Global warming - what to think ???

[BeltwayBandit]

A quick note for this discussion. This is not a comment about the research described above, but a commentary on abstracts. Abstracts are interesting in that they provide insight into the most new and breaking research. They should not, however, be given the same weight as journal articles as they are "works in progress" and are almost never subject to peer review. The research of many abstracts never ends up getting published.

 

This abstract and presentation is based on an article in the Journal of Geophysical Research. I'm currently trying to find a public copy, to no avail.

 

http://www.agu.org/pubs/crossref/2006/2005JD006118.shtml

[cabrerad]

This abstract and presentation is based on an article in the Journal of Geophysical Research. I'm currently trying to find a public copy, to no avail.

 

http://www.agu.org/pubs/crossref/2006/2005JD006118.shtml

 

 

Good enough. I understand the difficulty in finding public copies of some of these papers.

 

Article link-Canada, National Post

 

 

Some extra commentary on point #1 ("Statistics Needed") raised in the series quoted above.

 

The original paper (published in the social science journal "Energy and the Environment") by McIntyre and McKitrick (lets call them MM) critical of the the "hockey stick" graph in a Nature paper by Mann, Bradley and Hughes (lets call them MBH) questions the analysis used to come up with said graph. The analysis in question is called a Principal Components Analysis (PCA). In a nutshell a PCA analysis is a data transformation that reduces mulitdimensional data sets into components (the principal components) that describe the variance in the data. It is useful in that it allows you to focus on only those variables (principal components, PC1 through whatever) that explain most of the variance in the data (say you pick PC1-5 because PC6-12 are not significant and don't explain the data well).

 

To assist in visualizing this analysis (this gets a wee bit technical, but I am including it for completeness), think of a bunch of points on an scatter (XY, both variables random) graph. Say you need to draw a line that best fits the points on the graph. Now how do you describe your line and the confidence (think error bars) in this line? You can't do it with error bars/confidence intervals because both variables are random and this violates the assumptions of linear regression (It would also difficult to decide which variable to regress on the other). What you can do is define a confidence region, which is described by an ellipse. This ellipse describes the variation in both variables (unlike a regression) and has both a major (the line) and minor axis (perpendicular to that line) that best fits the data (these are called eigenvectors). The shape of the ellipse describes the variation about the line and is represented by the ratio of the lengths of the long and short axes (these are called eigenvalues and represent variance). A good correlation has a really long principal axis and a short minor axis. A bad correlation would have axes close in length (basically a circle) and would not explain the variability in the data well at all. Ok now with that done, imagine a data set in multidimensional space (more than two variables) and you are looking at PCA. Now, however, you are looking at calculating a principal axis through a hyperellipsoid (PC1). Then you look for the second major axis through that hyperellipsoid and you have PC2 and so and so forth.

Basically in PCA you take your data set and normalize the mean of the data set to a given value, say 0. The analysis looks for deviations from this mean and the principal componets are variables in the data set that best explain this deviation. You could also look at an interger or interval of a priori interest to your research to see if there is deviation from that value.

This is what MBH did in setting the mean over the 20th century to zero (to catch deviations from this period). Here is where MM had a bone to pick. They argued that the entire period should be set with mean O, not just the 20th century. In the MM analysis, the "hockey stick " component came up as PC1 in the MBH analysis and PC 4 in the MM analysis. MM decided that to exclude PC4 (and 3 and 5) and called them artifacts. This was a mistake. Subsequent analysis using MM's methodology has found all three excluded components to be statistically significant, so they should not have thrown them out. In addition, the MM analysis does not fit the entire data set when only the first two PC's are considered. When PC's 3, 4, and 5 are included it matches both the analysis of MBH and the data set. The only way the MM analysis fits the data set is to remove two of the tree ring data sets. You cannot omit data sets from an analysis without darn good reason to do so (statistical outliers are an example). This is an important check on the fit of the data and MM fall short here also.

 

Here is another paper, more recent than those of MM yielded similar results to the MBH paper accounting for some of the criticisms by MM and using a second type of analysis not using PCA analysis of the data sets in question:

 

Rutherford et al 2005

 

Pages 13 and 32 discuss some of the claims made by MM.

 

 

Now on to Wegman:

 

I finally got a hold of what his critique of the MBH analysis was. It also focused on the PCA analysis, specifically on centering of the PCA analysis. He was basically asked if the MM criticism had statistical merit. He stated that it did. That's not really much of a revelation, no one actually contests that centering affects PCA analysis. Here is a key point: As I described above, the method of centering the data did not affect the shape or statistical significance of the hockey stick in explaining the data. The only thing it did was switch around the principal compontents (all of which were statistically significant anyways). So centering had no affect on the end result. Moreover, the centering chosen by MBH had relevence to what they wanted to test. Wegman's criticism therefore seems to indicate a lack of understanding of the hypothesis being tested. Wegman made the point that there should be more talk between statisticians and climate scientists due to these questions and yes there probably should be. The key is working together, because a statistican alone is blind to what may be an important variable or a priori assumption in a given field.

Edited February 14, 2007 by cabrerad

[Aquariareview]

HOUSE HEARING ON 'WARMING OF THE PLANET' CANCELED AFTER ICE STORM

HEARING NOTICE

Tue Feb 13 2007 19:31:25 ET

The Subcommittee on Energy and Air Quality hearing scheduled for Wednesday, February 14, 2007, at 10:00 a.m. in room 2123 Rayburn House Office Building has been postponed due to inclement weather. The hearing is entitled

Edited February 14, 2007 by Aquariareview

[cabrerad]

HOUSE HEARING ON 'WARMING OF THE PLANET' CANCELED AFTER ICE STORM

HEARING NOTICE

Tue Feb 13 2007 19:31:25 ET

The Subcommittee on Energy and Air Quality hearing scheduled for Wednesday, February 14, 2007, at 10:00 a.m. in room 2123 Rayburn House Office Building has been postponed due to inclement weather. The hearing is entitled

[Guest Larry-T]

Almost all the published materials designed to cast doubt on the reality of global warming and the fact of human activity as a contributing factor are based on one of two foundations:

 

1. They use short-term data to attempt to divert attention from the bigger picture.

 

2. They fasten on one minor point of contention (why doesn't this particular glacier behave as predicted?) and try to use it to transfer doubt to the bigger picture.

 

Both of these are familiar to someone who has spent decades defending evolution against the well-funded attacks by creationists. Replace the Discovery Institute with the American Enterprise Institute and Exxon Mobile and you have a familiar picture.

[BeltwayBandit]

Almost all the published materials designed to cast doubt on the reality of global warming and the fact of human activity as a contributing factor are based on one of two foundations:

 

1. They use short-term data to attempt to divert attention from the bigger picture.

 

2. They fasten on one minor point of contention (why doesn't this particular glacier behave as predicted?) and try to use it to transfer doubt to the bigger picture.

 

Both of these are familiar to someone who has spent decades defending evolution against the well-funded attacks by creationists. Replace the Discovery Institute with the American Enterprise Institute and Exxon Mobile and you have a familiar picture.

 

I thought we were going to stick to a scientific debate, with peer reviewed papers. This reply is a generalized ad hominem attack with absolutely NO scientific foundation. So, care to try again to explain scientifically why we should trust results from an unvalidated model? Remember, your answer should be supported by peer reviewed literature.

[Guest Larry-T]

So, care to try again to explain scientifically why we should trust results from an unvalidated model? Remember, your answer should be supported by peer reviewed literature.

 

Again, you ask non-sequitor questions designed to divert attention from the wealth of data and the scientific consensus. The vast majority of scientists and the overwhelming majority of peer-reviewed articles agree that global warming is real, it is moving fast, and human activity has played a role in this process. The burden of proof is on those who would contest the scientific consensus. Models are not "valid or invalid." They are in a constant state of refinement. As the models used in the global warming studies become more and more refined, they point more and more to agreement with the current consensus.

[cabrerad]

sigh....

 

in case it was not noticed post 52 is a response to some of the literature out there....

[BeltwayBandit]

Again, you ask non-sequitor questions designed to divert attention from the wealth of data and the scientific consensus. The vast majority of scientists and the overwhelming majority of peer-reviewed articles agree that global warming is real, it is moving fast, and human activity has played a role in this process. The burden of proof is on those who would contest the scientific consensus. Models are not "valid or invalid." They are in a constant state of refinement. As the models used in the global warming studies become more and more refined, they point more and more to agreement with the current consensus.

 

No, I ask pointed and direct questions about the actual state of the science. Your appeals to authority while sounding impressive, are not the foundation of a valid scientific debate. Your inability to actually discuss the science, after specifically asking to do just that is rather dubious. Surely, you can find a peer reviewed article that would justify the use of unvalidated models. It shouldn't be that hard.

 

sigh....

 

in case it was not noticed post 52 is a response to some of the literature out there....

 

Yes, I am looking at that one now. My first comment is to note that Mann, Bradley and Hughes are all cited as co-authors of the paper. Hardly an independent verification.

[cabrerad]

#2 Part two of deniers:

 

"Warming is real -- and has benefits -- The Deniers Part II"

 

 

This part of the article profiles another "denier" (Dr. Richard Tol who released a critique of the The Stern Review of the Economics of Climate Change (Stern et al., 2006).

 

Here is a link to the Stern Report

 

 

Stern Report

 

Here is a link to Dr. Richard Tol's critique:

 

Tol essay

 

 

 

From my read of this essay, Dr. Tol contends with two points made in the Stern essay:

 

In this comment, I focus on two conclusions. Firstly, the Stern Review argues that

Edited February 15, 2007 by cabrerad

[Guest Larry-T]

No, I ask pointed and direct questions about the actual state of the science. Your appeals to authority while sounding impressive, are not the foundation of a valid scientific debate. Your inability to actually discuss the science, after specifically asking to do just that is rather dubious. Surely, you can find a peer reviewed article that would justify the use of unvalidated models. It shouldn't be that hard.

 

Actually, you haven't discussed science at all. You've cherry-picked ideological arguments from other sources and used them to raise doubts without using any science whatsoever. Are you a climatologist? What are your credentials? I don't claim to be one. My main field of study was Biology, but I haven't worked in a lab for many years. I know enough about the methodology to read the papers and see that those printed in journals such as Science and Nature have real substance, while those in the popular press and those coming from ideological think tanks do not. By focusing the discussion on quibbling about models you are essentially engaging and avoidance. Here's the main point - the overall trend of the amount of heat energy in the oceans and atmosphere has increased at an rate never before recorded. This closely tracks the amount of greenhouse gasses released through human activity. Localized and short-term swings have nothing to do with the basic facts of the matter. If we regard the data as accurate, the logical response is to reverse the increase in greenhouse gas releases and see if this slows or even reverses the trend. The main objections to this boil down to the following:

 

Since we don't have 100% certainty that this will work, we shouldn't go to the expense of trying. This basically says "Money is more important than anything else."

 

Unfortunately, even if we reversed the greenhouse gas release trend worldwide, it would take at least a decade, and probably longer, before enough data could accumulate to prove results. This is a longer attention span than most societies and any politician. This kind of "100% certainty before any regulation can even be considered" will lead to even worse consequences, but the short-sighted, greedy followers of the MBA culture don't care.

 

The greatest argument in favor of the scientific consensus being correct is the behavior of the insurance industry. Many of the largest reinsurance companies (those who control the overall behavior of the industry) are pulling out of risky (coastal) property insurance. This has a domino effect down to the retail level. Yesterday, State Farm Insurance announced that it will sell no new homeowners and commercial insurance in the state of Mississippi. While they cite a legal and political climate, this comes in addition to many insurance companies (including State Farm) refusing to sell or even renew existing policies in coastal Louisiana and in Florida. The underwriters can read maps and they are getting good technical advise on the future of these areas.

 

As far as I am concerned the SCIENTIFIC debate on global warming is over. The debate today, that that over evolution, is one of politics, economics, and doctrine. It has nothing to do with science.

[BeltwayBandit]

As far as I am concerned the SCIENTIFIC debate on global warming is over. The debate today, that that over evolution, is one of politics, economics, and doctrine. It has nothing to do with science.

 

Well alrighty then. That definately puts you in the ideological camp. Why say you want to discuss the science when you clearly don't. You have obviously confused ideological faith with science. Have a nice day.

 

I'll end with a few quotes from Einstein for you:

 

On your blind faith and lemming like mentality:

"He who joyfully marches in rank and file has already earned my contempt. He has been given a large brain by mistake, since for him the spinal cord would suffice. "

 

On your attempt to divert the scientific process by saying that the debate is over:

"To raise new questions, new possibilities, to regard old problems from a new angle, requires creative imagination and marks real advance in science."

 

And finally, on the power of one person in science:

"No amount of experimentation can ever prove me right; a single experiment can prove me wrong. "

[BeltwayBandit]

This is what MBH did in setting the mean over the 20th century to zero (to catch deviations from this period). Here is where MM had a bone to pick. They argued that the entire period should be set with mean O, not just the 20th century. In the MM analysis, the "hockey stick " component came up as PC1 in the MBH analysis and PC 4 in the MM analysis. MM decided that to exclude PC4 (and 3 and 5) and called them artifacts. This was a mistake. Subsequent analysis using MM's methodology has found all three excluded components to be statistically significant, so they should not have thrown them out. In addition, the MM analysis does not fit the entire data set when only the first two PC's are considered. When PC's 3, 4, and 5 are included it matches both the analysis of MBH and the data set. The only way the MM analysis fits the data set is to remove two of the tree ring data sets. You cannot omit data sets from an analysis without darn good reason to do so (statistical outliers are an example). This is an important check on the fit of the data and MM fall short here also.

 

 

I would be curious what method was used to determine the 'statistical significance' of the excluded datasets. And I do find it curious that Wegman, an independant statistician, agreed with the statistical conclusions of MM. You also brought up a good point about proxy selection. A large portion of the debate over MBH reconstructions is the use of the Bristlecone pine proxy records. I'll post up a bit more about that later, its another interesting discussion.

 

Here is a piece from McIntyre regarding the debate over verification statistics:

 

While climatologists may

[cabrerad]

My first comment is to note that Mann, Bradley and Hughes are all cited as co-authors of the paper. Hardly an independent verification.

 

The citation was intended to show a strong rebuttal to the critiques of the MM and the authors (not surprisingly) include Mann, Bradley and Hughes [side note here, just because a given author is on a paper is not sufficient enough reason to doubt the science contained within]. Here are some additional peer reviewed papers that yielded results supporting anthropogenic effects on increased temperature in the late 20th century (many of which do not include Mann, Bradley or Hughes).

 

Bauer, E., M., Claussen, and V. Brovkin, Assessing climate forcings of the earth system for the past millennium, Geophys. Res. Lett., 30 (6), doi: 10.1029/2002GL016639, 2003.

 

Bertrand C., M.F. Loutre, M. Crucifix, and A. Berger, Climate of the Last millennium: a sensitivity study, Tellus, 54(A), 221-244, 2002.

 

Briffa, K.R., T.J. Osborn, F.H. Schweingruber, I.C. Harris, P.D. Jones, S.G. Shiyatov and E.A. Vaganov, Low-frequency temperature variations from a northern tree-ring density network. J. Geophys. Res., 106, 2929 2941, 2001.

 

Brovkin, V., A. Ganopolski, M. Claussen, C. Kubatzki, and V. Petoukhov,

Modelling climate response to changes in land use, Global Ecology and

Biogeography, 8(6), 509.517, 1999.

 

Crowley, T.J., Causes of Climate Change over the Past 1000 Years, Science, 289, 270-277, 2000.

 

Crowley, T.J., and T. Lowery, How Warm Was the Medieval Warm Period?, Ambio, 29, 51-54, 2000.

 

Esper, J., E.R. Cook and F.H. Schweingruber, Low-frequency signals in long tree-line chronologies for reconstructing past temperature

variability, Science, 295, 2250-2253, 2002.

 

Gerber, S., F. Joos, P. Br

Edited February 15, 2007 by cabrerad

[BeltwayBandit]

In doing some other digging I re-discovered this excellent opinion piece by von Storch and Stehr. For those that don't know von Storch is a climate scientist and is familiar with the actual state of the science. Definately worth a read.

 

http://www.spiegel.de/international/spiege...,342376,00.html

How Global Warming Research is Creating a Climate of Fear

By Hans von Storch and Nico Stehr

 

 

"Silencing dissent and uncertainty for the benefit of a politically worthy cause reduces credibility, because the public is more well-informed than generally assumed. In the long term, the supposedly useful dramatizations achieve exactly the opposite of what they are intended to achieve. If this happens, both science and society will have missed an opportunity. "

 

The citation was intended to show a strong rebuttal to the critiques of the MM and the authors (not surprisingly) include Mann, Bradley and Hughes [side note here, just because a given author is on a paper is not sufficient enough reason to doubt the science contained within]. Here are some additional peer reviewed papers that yielded results supporting anthropogenic effects on increased temperature in the late 20th century (many of which do not include Mann, Bradley or Hughes).

 

 

Hang on lets put MBH to bed first, the issue MM had with MBH wasn't necessarily weather or not they detected anthropogenic effects or not. MM's issue was the misapplication of statistical methods used by MBH, the improper use of certain proxy series (notably bristlecone pines), and in general the lack of proper data archiving to comply with publication standards. It is irrelevant that other papers achieved similar results. The fact of the matter is, MBH achieved their results through a flawwed process. The MBH reconstructions stand or fall on their own merit. Just because you get the "right" answer with bad math dosn't mean your method was right, this is similar to the problem I pointed out earlier with the computer models. The "right" answer is only part of the issue, you have to get the intermediate steps correct in order for your work to be valid.

 

von Storch also found MBH to be flawwed, from the article above:

"In an article we published in the professional journal "Science" in October 2004, we were able to demonstrate that the underlying methodology that led to this hockey stick curve is flawed."

 

As for the other studies you listed, many of them use the same proxies that MBH used in their reconstruction. In effect, they are not independant reconstructions they are more like a rearrangement of the deck chairs on a famous ship.

[cabrerad]

Hang on lets put MBH to bed first, the issue MM had with MBH wasn't necessarily weather or not they detected anthropogenic effects or not. MM's issue was the misapplication of statistical methods used by MBH, the improper use of certain proxy series (notably bristlecone pines), and in general the lack of proper data archiving to comply with publication standards.

 

Please refer back to post 52 where I address their critique of the statistical methods and explain what PCA is and why the MM accusation falls short. In addition to my own explanation, I also provide a peer reviewed publication that addresses the criticism. Please note also that there seems to be only one peer reviewed (I am assuming peer reviewed as I have honestly never heard of Energy and the Environment) publication from MM . Part of their critique is also based on another manuscript that was submitted as a Brief Communication to Nature. The manuscript was sent out to reviewers who recommended it be rejected. MM claim it was rejected due to lack of space which is not even listed as a reason for rejection by Nature. On the other hand Mann has several peer reviewed articles about this issue. This leads one to question; if these critiques are so valid, why have MM not been able to publish them?

 

As for the claim that they lacked proper data archiving to comply with publication standards, this is untrue and evidenced by the fact they published in many journals (therefore meeting thier standards):

 

Please reference:

 

Corrigendum: Global-scale temperature patterns and climate forcing over the past six centuries

Michael E. Mann, Raymond S. Bradley and Malcolm K. Hughes

Nature 430, 105(1 July 2004)

doi:10.1038/nature02478

 

for the most updated information

 

 

 

It is irrelevant that other papers achieved similar results.

 

Not true. It shows several different studies have indicated anthropogenic effects on global temperature in the late 20th century. This is a very cogent to the discussion at hand.

 

The fact of the matter is, MBH achieved their results through a flawwed process.

 

As this statement is only backed up by one paper (addressed by the literature and which I expounded upon) and a very narrow interpretation of Wegman's comments, I counter there is not sufficient evidence for any such claim.

 

 

The fact of the matter is, MBH achieved their results through a flawed process. The MBH reconstructions stand or fall on their own merit. Just because you get the "right" answer with bad math dosn't mean your method was right, this is similar to the problem I pointed out earlier with the computer models.

 

 

Again, an interpretation that lacks sufficient support and is also not accurate. The MM critique the centering of the PCA analysis only provides an alternate interpretation of the centering that does not take into account the data being looked at and also neglects to address selection criteria for the principal components. To counter the MM claim that using non-centered PCA is bad math, you can reference Principal Component Analysis by Ian Jolliffe.

 

 

Principal Component Analysis, I.T. Jolliffe

 

I'm not sure MM can claim to have published a textbook on the statistic.

 

 

 

The "right" answer is only part of the issue, you have to get the intermediate steps correct in order for your work to be valid.

 

agreed

 

 

As for the other studies you listed, many of them use the same proxies that MBH used in their reconstruction.

 

and ground surface temperatures, continental borehole data, hemispheric surface air temperature trends, ice cores, corals etc. These many studies can't be simply discounted by saying they are using the same proxies that MBH used.

Edited February 15, 2007 by cabrerad

[quazi]

Global warming has finally been explained: the Earth is getting hotter because the Sun is burning more brightly than at any time during the past 1,000 years.

 

 

 

http://www.telegraph.co.uk/news/main.jhtml.../ixnewstop.html

 

 

 

I realize it may be hard to believe, but the SUN may be causing global warming. Now what do we do? Can some one just turn the Sun down a few degrees?

Edited February 15, 2007 by quazi

[modelrr]

http://www.americandaily.com/article/1128

that's what this article says too

[cabrerad]

Global warming has finally been explained: the Earth is getting hotter because the Sun is burning more brightly than at any time during the past 1,000 years.

 

 

 

http://www.telegraph.co.uk/news/main.jhtml.../ixnewstop.html

 

 

 

I realize it may be hard to believe, but the SUN may be causing global warming. Now what do we do? Can some one just turn the Sun down a few degrees?

 

 

 

 

This article also mentions the effects of CO2 and the critiques of a few scientists. And it is quite a stretch to say global warming has finally been explained due to one paper coming out when there are a heck of a lot of other papers out there (hint..hint.. read the previous threads)

and

 

QUOTE

http://www.americandaily.com/article/1128

that's what this article says too

 

 

This article is a pretty strong opinion piece with an inflamatory title and should be taken as such. The first sentence alone indicates an elementary lack of understanding of global climate change (unless the author knows this and just wants to fan the flames so to speak).

 

 

With that said, current climate models do take into account solar activity

Edited February 15, 2007 by cabrerad

[BeltwayBandit]

You forgot the paper by von Storch and the review by Wegman which found the MBH papers to be flawed.

 

I'll also add, Mann is an editor and reviewer for Nature. Its akin to the situation of the fox guarding the henhouse.

[modelrr]

The title indicates this is an opinion piece (not to mention the publication). The first sentence alone indicates an elementary lack of understanding of global climate change.

I know.. I just thought it was an interesting idea

[cabrerad]

You forgot the paper by von Storch and the review by Wegman which found the MBH papers to be flawed.

 

No, I didn't forget., I was addressing MM first and want to read the paper by Von Storch before replying.

 

I'll also add, Mann is an editor and reviewer for Nature. Its akin to the situation of the fox guarding the henhouse.

 

 

I'm sorry, that just isn't true and is simply conjecture on your part. The journals take exceptional care to keep out submitting authors who also happen to be editors of a journal from the review process. In fact it would most likely be more difficult for an editor to get a paper in the journal due the other editors wanting to avoid a perception of bias. Reviewers are selected from a group of the author's peers (i.e. there are lots and lots of reviewers for every journal) by the editor (which in this case is clearly not Mann). It is far from easy to have a paper accepted by Nature. Also, I can't find any reference to Michael Mann ever being an editor for Nature in the magazine itself (I'm looking at the editorial table of the current issue right now) or online.

 

Also, please note that the National Acadamies of Sciences had this to say about MBH:

 

There is sufficient evidence from tree rings, retreating glaciers, and other "proxies" to say with confidence that the last few decades of the 20th century were warmer than any comparable period in the last 400 years, according to a new National Research Council report. There is less confidence in reconstructions of surface temperatures from 1600 back to A.D. 900, and very little confidence in findings on average temperatures before then.

 

 

opening statement

 

 

an excerpt:

Let me summarize five key conclusions we reached after reviewing the evidence:

 

1. The instrumentally measured warming of about 0.6

[BeltwayBandit]

My mistake. Mann is an editor for Journal of Climate, and has only been a reviewer for some of the other journals.

[cabrerad]

My mistake. Mann is an editor for Journal of Climate, and has only been a reviewer for some of the other journals.

 

 

Not a problem. I'll give Von Storch a read soon.. but it's soccer tonight...gotta love those 10:40 pm indoor games...

[rsarvis]

Sorry if this open letter has already been posted. Interesting, though.