Spencer and the Mystery Journal

Hi William,

I don't claim to know what journal he has published the paper in but I don't have any problem understanding what he is saying.

If the journal was E&E he would have no problem getting the paper published and therefore no particular reason not just to tell us this fact at the outset.

He appears to be concerned that if he reveals the name of the journal before the paper is actually published, people are going to write angry letters to the journal editor in the hope that its publication could still be blocked.

I also doubt for the same reasons that it is a minor journal after bigger journals rejected it as with Lindzen & Choi.

I am not offering an opinion on whether Spencer's concern is justified, but I don't have any problem understanding what he is saying.

-Alex

[I don't understand your comment. What Spencer is saying is completely obvious, as is the fact that it is nonsense. What made you think that his meaning was in any way obscure? -W]

More fodder for rejectionists. I'd have thought Spencer would have to provide it, given his role at the Marshall Institute and his pledge to thwart efforts to curb CO2 emissions.

Hank Roberts unearthed a top notch bit of work by Ullrich Ecker.

Interesting to see if Dessler gets to reply

William, okay well maybe your conclusion that perhaps the paper is published in E&E or some minor journal was supposed to be satirical.

[Its not a conclusion, it a question. You can tell that because the sentence ends in "?". Also, I can't wait for the answer so... does anyone know? would have indicated to most readers that I don't know -W]

J Bowers, I am curious to know what Spencer's role at the Marshall Institute actually is. Do you happen to know?

Alex Harvey, Spencer's on the board of directors of the Marshall Institute. He's also on the board of advisors for the Cornwall Alliance for the Stewardship of Creation, through which he made his pledge.

Is it possible to write and publish a rebuttal of Spencer's rebuttal even before the latter's published? It might be an interesting exercise...

-- frank

And the funny comments have already begun!

It seems the journals are looking for a graceful way to back out of the corner they painted themselves into, by publishing properly executed scientific work.

Ha!

BTW, how many times does Dr. Roy Spencer⢠have to completely and utterly destroy all competition before the scientific world acknowledges that he's the smartest of them all? From the way he's treated, you'd almost think that he's not the smartest. Almost.

Spencer's starting to sound more and more like MIchael Behe every day ...

Spencer's starting to sound more and more like MIchael Behe every day ...

Funny, that:

http://tinyurl.com/5anfnu

Oh, yes, it's no secret that Spencer's a creationist.

I was speaking to his increasing megalomania that leads him to apparently seriously believe that he's outsmarted every mainstream climate scientist in the world and has shattered the field to bits with the startling conclusions of his "research".

This is Behe's schtick regarding modern biology.

There was a time when Spencer - on the surface, at least - was a serious scientist. Same with Behe, in his field. They've both followed a similar path towards crankdom in support of their ideological beliefs. Behe was earlier, that's all.

I see you put "Dr" in quotes. I don't blame you. Dr Spencer is one of those folks with a latent inferiority complex who must forever be parading around the fact they have a PhD. I know of a man who signed a speeding ticket from a cop with "Dr So-and-So". It was posted on a lab bulletin board and others rightly laughed. In Dr Spencer's case we have the additional oddity: why does he thinks his "Dr" is so impressive when he clearly thinks most of the academy that awards these degrees is so wrong in so many fields (inc. evolutionary biology). What an intellectual midget is this Dr Spencer.

It's self explanatory really.

Our early-2010 paper showed extensive evidence of why previous attempts to diagnose feedbacks (which determine climate sensitivity) have likely led to overestimates of how sensitive the climate system is to forcings like that from increasing CO2. The basic reason is that internal radiative forcing from natural cloud variations causes a temperature-radiation relationship in the data which gives the illusion of high climate sensitivity, even if climate sensitivity is very low.

A sad case because Spencer is a technically competent fellow who has made useful contributions in his specialty. But as his embrace of creationism shows, he's willing to throw science out the window when it suits his personal agenda.

Perhaps we shouldn't be all that surprised that scientists can lose their way amongst the same human foibles (pride, egotism, etc.) as other mortals.

OT, but I know you've posted on Woods's experiment before. Woods's greenhouse raises it's head again.

[Gosh, that is wildly exciting - thanks. I've blogged it -W]

The link is missing for Wood's experiment. It can be found here.

Cheers, Alastair.

Steve Balibar:

â..Spencer is a technically competent fellow who has made useful contributions in his specialty..â

dhogaza

âThere was a time when Spencer - on the surface, at least - was a serious scientist.â

Spencer gets more credit as a âcompetentâ scientist than he deserves. Spencer (and Christy) made breakthroughs in the application of MSU to characterizing tropical cyclones and then tropospheric temperature. But they messed up temperature measurements big time from the start, despite their insistence on the precision and accuracy of their measurements [1,2,3]. Already in 1991 it was pointed out [4] that their analyses were insufficient to distinguish the cooling they would soon try to sell [5] from warming that would be consistent with surface measurements.

And it was repeatedly left to other to sort out the various messes in the analysis of MSU data: that the method of averaging different satellite records introduce a spurious cooling trend [6], that disregard of orbital decay introduced another spurious cooling trend [7]; that MSU-2 showed a spurious cooling trend due to spillover of stratospheric cooling into the tropospheric temperature signal [8], and later still that the diurnal correction applied by Christy and Spencer (a sad litany of incompetence) was of the wrong sign and gave yet another spurious cooling trend [9].

Spencer (and Christy) spent the better part of half a career getting this stuff wrong. Seems like they canât bring themselves to cast off their preferred interpretation that the earth doesnât respond much to enhanced greenhouse gas concentration, and have given up any pretense at addressing this properly.....pretty sad.

[1] RW Spencer and J. Christy (1991) Precise monitoring of global temperature trends from satellites. Science 247, 1558-1562.

[2] Christy, Spencer and WD Braswell (1997) âHow accurate are satellite âthermometersâ? Nature 389, 342.
âWe believe that lower-tropospheric temperatures measured directly by satellites have excellent long-term accuracy, as seen by comparisons with independent atmospheric measurements from weather balloons.â

[3] JRC and RWS (2003) Reliability of Satellite Data Sets Science 301, 1046-1049.
In which our intrepid duo insinuate that their data is accurate and models wrong.

[4] B.L. Gary and S. J. Keihm (1991) Microwave Sounding Units and Global Warming Science 251, 316 (1991)

[5] RWS and JRC (1993) Precision lower stratospheric temperature monitoring with the MSU - technique, validation, and results 1979-1991. J. Climate 6, 1194-1204.
âThe largest globally averaged temperature variations during 1979-91 occur after the El Chichon (1982) and Pinatubo (1991) volcanic eruptions. These warm events are superimposed upon a net downward trend in temperatures during the period.â

[6] J. W. Hurrell & .K E. Trenberth (1997) Spurious trends in satellite MSU temperatures from merging different satellite record. Nature 386, 164 â 167.

[7] F. J. Wentz and M. Schabel (1998) Effects of orbital decay on satellite-derived lower-tropospheric temperature trends. Nature 394, 661-664

[8] Q. Fu et al. (2004) Contribution of stratospheric cooling to satellite-inferred tropospheric temperature trends Nature 429, 55-58.

[9] C. A. Mears and F. J. Wentz (2005) The Effect of Diurnal Correction on Satellite-Derived Lower Tropospheric Temperature, Science 1548-1551.

I was about to post that Remote Sensing is a well-established journal, but then I realized I was thinking of the International Journal of Remote Sensing.

I wonder why Dr. Spencer decided to submit to this brand-new journal, instead of one of the three older/more widely-read ones. Perhaps they were desperate for material?

Hi William,

Remote Sensing looks like a journal for experts in remote sensing, which seems appropriate enough for an article about remote sensing.

What do you mean by "trivia like impact factors"?

-Alex

[If you don't even know what an impact factor is, you need to... well, look it up would be the obvious thing to do -W]

Published in Remote Sensing on Tuesday!

[Thanks. Blog updated after someone pointed me at the Heartland press release: its all so fitting -W]

Will the name be changed to Remote Viewing at any point on the near future?

> which seems appropriate enough for an article about remote sensing.

Alex my university department here has half a dozen remote sensing specialists, each of whom would have declined to review this paper as its core subject is climatology, not remote sensing.

Spencer used the popular crank trick of trying a journal just enough removed from the field that they would not be expected to have actually competent reviewers on their list for this. The journal ought to have had the sense to decline too -- as the better established ones would have done (and who knows, perhaps did).

At last! The final final final (this one really is) final final (we've got them now!) final (now it's REALLY final) final final nail in the coffin of the AGW conspiracy!

[Like an ice-pick to the head -W]

And, of course, the headline writer gets the crucial fact:

New NASA Data Blow Gaping Hole In Global Warming Alarmism
By James Taylor | Forbes â Wed, Jul 27, 2011
"James M. Taylor is senior fellow for environment policy at The Heartland Institute and managing editor of Environment & Climate News."

Hmmm. One of the head honchos of Remote Sensing, Shu-Kun Lin, Ph.D. (President and Publisher), also founded/co-founded International Journal of Environmental Research and Public Health, which published 'A Multidisciplinary, Science-Based Approach to the Economics of Climate Change. Carlin (2011)', and is also part of the MDPI chain. Remember that paper which made a splash in the deniaverse?

Why aren't these guys (Spencer, Christy, Pielke Sr.) publishing in MDPI's International Journal of Environmental Research and Public Health? It seems so much more appropriate. Is it so Heartland can say "peer reviewed paper" for a change?

I don't quite understand how someone like Spencer, who has a long sordid history of relying on bad data and analysis, is supposed to be taken seriously.

Spencer in 1997:

"The temperatures we measure from space are actually on a very slight downward trend since 1979 in the lower troposphere. We see major excursions due to volcanic eruptions like Pinatubo, and ocean current phenomena like El Nino, but overall the trend is about 0.05 degrees Celsius per decade cooling"

"We've concluded there isn't a problem with the measurements"

"Instead, we believe the problem resides in the computer models and in our past assumptions that the atmosphere is so well behaved."

Re #2.
To JB and Hank. Please don't encourage them by teaching them about this! On second thoughts, they probably knew it already as do most propagandists.

For those curious to read the original paper, Mr. Taylor has now provided working links. They are:

http://www.mdpi.com/2072-4292/3/8/1603/pdf

and

http://www.drroyspencer.com/wp-content/uploads/Spencer_Misdiagnos_11.pdf

Barry Bickmore takes a bite

> MDPI
Get your papers in now
http://www.mdpi.com/about/apc/
before prices go up
Changes of Article Processing Charges (APC) Effective 1 October 2011 and 1 January 2012 Respectively

Assuming somewhere in here is a scientific dispute, sensitivity has a given likely range of 1.5 to 4.5C. Spencer is presenting what he considers evidence for the low end. From what I've read here, I can see Spencer must be a totally bought and paid for buffoon. He's so silly with his little journal paper and his self promotion. He couldn't have gone to any of the right schools.

[I think you're right in general terms, though perhaps a little harsh in your language. If you're interested in Spencers errors, others have already mentioned the MSU stuff, but Barry Bickmore has a nice explanation of why his modelling is worthless -W]

There is a clear political division in the support and dismissal of Spencer. It really is necessary to remove climate mitigation from politics. The new buzzword is climate pragmatism. Pragmatism incorporates the variety of valid reasons framing and what I've been calling the focus model. There's a report which I have not read. It is part of a major shift in social science since the Copenhagen collapse.

The "climate is the most immediate and important threat" framing is not persuasive. Scholarly papers inform that the information deficit model is not applicable to the policy debate. Within the variety of reasons framing, the focus model identifies the goal, which the pragmatists want everybody to agree is decarbonization.

I wrote my own response to Spencer's latest work:

http://circleh.wordpress.com/2011/07/28/roy-spencer-pulls-another-misle…

Has anyone kept track of the number of times and ways Spencer has proved everyone not named Roy Spencer, Ph.D wrong?

You can tell he has everyone else on the ropes. Read the fear from the Climategate emails as Ben Santer and Michael Mann, in private and therefore candid emails, admit to each other that Roy Spencer, Ph.D has the superior knowledge!

Christy and Spencer have made a scientific career out of being wrong.

Christy and Spencer continue to lose more and more scientific credibility with each awful paper they publish.

See that? Roy Spencer, Ph.D has to have said something right for them to call him wrong.*

* According to creationist logic

After all the subterfuge (Spencer) and ludicrous hyperbole (Forbes), has anyone noticed that Spencer's paper has essentially nothing to say about (Charney) climate sensitivity, and only purports to show that it's difficult to determine v. short-term feedbacks from measuring radiative feedback responses to changes in surface temperature (due e.g. to fluctuations arising from ENSO etc.).

This isn't very controversial (other than the dubious concept of "internal radiative forcing"), especially when the feedback under consideration is cloud response (Dessler has indicated that the cloud feedback response may be positive using the method of assessing short term response to surface temperature variation, but with v. high error bounds)...the situation with respect to water vapour feedback is more straightforward..it's certainly positive as Dessler also has shown when considering rapid response to temperature variation, and many others (e.g. Soden) have shown when considering long term water vapour response to greenhouse-forced tropospheric temperature rise...

So this "storm-in-teacup" farce is an attention - seeking (Spencer), agenda-supporting (Forbes et al) exercise. Who'd have thunk it!

Is chris correct that the science in the paper, as opposed to the press release science, is not very controversial?

[As Gavin says, it looks to be just the same broken analysis repeated. That could be "controversial", like publishing Creationism would be -W]

Gavin Schmidt seems to sum it up neatly HERE.

Why not read the paper Paul and compare with the press release.

[Well, to be fair few other people are bothering. I certainly haven't -W]

There is nothing specifically concluded about the Earth response to radiative forcing other than a sentence in the abstract about satellite-based metrics in 2000-2010 departing in the direction of low climate sensitivity compared to model computations but that this can't be accurately quantitated in terms of feedbacks that determine climate sensitivity. That seems to be the basis of the Forbesian hyperbole.

This is further qualified in the body of the paper, where Spencer states that this discrepency isn't necessarily real (it's "nominal"): e.g.

"Yet, as seen in Figure 2, we are still faced with a rather large discrepancy in the time-lagged
regression coefficients between the radiative signatures displayed by the real climate system in satellite
data versus the climate models. While this discrepancy is nominally in the direction of lower climate
sensitivity of the real climate system, there are a variety of parameters other than feedback affecting
the lag regression statistics which make accurate feedback diagnosis difficult. These include the
amount of non-radiative versus radiative forcing, how periodic the temperature and radiative balance
variations are, the depth of the mixed layer, etc., all of which preclude any quantitative estimate of how
large the feedback difference is. More recent work which attempts to minimize non-feedback
influences [14] might well provide more accurate feedback estimates than previous studies."

In other words nothing of significance to report, since the analysis is underdefined in Spencer's view...that indeterminacy actually seems to be the point of the paper.

There is a controversial element to the paper. Spencer proposes a concept of "internal radiative forcing" which seems to be a sort of "cart-before-the-horse" notion that (I think!) variations in clouds (say) act as a radiative forcing. The accepted way of thinking about (persistent, non-random) elements of cloud variation, would be that these are feedback responses to changes in surface temperature, or external forcing. e.g. Dessler's recent paper in Science (cited by Spencer) suggests that clouds respond (positive feedback) to changes in surface temperature arising from ENSO variability. Spencer proposes (I think... it's a slightly odd concept) that the clouds can change (persistently) by themselves and result in (persistent) changes in surface temperature (i.e. "internal radiative forcing!!)...Spencer would say that it's cloud changes that cauee ENSO variation..

chris,

I'm in the 3C sensitivity camp (except for taking 2.7C in the office pool). Won't read the paper. Wouldn't understand a word of it.

A controversial element to the paper - as opposed to elements that are merely said to be incorrect - is internal radiative forcing. A controversial element was the PR roll out of the paper and the reaction to that roll out.

No worries Paul, I totally agree that the whole episode (Spencer's obfuscation, Spencer's puerile coyness, the dreary journal, the pathetic PR roll) is contrived claptrap, and it's pretty sad that people (mostly in the US, but there are nutters everywhere), buy into this stuff even to their own detriment..

In fact you don't have to understand very much of the science to see that the press release bears little relation to the analysis presented and its interpretations. Spencer is presumably complicit in that deceit.

It's what I call the "fu*^ you" approach to misinformation in which the perpetrators know that what they're selling is horsecrap, and they know that you know too. The point isn't so much to create an alternative interpretation of particular science, but to reinforce "battlelines". Appropriately enough it's a common tactic amongst Creationist fanatics as shown rather delightfully in the Dover trial!

Incidentally, the paper isn't the worst thing that's ever appeared in the scientific literature, and I wouldn't say that it doesn't deserve its place in what seems to be yet another of the low ranking open-access "you pay, we publish" outlets.

"I cannot believe it got published." -- Kevin Trenberth

Without benefit of reading the paper I'm guessing that Gavin is right. But I particularly enjoyed this quote from Gavin.

"they appear to be calculating regressions on smoothed data (without taking into account the decrease in degrees of freedom)"

Remarkably he didn't raise this same objection about Rahmstorf and Vermeer's two sea level papers.

[? You'll have to give some pointers on that I'm afraid -W]

W,

I wrote a series of posts on this general topic some time ago. I sort of brushed over the smoothing issue, but it is an important reason why there is so much auto-correlation. In any event both the Science and PNAS papers rely on a method that first smooths the data using an unpublished algorithm. Even at that the 2007 paper didn't really work as I pointed out. Two comments were published at the time in Science, but they were brushed aside by Rahmstorf using what was essentially slight of hand. Since that time the paper has been sited numerous times.

http://nierenbergclimate.blogspot.com/2009/11/follow-up-on-rahmstorf-20…

[I may have to read this to understand. This is the R paper that used fitting to the past to predict the future, yes? Reading your first post, I find "...we do not agree that simplistic projections of the nature presented in [R07] substantially contribute to our understanding of the uncertainties in the nonlinear relationships of the climate system." very odd as a comment. It seems orthogonal to what I think the Rahmstorf paper is about. Also, your first post would be much better with pictures, and links to Rahmstorf. Oh, and googling "ssatrend" now produces http://www.glaciology.net/software/ssatrend-m as the first hit -W]

Welcome to CafeMom - Global climate change

"Some of you may have heard of a pack of paid shills called The Heartland Institute...
[...]
Those of you who remember polynomial graphs from school will know that the more variables you're allowed to change in an equation, the easier it is to make your equation fit closely to a set of data points. In fact, if you have the same number of variables as datapoints, you can always get an exact fit. This is why legitimate scientists, if they use regression analysis upon a bunch of data to come up with a curve that fits it, then carry out a confirmatory analysis using different data, to double check that the fit wasn't just coincidence.

Spencer, apparently, is fond of skipping this step. He specialises in taking the inputs he wants to be the cause of temperature records (for instance: cloud cover), then using lots and lots of free variables to combine these inputs to produce a a result that is something similar to what's been observed.

To quote one reviewer:

âWell, give me more than 30 parameters, and I can fit a trans-dimensional lizard-goat and make rainbow monkeys shoot out its rear end.â"

Mum always knew best.

W,

Hmm, I thought the series of posts was clear as at least what paper I was talking about. In answer to your question, yes it is a paper that tries to create a regression model of the rate of sea level rise based on the global temperature. The theory is that the rate of sea level rise will be proportionate to the difference between the current temperature and some base temperature. To do this he first smoothed all the data using SSA, which is already wrong before you do regression, and that was the point that Gavin was making. Although as I point out he didn't seem to mention it to Rahmstorf or Vermeer.

The site you refer to is relatively new from Grinsted. But note the warning, which is what Dr. Grinsted told me when I inquired.

"Please be aware that this is a work in progress. In particular, you should be aware that:

Data boundaries have an effect on the smoothed series. You should examine the code to see how I deal with it and whether another approach might be better for your data. (Note, In the new code i have made it easy to add new boundary treatment methods.)
A new method for calculating the confidence intervals is being used. It needs further testing. Please beware."

This is hardly a well understood smoothing algorithm even now. But even if it was Gavin's general comment would apply.

[That sounds correct, yes (without having time to examine the details). But I would say that Rahmstorf's general point still seems reasonable: that the historical temperature-SLR relationship is useful for predicting the future, or at least exploring it -W]

As Grinsted says on his web page:

"There were some blog-comments on climateaudit etc. about ssatrend padding uncertainties being wrong. However, the comments were all over the place on multiple blogs, so I had no energy to find out what they were talking about."

This situation appears to be getting worse now. I tried looking for likely references; blog science has a lot besides NN's pages.

So far I don't see anyone has been able to pull the comments together coherently or get a statistician to publish on it.

Discover: No, new data does not âblow a gaping hole in global warming alarmismâ

W,

Yes I think the relationship of global temperature to sea level rise is interesting. But the problem was that Rahmstorf's model was fundamentally wrong. Sea level rise is made of two components steric (caused by temperature increase), and volume (caused by ice melt and other factors). His model tried to relate sea level rise to a single term having to do with temperature.

[I don't think that distinction makes sense. The component you've labelled "volume" comes fro ice melt and is related to temperature, too -W]

This could possibly explain the rate of volume change but obviously not the steric change which has to do with the rate of temperature change.

[Errm, why? Steric is caused by temperature increase (as you said at first) not rate of temperature increase (as you've now switched to) -W]

If you could remove the steric change then you might be able to model the volume change that way. This might be possible by using a model of the steric change based on sea temperatures. I tried that, but was unable to get a good fit. He managed to get a good fit on a flawed model through smoothing, smoke and mirrors.

Vermeer tried to fix this by adding a second term. But the smoothing issue remained, an he didn't do any out of sample testing. (Weirdly he ran a test against a model that only works for steric changes.) It was, in my opinion, a very weak paper that got published in PNAS.

The net of all this is that I am interested in what people who actually study sea level will come up with as a consensus in the next IPCC report. Sadly I think they will pay some deference to the Rahmstorf and Vermeer papers because of all of the uproar last time over the IPCC sea level forecasts. The AR4 conclusions were remarkably consistent with the prior three assessments.

[AR4 was deficient, I think, in that it downplayed other contributions to SLR; and R+V was part of the response to that. But there have been other papers since saying much the same. So yes, I'd expect more on that from AR5 -W]

*[ That's enough ripping off my best imitation of John Maddox doing Private Eye -Ed.]

http://www.nature.com/nature/journal/v475/n7357/full/475423b.html

NN, FYI two papers in the last month have confirmed that in the Eemian the rise was mostly melt, which makes you a lot happier about R+V, yes?

[I would imagine not, since NN's complaint is about the procedures R+V used. Just because they produce the correct end result would not be good enough -W]

And even happier since all of this meshes nicely with Hansen's assessment that models currently mix heat into the deep oceans too quickly, yes?

Hey Steve,

Do you have a link to those papers? How could significant rise be anything but volume? And how does that relate to Hansen? Heat being mixed would have to do with steric sea level I think.

BTW I'm not happy or sad about any of this, other than crappy stuff getting published, which I guess makes me sad. Are you happier if it turns out we will be all under water?

[Snipped. Sniping -W]

[Snipped. Response to snipped sniping -W]

Shorter:

Sea level rise is made of two components, steric and volume, which are caused by temperature increase.
The AR4 and prior three assessments lacked information sufficient to assess the effects of temperature increase on sea level.

NN -- "How could significant rise be anything but volume?"

Localised sea level rise due to gravitational effects? Jerry Mitrovica's been doing some interesting work on this.

Briefly putting on the homework-help hat: kids, here's how to start, if someone's referring to stuff without citing it. Take the description you found and paste it into the Google search box and read the first page or two of results, it won't take too long. This seems likely to describe both recent papers mentioned above:

http://www.bitsofscience.org/eemian-sea-level-rise-greenland-antarctica…

"Posted on July 29, 2011 by Rolf Schuttenhelm
Eemian sea level rise Antarctica
Last week we learned 5 percent of the Eemian sea level rise was thermal expansion of the oceans. Today we learn the slightly higher temperatures led Greenland to âonlyâ add an extra 1.6-2.2m...."

W,

Let me explain about why there is a distinction. Assume that the temperature ante is 0C. The global temperature rises to 1C. This will cause thermal expansion while the temperature rises to 1C but then it will stop. This effect is well understood and modeled. So in this sense the sea level (not the sea level rate of rise) is determined by the temperature.

[By change in temperature from an equilibrium value, yes. Though of course there are long lag effects: it takes make hundreds of years for temperature changes to penetrate the deep oceans -W]

However what isn't as well understood is the amount of volume change that will occur and over what period of time. It clearly doesn't happen just during the period of temperature of increase, [Just like thermal expansion, of course, though the lags may be longer in this case-W] but will be over a very long period. It is also reasonable to assume that at 2C the rate of volume change will be higher than it would be at 1C [Just like the thermal expansion case -W]. Of course over a very long period you would once again reach equilibrium and so the total volume increase would in some sense be a function of temperature, but over shorter periods like decades or centuries it is the rate of volume change that is proportional to temperature.

[No, not following you at all guv. I'd say that to first order, rate of volume change is proportional to temperature-change-from-equilibrium -W]

So in summary over decadal scales steric sea level changes will be a function of delta t, but over those same time scales it is the delta of volume that will be a function of t. Rahmstorf's model assumes that the delta of the sum of steric and volume over short periods will be proportional to t which is not correct.

[I don't think that holds water (ho ho). Have you even got your terms right? "steric sea level changes will be" and "delta of volume" sound like you're trying to distinguish, but "delta of volume" is just "volume changes". And no, "delta of volume" isn't a function of absolute temperature. Of course it isn't. It can't be -W]

On the subject of the Eemian sea level rise. As I said this amount of sea level rise would have to be mostly volume changes. These papers agree with that statement. They are just discussing where the volume came from. Their conclusion is that less came from Greenland and more from Antarctica than was previously believed. But everyone who worked on AR4 already knew what the total sea level rise was during that period, and there was no change to that conclusion.

> model ... is not correct

That's not a problem, though.
You know the Box quote about models.

Did you follow the cites to the earlier modeling approach, going back to this one?
Greatbatch, R.J. (1994), A note on the representation of steric sea level in models that conserve volume rather than mass, J. Geophys. Res., 99, C6, 12767-12771 ?

Nicholas, we seem to be talking about some rather interesting analyses by Rahmstorf (and Vermeer), on a thread about a truly dismal paper by Spencer and Braswell. One might think you're diverting attention from a true stinker to do a bit of "fault-hunting" on some decent science! Or do we simply all agree that Spencer and Braswell is sufficiently dismal that it doesnât need any more attention?

There are some problems with your analyses of Vermeer and Rahmstorf (V&R).

1. "â¦This will cause thermal expansion while the temperature rises to 1C but then it will stop. This effect is well understood and modeled. So in this sense the sea level (not the sea level rate of rise) is determined by the temperature."

Yes and no; and the same could be said about the steric (land ice) contribution to sea level rise, which will also rise and eventually stop. It's the difference between equilibrium and transient responses and their time-dependencies. Remember that the second term in the model equation ( dH/dt = a.(T - T0) + b. dT/dt ) is a term that varies with the rate of temperature change. Therefore in the time series of interest, time-dependent changes in T (characterised by rates) will result in a contribution to the rate of change of sea level (it's contribution will be less long lived, of course, but since temperature is continually changing (rising mostly!) through 21st century, there will be a continual contribution from the second term.

Incidentally, although one might assume that the second term in the equation accounts for "rapid" changes in sea level change due to the thermal response, we have to be careful to recognise that this isnât necessarily the case. This is a very simple model after all! The terms are chosen to accommodate slow and rapid response components of sea level rise but the fact that addition of the second term greatly improves short term sea level responses to temperature (in fits to real and model data) doesn't mean that that the term discretely or fully accounts for a real-world thermal response.

2. "But the smoothing issue remained, an he didn't do any out of sample testing. (Weirdly he ran a test against a model that only works for steric changes.)"

Iâm not going to comment on the smoothing issue other than to note (i) that smoothing is O.K. to minimise "noise" if done with some thought (although one should recognise that resulting regression coefficients may be "over-good" as a result) and (ii) V&R described quite a bit of analysis of effects of different smoothing in their supplement.

However re your complaint of model testing, perusal of the paper indicates that in addition to the out of sample testing shown in Figure 1, V&R also did out of sample testing with the millennial scale model shown in Figure 2 and the millinial timescale ECHO-G and ECBilt-CLIO models.

3. âSadly I think they will pay some deference to the Rahmstorf and Vermeer papers because of all of the uproar last time over the IPCC sea level forecasts.â

I wouldn't be sad. As usual the IPCC will accommodate all the relevant information that pertains to the topic at issue (centennial scale sea level rise in response to mankind's enhanced greenhouse forcing). They will very likely expand the upper range of likely sea level rise, not just because of the enhanced projections from V&R, but also because independent empirical modelling supports higher sea level responses (e.g. Grinsted et al (2010), and projections based on renewed understanding of non-linear ice sheet responses support higher 21st century sea level responses (e.g.(Pfeffer et al (2010), (Price et al (2011)). These analyses tend to support 21st century sea level rises in the range 0.8 - 1.3 m under likely emission scenarios.

Of course any projection is subject to uncertainty, but we'd be foolish not to make contingency for these. Those that are able to address future sea level rise with minds concentrated by mature consideration of reality, are certainly planning for sea level rises of the sort projected by V&R. For example the Dutch Delta Commision report advises: "a regional sea level rise of 0.65 to 1.3 m by 2100, and of 2 to 4 m by 2200 should be taken into account."

V&R is not perfect. We can certainly "hunt out" flaws. But it addresses something of a "no-brainer" IMO; that the rate of sea level will vary according to the difference in surface temperature relative to some "equilibrium" temperature. Of course it's unlikely to be the perfect way of addressing this- it's a model!

[Apologies for leaving this in the queue for a while -W]

For Nicholas Neirenberg.

Sorry, should be 'Nierenberg'.

Nicolas Nierenberg - I'm a bit confused by your use of terms. You're talking about steric vs. volume but steric changes are also realised as alterations of volume.

What's different about sea level rise from glaciers/ice sheets etc. is that there is a change in the mass of the oceans. Talking about density vs. mass would make more sense to me but perhaps your usage is idiosyncratic to sea level research?

So, the folks with their own theories deserve the attention they could attract on their own blogs, wouldn't you say?

But what about Spencer?

W,

I was prepared to write yet another long response but it turns out that Church and White already beat me to it. While they don't talk about the other issues with Rahmstorf, and Rahmstorf and Vermeer they do discuss the issue of lumping Steric and Volume together. In any event it is a terrific summary of the current state of science on sea level and I'm happy that this conversation caused me to go out and find it.

http://www.tos.org/oceanography/archive/24-2_church.pdf

Paul, yes I was incorrect to say volume. Of course volume changes, that is the whole point. I meant mass. Synapse misfire. Thanks for the correction.

I think the cooling of the ocean surface by evaporation would about balance increases in ocean temperature from other forcings. As such little can be gained from the poor ocean temp data and trends may be an ink blot test -- you see what you want.

Church and White:
"In summary, although semi-empirical
models warn that larger rises in sea level
than suggested by current process-based
models may be possible, they should be
used with caution until there is adequate
evaluation of and accounting for
the above concerns....
... It is important to recognize that
there are important thresholds, such
as those leading to ongoing melting of
the Greenland Ice Sheet and meters of
sea level rise. These thresholds could be
crossed in the second half of the twenty-
first century if greenhouse gas emissions
continue unabated."

Shorter:
Their models are wrong;
their models are useful.

Hank,

I'm sorry but I disagree with your summary. The ellipses tell the story. The first part of your quote is from the section discussing the semi-empirical models. That section finishes by saying that they should viewed with caution unless numerous concerns are corrected (which is a gentle way of saying they are wrong). The second portion of your quote was not from the section about semi-empirical models but was from the summary of the whole paper. This is a cautionary note about sea level rise in general and has nothing to do with the semi-empirical models.

Nicolas Nierenberg - Ok, I just wondered if your usage was standard.

On semi-empirical models I agree current efforts are of questionable use for 21st Century projections. It's not clear exactly what they can tell us about future climate-SLR relationships given that the dominant sources of sea level change are and will be shifting gears. For example mountain glacier & ice cap melt are a source of something like 25% of recent SLR but can't possibly represent 25% of a 1.5-2m rise since the available mass can only provide about 30cm in total.

I wonder to what extent calculations could be improved by focusing on empirical relationships between temperature and the major sources, rather than just resultant sea level. I guess that would be difficult due to lack of reliable long-term data.

These models might hold over periods of relatively little change like the last two thousand years but the climate projections for the next century are suggestive of a completely different mode of operation so it's not clear how much can be learned from the recent past. Having said that I actually think they might turn out to be about right because the implicit underestimation of ice sheet change and overestimation of glacier melt/thermal expansion could cancel each other out.

I looked carefully at that text when I edited the quote.

I could have pasted the entire section of text in.

I suggest people actually read the source.
I don't think your interpretation is right.
You're just doing the "uncertainty" rag about
the work done with previous and current models.

It's not bad work and the authors didn't do it wrong.

You've read enough to understand that these models are being improved slowly.

Improving something like this doesn't mean the earlier work is wrong or that the authors did it wrong.

They're approximations. They're improvements on prior work.
Greatbatch, R.J. (1994), A note on the representation of steric sea level in models that conserve volume rather than mass, J. Geophys. Res., 99, C6, 12767-12771

Hank,

I am not objecting to all models. I am objecting to the semi-empirical models by R+V, which if you read this text you can see are in a category by themselves.

And by the way you didn't "edit the quote." You strung two completely different quotes together to make it sound like they were talking about the same thing.

In any event I also suggest people read the original. As I mentioned before it is quite a good summary of the current state of the science of predicting sea level rise.

In one of my favorite parts they correct the mistaken impression that the IPCC ignored dynamic ice flow changes, or that they didn't include it in their estimates.

NN -- "In one of my favorite parts they correct the mistaken impression that the IPCC ignored dynamic ice flow changes, or that they didn't include it in their estimates."

Climate Change 2007: Synthesis Report

page 45:
"Model-based projections of global average sea level rise at the end of the 21st century (2090-2099) are shown in Table 3.1. For each scenario, the mid-point of the range in Table 3.1 is within 10% of the TAR model average for 2090-2099. The ranges are narrower than in the TAR mainly because of improved information about some uncertainties in the projected contributions.12 The sea level projections do not include uncertainties in climate-carbon cycle feedbacks nor do they include the full effects of changes in ice sheet flow, because a basis in published literature is lacking."

Nicholas, the two paragraphs from which I quoted excerpts are not unrelated.

Read the text in between; they make the connection clear:

"... major deficiencies in our understanding remain, and current projections still cover a broad range of values regardless of emission scenarios....
It is important to recognize that there are important thresholds ...."

Shorter: These models don't support confidence in the possibility of running up to just before the edge of trouble and stopping short.

PS, nonlinear doesn't always mean less.

Church's examples:

"Two such nonlinear scalings are the reduction of glacier area as the glaciers contract and a reduction in the efficiency of ocean-heat uptake with global warming. Again, both of these physical effects would reduce the semi-empirical model projections."

But those assumptions aren't the only possibilities being discussed in recent publications:

"the quasi-linear relation fails as soon as global warming
starts to decelerate, i.e. around 2100 for RCP8.5, and some
time earlier for the lower scenarios. As suggested by Ver-
meer and Rahmstorf (2009), validity of semi-empirical pro-
jections of sea level change based on this relation might be
extended by taking rapid adjustment processes into account."
http://www.earth-syst-dynam.net/2/25/2011/esd-2-25-2011.pdf

Why, there are even papers out there suggesting we may live in the best of all possible climate systems and that it can't get much worse (the "optimists and pessimists agree" perspective):

http://www.fisica.edu.uy/~barreiro/papers/BarreiroCherchiMasina2011.pdf

"... results indicate that the climate warms only if the OHT increase does not exceed more than 10% of the presentÂday value in the case of a strong cloudÂSST feedback and more than 25% when this feedback is weak. Larger OHT increases lead to a cold state where low clouds cover most of the deep tropics increasing the tropical albedo and drying the atmosphere. This suggests that the presentÂday climate is close to a state where the OHT maximizes its warming effect on climate and pose doubts about the possibility that greater OHT in the past may have induced significantly warmer climates than that of today."

That sound familiar?

J Bowers,

I didn't say that the summary included it, I said that the IPCC report included it. Check out table 10.7 page 820, and the bottom of page 821. "We also evaluate the contribution of rapid dynamical changes under two alternative scenarios." Under one scenario the estimates in 10.7 are decreased by .02m. In the second case they add .1 to .2 m to the estimates in 10.7. So while there is even more uncertainty in these figures they increase the top of the envelope to about .8m when looking at the A1F1 scenario. This is the figure in the Church and White paper.

Hank,

Again the part you have added is from the summary. It doesn't refer to the semi-empirical models. This is the summary of a long paper. Just because the part on semi-empirical models appears at the end of the paper doesn't mean that the summary really just refers to them.

> doesn't mean the summary really just refers to them

No, of course not.

But "them" are the only "improved ice sheet models" discussed.

First, Church et al. discuss the work done up til now.

They say

"The sea level projections of the IPCC WG1 Fourth Assessment Report (AR4, completed in 2007) were obtained by applying methods available in the preceding year to the latest results ....
Here, we give an overview of the IPCC AR4 projections and the limitations of these projections. We then discuss progress since the AR4 and prospects for improved global and regional projections. We also offer some cautionary comments on the use of semi-empirical models without adequate understanding of their potential limitations."

Then they go through the various areas studied.

Then they clarify the problem as of the last IPCC report:

"... during the twentieth century, our lack of ability to adequately close the sea level budget over decadal periods, and the observation that sea level is currently rising near the upper end of the IPCC projections has led to concern that the IPCC projections for the twenty-first century may be underestimated (Rahmstorf et al., 2007; Figure 1b). This concern has, in turn, led to the development of semi-empirical models (Rahmstorf, 2007; Vermeer and Rahmstorf, 2009; Grinsted et al., 2010) in an attempt to bypass our lack of process understanding. These semi-empirical models scale observed sea level rise to some other physical parameter such as global averaged temperature or radiative forcing. They give higher rates of rise and a wider range of projections (about 50â180 cm) by 2100.
A number of concerns have been raised about these semi-empirical projections.... "

Then they discuss ways to improve the new approach.

"This is how it works: you put your model out there in the coliseum, and a bunch of guys in white coats kick the shit out of it. If itâs still alive when the dust clears, your brainchild receives conditional acceptance. It does not get rejected. This time."

-- Peter Watts http://www.rifters.com/crawl/?p=886

Does this sort of remind you of the previous discussion? You know, the one about "there were no explicit death threats to climate scientists, just some un-nice words"

As in that case, there were

It's an interesting tactic, to say something ain't so, prove it to me and then just a little bit of a twist. Fortunately folk like Hank and Body are indefatigueable (or however you spell that). Normal people just walk away rolling their eyes

In # 36 chris points out that

There is nothing specifically concluded about the Earth response to radiative forcing other than a sentence in the abstract about satellite-based metrics in 2000-2010 departing in the direction of low climate sensitivity compared to model computations but that this can't be accurately quantitated in terms of feedbacks that determine climate sensitivity. That seems to be the basis of the Forbesian hyperbole.

To Eli this has the stink of Referee #3 sayin take that crap outta there

Heads up- #75 is a spammer for an alleged weight loss drink. These spammers who just randomly copy and paste a sentence or two from the thread have really increased in numbers lately, if not intelligence.

[Thanks. Gone -W]

> bodyslim ... intelligence

That's a spambot, aided and abetted by the secret masters of Scienceblogs.com; the link goes through a redirect that prevents you from checking the URL in advance so you can search for where else it's appeared. Don't click unless your malware filters are working. Ya never know. Neither does the secret management at scienceblogs.com, obviously.

Hank, some programmes (*cough* IE 7 *cough*) actually do show the URL if you hover your mouse over the link (just not in the bottom left)

[Infidel! You pollute the purity of the chrome/firefox fraternity with your dirty speech. We may be forced to burn you - for your own good, you understand. Wash your mouth out with soap and repeat the mantra "open source" until you are Enlightened -W]

I just found out that Google Chrome does the same...

(dare I ever return when using the heretic program?)

[In theory, I imagine, I could devise a browser-dependent site... but no -W]

Hank,

I'm sorry but the R+V stuff are specifically not "ice sheet models." That's what they mean by "bypass." They are an attempt to estimate sea level rise without ice sheet models. The new/updated ice sheet models are things like Charbit et al. 2009, and Ridley et al. 2010. (see page 138). For Glaciers and Ice Caps see page 136 and 137 for the set of new models in that area.

So if we agree that C+W were referring to "ice sheet models" in the last section, then we should agree that they were explicitly not referring to R+V (Semi-empirical models of sea level rise).

Cross Posted

Eli,

That is indeed a death threat. If that had been presented in the original discussion then I would have simply agreed. I didn't doubt that death threats had been made, I just said that the things being listed weren't death threats. That, in fact, is how the whole discussion began.

> So if we agree that C+W were referring to
> "ice sheet models" in the last section

But we don't agree on that.

See Chris's post above, which came in belatedly:

http://scienceblogs.com/stoat/2011/07/spencer_and_the_mystery_journa.ph…

"... do we simply all agree that Spencer and Braswell is sufficiently dismal that it doesnât need any more attention?"

Re #59 ("Apologies for leaving this in the queue for a while").

hmmm..If I knew I was going to been in a queue for three days, I would have brought my tent.

Chris,

Thanks for your comments. I will go back and look at the other models they used for out of sample testing in V+R, it was my impression that the models were Steric only but that may be wrong. On the other hand the out of sample testing reported in R2007 was just wrong.

Smoothing was never my primary argument, but smoothing before regression is difficult to explain. Why not just use the underlying data? But in any event it was the juxtaposition with Gavin's comment about the Spencer paper that caused me to raise the issue. In any event it doesn't seem that that is the strongest objection to that paper either.

In the end Church and White essentially come to the same conclusion that I did. The semi-empirical models are not a good approach, at least as implemented so far. The best approach is to separate the modeling of Steric from the modeling of mass changes. They are simply too different to put into one simple model.

You note that it would be foolish not to make provision for the sea level rise projected in V+R and refer to the Dutch Commission. But this is the whole problem. It is foolish to make contingencies because of one or two flawed outliers. The dutch commission estimates were in fact based on R2007. They should be using the estimates of experts on sea level when doing this kind of work. The current IPCC report is one example of that. The paper that I referenced by Church and White is another example.

[There I think there is room for disagreement. AR4 may not be so useful for those estimates, because of what they excluded -W]

Hank,

Chris's post is excellent, but it didn't have anything to do with whether V+R is an ice sheet model, it isn't. I'm sorry we disagree about something that is so straight forward.

Nicholas, I disagree with how you characterize progress.

You're claiming Church rejects semi-empirical models; Church is saying that's an attempt to improve on prior modeling, and needs further improvement, because tipping points are close enough to surprise us before we can plan what to do.

I'm not going to argue about whether to something an ice sheet model, or a sea level model, or an ocean temperature model -- you can't separate then neatly.
http://www.pik-potsdam.de/sealevel/en/data.html

"You note that it would be foolish not to make provision for the sea level rise projected in V+R and refer to the Dutch Commission. But this is the whole problem. It is foolish to make contingencies because of one or two flawed outliers. The dutch commission estimates were in fact based on R2007. They should be using the estimates of experts on sea level when doing this kind of work. The current IPCC report is one example of that. The paper that I referenced by Church and White is another example."

Nicholas, if any of that were remotely true then it might be a "problem". However your information is incorrect.

The Dutch Delta Commission recommendations were based on a commissioned report involving consulation "of experts on sea level" to address realistic bounds of high sea level scenarios that are obviously of concern when real-life planning issues are at stake. These included Hans von Storch and John Church (and Rahmstorf) who are co-authors on the report along with 18 other experts on sea level [see P. Vellinga et al (2009) Exploring high-end climate change scenarios for flood protection of the Netherlands.]

And the recommendations for 2100 sea level rise were certainly not based on R2007. They were based on the re-estimation under high end scenarios and summation of individual contributions from thermal expansion, Antarctic and Greenland ice sheet contributions, glaciers etc. just like in the IPCC AR4.

And nor did I say or imply that "it would be foolish not to make provision for the sea level rise projected in V+R.." I described the fact that there are now several studies based on semi-empirical models and consideration of ice-sheet dynamics that support higher likelihood of greater sea level rise by 2100 and stated that "â¦we would be foolish not to make contingency for these.". Thatâs the point. In addressing the likelihood of future scenarios we take all of the information at hand into account. That's what the Dutch Delta Commission are doing. It's a particular reality for them and they have an interest in addressing the issue maturely (and with a focus that those who currently find it convenient to accommodate misrepresentations of science in support of tedious agendas, may eventually be pursuaded to adopt.)

...which brings us back to Dr Roy W Spencer and his dismal paper..

Chris,

I went back and looked at V+R in PNAS. I was right. In all cases they validated against Steric only models. Notice the term "Modelled (sic) (no ice)" in the charts. A model of sea level excluding ice is by definition not taking into account mass changes. This makes sense because they trained their parameters on a Steric only model.

"The parameters are fitted to the global temperature and sea-level output from the climate model for 1880â2000, resulting in a = 0.080 ± 0.017 cm·Kâ1·aâ1, b = 2.5 ± 0.5 cm·Kâ1, and T0 = â0.375 ± 0.026 K (temperature relative to the reference period 1951â1980)"

So they proved that using temperature they can explain Steric sea level increase when compared to models that also use temperature to predict Steric sea level. Not a very interesting result.

You will also note that buried in their addendum is the fact that using the first half of the actual data their model does a lousy job of predicting the second half. Take a look at the coefficients to see that.

On the Dutch commission report I see that you are correct and I was wrong. I can't figure out where I remember seeing the direct footnote to Rahmstorf 2007. Maybe I'm mixing it up with the "Copehnhagen Diagnosis." BTW it is perhaps ironic that in this assessment the Rahmstorf model is only used in an assessment of potential Steric sea level increase. In this mode it had a minimal impact on the projections which are mainly increased from the IPCC by dramatically different assumptions about the contributions from Antarctica.

... which brings us back to Dr. Roy W. Spencer and his dismal paper....

Anyone know if Spencer is a Christian Reconstructionist or Dominionist, and whether the Cornwall Alliance is of that evangelical brand? Genesis 1:28 seems to fit their views on environmentalists very nicely, who the CA recently declared to be the antichrist. Fits the bill for the Teavangelical Taliban's undeclared war on science, too. I'm not referring to all evangelicals, by the way.

Found an answer to my own question about the Cornwall Alliance. They are guided by the Dominion Mandate, I think first declared by Henry Morris. Ergo, it's plausible that Spencer and McKitrick, with ties to the Cornwall Alliance, also follow the same beliefs.

Turns out Rick Perry and Michele Bachmann are also Dominionists. Daily Beast: A Christian Plot for Domination?

Does this finally bring us back to Dr. Roy W. Spencer and his dismal paper?

Yes, Hank, I think it does.

A very interesting development over at Remote Sensing - The editor in chief has resigned because of SB2011.

Alas, poor Woy, we knew him well...