article about the ocean warming data:

http://earthobservatory.nasa.g.....intall.php

Don’t really see anything in there saying the Argo data show ocean warming even after 3 separate adjustments. I DO read a LOT about how they are attempting to get the data into agreement with the models and other data like SEA LEVEL, which hasn’t been rising as fast the last 2 years!!!!

http://sealevel.colorado.edu/

By the way, since the Team has told us that warming AND cooling is CONSISTENT with their models, exactly what ISN’T??? Kinda takes the power out of the CONSISTENT WITH statement doesn’t it??

@Kuhnkat
———————————————-Beginquote
…
By the way, since the Team has told us that warming AND cooling is CONSISTENT with their models, exactly what ISN’T??? Kinda takes the power out of the CONSISTENT WITH statement doesn’t it??
———————————————-Endquote

And as stated earlier: designing good (real world) tests for the models becomes difficult indeed!

Cassanders
In Cod we trust

In all these discussions there appears to be a consensus among the commentators that a 2C/century increase is something to worry about rather than to be happy about. I think more (and less biased) research is warranted to establish the pros as opposed to current focus on the – in my opinion less worrisome- cons. Surely, everybody agrees that a 2C/century drop would really be worrisome.
.
What makes you think that ?
If there is an example of NON consensus then it is this point .
On most climate related blogs , this one included , scientifically oriented people and actual scientists agree about many aspects of the physics involved in the atmospheric processes .
Most agree that there is a temperature increasing effect of IR active gases , the most important of which being H2O .
Most will also agree on a number “all things being equal” somewhere between 0.5°C to 2°C per doubling .
.
Disagreements will begin to appear when feedbacks are included or when the climate “noise” (what an irrelevant and misleading term to describe weather ! It is even worse than the “greenhouse effect” which has nothing to do with the real greenhouse effect) is considered .
.
And the biggest disagreements will appear when the question is whether 2°C in a century matter or are “worrying” in any significant way .
While the laws of nature won’t change in 1 century , everything else will .
The population and its distribution , the economical growth , the energy sources (f.ex nobody will talk about fossile energy because there will be hardly any left) , wars , revolutions , transport means etc etc will have changed and some in ways we have no clue at all .
So while there are many things to be worried about at century scales (population growth and energy supply being the most important) , a variation of some average by 2°C up or down is clearly not on the list .
People who focus on this negligible issue are either deluded or have an agenda .
Sometimes both .

Nathan–
A) These people are expressing their own opinions. There are many people in comments and they express a range of ideas. The are not all representing my work, opinion or anything else.

Nothing in this suggests tetris developed his opinion about action based on my results or blog.

As you correctly note, there is a lot of political credibility at stake. The implications of allocating billions of tax payers dollars on purported mitigation policies that are based on questionable information are far reaching.”

Rest assured that tetris thought this before I even began blogging.

Chris thinks 2C/century is not alarming. Once again, his idea of this is not hte result of my indicating that the IPCC projections of about 2C/century are inconsistent with data on the high side.

So, these examples simply do not support the notion that what I blog about causes these people to think what they do.

You can use the opportunity in comments to try to convince Chris that 2C/century is alarming– and even 1.6C/century is alarming. Or not.

But trying to suggest that they think these things because I discuss whether or not the data support any particular prediction is silly.

B) Now that you have clarified, the same answer holds. I did not do that. No one who had mastered 8 grade algebra would multiply a 9 year trend by 10/9th to get a decadal trend.

So why are you asking this? Are you being disingenuous? Or do you really not know? Because at this point, I honestly can’t tell.

TomVonk:

Hear-hear. A measure of perspective is too often missing in these types of discussions.

I do not know if the most shrill contributors (on both sides) are simply deluded or if they have an agenda, but those who object to legitimate efforts to validate or invalidate the predictions of climate models, or who object to the weighting of costs versus benefits of public action on carbon dioxide emissions are doing a terrible public disservice.

tetris (Comment#10497)

Anecdotal but real nevertheless, in the mid 1960s the English Channel froze over several times and the Dover-Calais ferries got stuck… Now that’s cold.

What??? Where did you pick up that little gem of disinformation?

Well said TomVonk.

My one complaint is that your sentence “Most will also agree on a number “all things being equal” somewhere between 0.5°C to 2°C per doubling.” given that the mean across the 11 or so most common published estimates is closer to 2.5, with a range of 1.5 to 4.5.

http://www.cgdev.org/userfiles.....ange/1.JPG

Hare and Meinshausen is a good paper to wrap your head around various sensitivity estimates and mitigation scenarios: http://www.earthscape.org/r1/ES16123/ (for the paywall-free working paper version).

An addendum: this refers to equilibrium sensitivity, of course. Expected 2100 mean temps resulting from a doubling of CO2 is 2 degrees C with a range of 1.5 to 2.5 (again, see the H&M paper). That said, it does matter a bit -when- the doubling occurs. If concentrations were to magically hit 2xCO2 tomorrow and stay stable from than till 2100, it would be different :-p

Zeke:

I think TomVock meant the direct radiative effect of doubling CO2, before including the feed-back/amplification factors assumed in the climate models. Without feed-back, a sensitivity number centered close to 1C for double the CO2 is probably about right, and 0.5C to 2.0C certainly covers the credible range.

kuhnkat (Comment#10543),

Lemme see, in 2008 several glaciers that had been retreating showed growth

As ever, some glaciers have gained mass whilst others have lost it. The trend is negative, and has accelerated whilst we’ve been having all this ‘global cooling’ -

http://www.sciencedaily.com/re.....090002.htm

studies showed most of the galloping Greenland Glaciers had slowed to normal

“Normal” meaning a continuing loss of ice mass from Greenland.

Arctic Ice larger than previous record low

Whilst likely recording record lowest volume.

unseasonably cool temps in Australia and other areas, major flooding and low records in midwest and eastern US and some other areas

Along with anomalously warm temperatures in Australia and other areas, major droughts and high records in all sorts of areas that lie outside the 2% of the globe’s surface covered by the US.

some European Ski resorts open 11 out of 12 months

Offering glacier skiing, I presume. What’s new about that?

Steig showed Antarctica has been cooling for 30 years even after Mannian adjustments to the data…

No he didn’t – he showed East Antarctica cooling, West Antarctica warming.

If that is what you consider a “pretty Warm Year” I hate to see what you consider cold.

Well, anything other than one of the top ten warmest years on record would be colder, wouldn’t it?

Just to add one more rebuttal to Simon Evans’ list:
“SEA LEVEL, which hasn’t been rising as fast the last 2 years!!!!”
2008 is about 3.8 mm higher than 2006, according to the sealevel.colorado.edu data: 3.8 mm in 2 years is… slightly less than 2mm per year… certainly, that’s slower than the 3 mm per year trend since the mid-90s, but higher than the average 20th century rate of 1.7 mm/year.

No he didn’t – he showed East Antarctica cooling, West Antarctica warming.

.
Actually, his satellite recon shows a trend not statistically different from zero in East Antarctica from 1969-2000 and warming prior to and after.

Whatever the temperature does in the future it will not affect the temperature trend due to WM-GHGs. That would be to reverse cause and effect. If the temperatures stall for another decade it would not affect the underlying trend.

What it may affect is our estimation of that trend, in the light of new science, or better understanding. The underlying trend is whatever it is, falling temperatures will not change it. It just happens that we do not know what it is. There is a lot of hard science in the atmospheric radiation equations that is not going to go away whatever the temperatures do.

Given that the world, and the science (models) seem to be diverging and that we would have to wait perhaps another 5-10 years for it to be irrefutable, would it not be a good idea for people to say that there is more in Heaven and Earth and start beavering away, hopefully at least to get as far as to say that this is interesting and really needs looking at. To take it on the chin and say “well it may just be noise but it is far too important to ignore”. That is why I do not get a warm feeling from the modellers and the “the science is settled” mindset. Science is never settled.

The stupid thing is that even 20 years of stagnation would not overturn the known science, it would neither threaten the climatologists nor the
models. It is what should get their juices running, (the climatologists not I hope the models).

If, as is quite plausible, we are experience some cyclical behaviour then in time warming will return. Each year that goes by with no warming does not refute the science nor the effects of WM-GHGs but it may just help us improve the science and the models and probably lead to the conclusion that the underlying trend is smaller than the models currently suggest.

To be honest, I do not get the reason that this has to feel like trench warfare.

Alexander Harvey

Zeke:

I think TomVonk meant the direct radiative effect of doubling CO2, before including the feed-back/amplification factors assumed in the climate models.

Right , that is what I meant .
But the numbers were not important – my point was more the temporal scale – to get the whole thing in a right perspective .
As I have left Academia a certain time ago because I wanted to be more in contact with the reality , I have lost the obsession about the accuracy to the third decimal .
Or even to the first for that matter .
That’s why , even if I have a precise and argumented physicist’s opinion about the climate dynamics , I am not interested in the average (!) temperature in 1 or 2 centuries .
I am much more interested in the question if this particular parameter matters for man when it fluctuates in a certain band of orders of magnitude .
.
And here the answer is a very clear no , it does not matter at century scales .
Why ?
Mostly for 2 reasons :
.
First is more philosophical . There doesn’t exist an average (!!) temperature that would be optimal for man .
As long as it stays in a band like – 10°C in Finland and 40°C in Las Vegas where (luckily) it has been staying for the last millions of years and probably even billions , everything is acceptable within the band .
I am ready to concede that Ice Ages are a bit less confortable for a civilisation without technology and energy but if that is what we get then that is what we’ll live with .
Obviously it is not because some average (!!!) temperature is now 14.586 °C that
a) it is the paradisiacal value
b) it will stay at this paradisiacal value for centuries and centuries . Amen .
.
Second is connected to prospective studies where I happened to work several years .
In prospective even restricted to specific fields (kind of “everything else equal” approach) , 50 years are stretching it to its limits .
100 years would make everybody exploding with laughter .
Trying to elaborate a set of scenarios of the WHOLE mankind with a 100 years or more horizon would be a pleasant joke .
There is not a single variable that could be better than extrapolated over such huge time intervals .
The differences between extremal scenarios would be so enormous that the conclusion would be “about anything can happen” and that would be the right conclusion .
So while there are few qualitative certainties like “the growth rate of the population is the dominating parameter” and “the purchasing power varies like the GDP” , “the technology varies by ruptures and not continuously” , “average life expectancy is not constant” etc most of the parameters and their values is unknown .
.
Under those conditions , questions about +/- 2°C in some average at a century horizon don’t matter and are a second or third order correction to the main effects mentionned above .
And even if somebody wanted to account for this small correction in a scenario , he would not be interested by averages but by local values because both mankind and its economical activities are local .
It is not today and probably not after tomorrow either that some GCM will give me a local prediction of temperatures and precipitations at century scales with +/- 20 % so the question is irrelevant anyway .

Alexander Harvey (Comment#10622):

I agree completely that the true impacts of CO2 (and other emissions) on the Earth’s climate are what they are, and natural climate variation superimposed on those impacts has no influence. Indeed, natural variations ought to help to refine/optimize the performance of climate models.
.
But the reason climate change discussions seem so much like trench warfare is that so much is at stake. Many politicians propose draconian reductions in energy use and enormous public expenditures based mostly on the projections of IPCC climate models. If those reductions in energy use and public expenditures take place, and the dire projections of the models turn out to be far from correct, then we (humankind) will have made a very costly mistake.
.
Climate modelers (and certain politicians) have a personal interest in making sure the projections of the models are not refuted, and these people object strenuously to any data and/or analysis which suggests that models are making unrealistic projections of future warming: “You don’t know what you are talking about… The models are right….. The science is settled”. Climate modelers are not saints; they are subject to the same human follies as the rest of us, such as expectation bias and fear of being shown to be wrong. Were they modeling protein conformation or turbulent air flow, instead of justifying large and costly worldwide economic changes, they would no doubt be as defensive about the validity of their work, but nobody outside their field, least of all politicians, would care.

So I believe it is critically important that technically trained people who do not work on climate models, and so who have no personal interest at stake, critically examine the model projections versus the real data to make sure we (humankind) do not make a very costly mistake.

SteveF:

Thanks for your comments.

Being a simple person I think:

Let the politicians do the politics (nobody ever believes them anyway).

Let the modellers build and run the models.

And let the environmentalists get alarmed about the environment.

That said, damage due to local climate projection, (frequently done by people who are not experts) is not just in the future, it is happening now. The prospect of rapidly increasing sea levels is blighting the lives of coastal dwellers in my locale. It might not be much in the scheme of things but if it is your house and it has sufddenly become worthless it is not a joke.

Alex

[...] the method described in Santer. Chip Knappenbeger describes Michaels’s method in comments here. Written by: lucia « GISS Feb Reported: Trend Since Jan 2001 still [...]

[...] I’m using and 2 standard deviations from the mean. Michaels appears to be using a slightly different methodology that involves grouping the runs from a single model together before calculating the 95% bounds. [...]