« Designed To Confuse | Main | Secret Write Off »

March 08, 2009

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

You claim to be both a skeptic and a critical thinker, yet you have completely bought into the nonsense spouted by the IPCC and some climatologists that we are undergoing anthropogenic global warming that threatens the planet. This, just thirty years after similar "experts" on climate were claiming that an ice age was just around the corner.

Either read the actual reports on climate objectively, or stop claiming to be a critical thinker and skeptic. Here's a good starting point: find any literature that validates the concept that the greenhouse effect (tested in a small, closed system) applies to the entire planet and its atmosphere. Good luck with that literature search, because no such studies have been done. Next, find literature that supports the idea that the impact of people on climate is greater than the impact of solar energy output changes on climate. Good luck again, because all you will find is junk science and mathematical climate modeling with weighting factors chosen arbitrarily. For a third challenge, read the most recent IPCC report and look at how badly the climate models performed. The "meta-model" did best, but it overestimated temperatures in the polar regions by six degrees centigrade! Despite that huge error, the IPCC used that model to predict devastating polar ice melting. A model that cannot predict polar temperatures should not be used to predict global warming.

Finally, quit arguing based on authority. It matters not who belongs to IPCC if the entire organization spouts nonsense. They have a vested interest in promoting panic, because it gives them jobs, publicity, and a sense of importance. Having a science PhD doesn't automatically turn one into an objective, trustworthy, and skillful scientist.

Why is it so, that
I see this kind of "GW denial debunking" style post it leaves some bad taste in my mouth? Be it on Phil Plait's site or here.
Guys, are you not applying double standards here? And that given that I am more inclined to believe in AGW than not, though post like that rise my doubts.

1) Care to clarify are those 70 not world's elite scientists, or just not worlds elite climate scientists?

2) If so it happens, that non expert in climate sicence makes valid scientific claims - should those claims be dismissed just because s/he is not a climatologist?

3) Why do you keep talking about people? There is no need for that if there is a solid evidence for AGW. Is it there?
From your recent post:
"For those interested, here is some. The links are not direct to peer reviewed papers (although there are links in some of the articles to peer reviewed work). Obviously it’s not comprehensive. But it’s a good place to start and you certainly can’t claim there is “none.” Unless you’re a denier."
Uhm. So that should be somehow convincing? And if I am not convinced, then I am a denier? How about the word "skeptic"? (For the record, I am neither. I just don't know, and you don't help me.)

4) In recent post you also talk about " a list of significant predictions of enhanced greenhouse gas warming that have been made and confirmed."
How about list of predictions that were not confirmed? Predict a lot and you will hit some, how do I know it was not a pure luck? And by the way, do those models put an "A" to "AGW"?

So my point is—if when you have strong evidence, by all means, bring it on. Until
then it is fine to have beliefs, but maybe it is not so right so call all skeptics "deniers", what do you think?
I don't deny anything, I just find your this and las posts about GW not convincing and substandard.


Dr. T, go back to dragooning kids to play your enormous piano; you're doing a very poor job as a troll.

Rimantas:

I'm assuming you actually want to learn about AGW, so:

1) What books have you read on it?
2) Scientific journals?
3) Lectures by climate scientists?

Given that I can propose some appropriate recommendations. It of course is quite plausible to answer "haven't read a book, a journal article, or heard a lecture."

I would *not* recommend to anyone that they try to start to learn climate science from blogs, even good ones. That's a good way to get confused, or worse.

Calling yourself Dr. T and spouting FUD makes me think of the worst insult I have ever typed:
"Fred Singer".

I feel really dirty and disgusted now. I'll have to take a shower.

Dr. T, I recently regressed NASA-GISS global annual land-sea temperature anomalies on ln CO2, Dust Veil Index, and four measures of solar activity -- TSI (Total Solar Irradiance, also known as the Solar Constant), Sunspot number, years since maximum, and years since minimum. I did the solar indices one or two at a time to avoid multicolinearity. Guess what? None of the solar indices was significant, and CO2 accounted for 76% of the variance. So I guess the human effect on global warming is greater than that of changes in solar activity, isn't it, at least for the last 128 years. So will you change your mind now?

Dr T - you do realise that the anti-AGW camp bases a lot of its opposition on the belief we are about to enter a terrible cooling phase don;t you... eg Archibald?

@fraggle, did you prove the causation, or just saw the correlation?
I don't think that someones guess is solid base for an opinion change.

Finally, quit arguing based on authority.

Try understanding what an argument from authority is.

More than 70 of the world’s elite scientists, economists and others specializing in climate issues will confront the subject of global warming.

Is an argument from authority.

Pointing out that they aren't the world's elite anything is not.

"While these two legitimate climate scientists express doubt about the dangers of climate change, they both acknowledge that the world is warming and that it is due to human activity, primarily greenhouse gas emissions."

I don't think so. Roy Spencer recently said "I now believe that most of the warming in the last 100 years was due to natural cloud variations caused by the Pacific Decadal Oscillation".

"Dr. T, I recently regressed NASA-GISS global annual land-sea temperature anomalies on ln CO2, Dust Veil Index, and four measures of solar activity -- TSI (Total Solar Irradiance, also known as the Solar Constant), Sunspot number, years since maximum, and years since minimum."

The standard retort to this type of analysis is that it doesn't control for coincidence. Instead of CO2 you could use "number of pirates" or many other factors.

You could include "year" as a variable in your regression model, but that assumes the coincidence of trends is given by a linear equation on "year."

What I've suggested is using a little-known technique for controlling coincidental trends called "detrended cross-correlation analysis." It works.

That, of course, would simply serve to prove a point. The models that are actually useful in making predictions don't assume CO2 directly influence temperature. It's a bit more complicated that this. Equilibrium temperature depends on climate forcings, logarithmically.

Skeptico - thanks for linking to my list.

Rimantis,
I've written up the people who attended last year's conference and who signed the resulting "Manhattan Declaration" as part of my larger list of climate scientists and others making claims about climate science:
http://www.eecg.utoronto.ca/~prall/climate
Look at the longer list (now past 2000 names, nearly 1500 ranked by cites and linked to homepages.) I've noted who served on the IPCC (all 619 AR4 wg1), who signed 'activist' declarations such as the Bali Climate Declaration or the 2008 statement from tne Union of Concerned Scientists (of whose 1700 signers I have only those I already came upon independently). I've also noted who signed any of ten "skeptics" declarations, from Leipzig in 1995 through the Manhattan Declaration last year.
Signers of skeptic statements show up very little in the top 100 or even 500 - down around 4%. The most highly cited climate skeptics, Freeman Dyson, Antonio Zichichi, and William Happer, got all their citations in their own fields (e.g. particle physics) and have no published research *on climate* of their own.
Of the top 100, over 60 either served on IPCC AR4 wg1 and/or signed an activist declaration, and three signed skeptics' statements. Looking through the remaining 30-some non-signers, you'll find no sympathy for inaction on GHG reductions - names such as Wally Broecker, Reto Ruedy, James Lovelock, Willi Dansgaard, Jorge Sarmiento, Andrew Lacis, and Ralph Cicerone. See any 'skeptics' here? Have a look at their work and let me know if you find anything.

Oh and while you're at my long list, scroll to the end of the rankings by citations (starting around #1425) and note the dense packing of self-declared climate skeptics who have either no peer reviewed publications at all, or fewer than four works, with zero to a handful of cites to these.
This is not an 'elite'.
P.S. I've never seen a normal scientific conference refer to the conferees as "elite" - maybe a hacker conference might claim to be 3733T, but...

Fake Experts? I’m not as interested in the college degree as in the facts. Dr. Hansen’s degree was in astronomy: does that make his papers on climate irrelevant? I find the AGW bloggers more inclined to appeal to authority than the skeptics. “The consensus is” “Shut up - the science is settled” etc.

The facts that Sketico fails to address include:
1. Proof of the positive forcing factor for CO2 of about 3, used in the climate models.

2. Explanation of why global warming has ceased for almost ten years despite the increase in CO2 levels. See http://rankexploits.com/musings/2009/giss-feb-reported-trend-since-jan-2001-still-negative/ and also note the second graph from that link. The previous response was to doubt Lucia’s ability. If that is the case, care to point out the mathematical error?

3. Mann’s hockey stick, that started the whole AGW scare, is wrong. See http://www.people.cornell.edu/pages/asu1/do-we-know-that-1990s-warmest.pdf
Those that claim it has been independently verified have missed the point that all papers coming up with this result are based on a few very suspect bristlecone temperature proxies, and in some cases the calculations (like Mann’s) were not robust.

4. Most of the land based temperature stations don’t meet their own organization’s specifications.
See http://www.surfacestations.org/ Most are sited where there is a strong chance of “urban warming” effecting the readings, and yet much is made of change of as little as 0.1C


It is probably worth viewing Dr. Lindzen’s short presentation on the politics of global warming, as this directly addresses the point make by Skeptico that started this thread. http://www.youtube.com/watch?v=OS-cLp1PEGQ&eurl=http://motls.blogspot.com/search/label/climate

To forestall the usual ad hominems, I have been following the subject closely for more than ten years and have read the IPCC report. Yes, I have technical background.

Explanation of why global warming has ceased for almost ten years despite the increase in CO2 levels

Simple: it hasn't ceased. Global warming has continued.

Parallel, for someone who has been following the AGW debate closely you must be a very inattentive reader.

See, for example: The Hockey Stick Holds Up, Nature Reports Climate Change
Published online: 18 December 2008 | Corrected online: 6 January 2009 | doi:10.1038/climate.2008.142


A follow-up to the infamous 1998 'hockey stick' curve confirmed that the past two decades are the warmest in recent history. Climatologist Michael Mann's contentious graph has become a symbol of the fierce debates on evidence for global warming, to the extent that an independent investigation into the study was performed at the request of US Congressman Joe Barton. The 2006 report that resulted from the Barton enquiry criticized Mann and colleagues for their reliance on tree-ring data from bristlecone pines as a proxy to reconstruct Northern Hemisphere temperatures over the past 1,000 years. Although their earlier work had been largely vindicated, in September the same team revised their global surface temperature estimates for the past 2,000 years, using a greatly expanded set of proxies, including marine sediments, ice cores, coral and historical documents (Proc. Natl Acad. Sci. USA 105, 13252–13257; 2008). The team reconstructed global temperatures with and without inclusion of the tree-ring records: without their inclusion, the data showed that recent warming is greater than at any point in at least the past 1,300 years; inclusion of tree-ring data extended this period to at least 1,700 years. According to the Christian Science Monitor: "It still looks a lot like the much-battered, but still rink-ready stick of 1998. Today the handle reaches further back and it's a bit more gnarly. But the blade at the business end tells the same story."

You say that global warming hasn’t ended. Do the slopes of the temperatures, from the two main data bases shown on the linked graphs, really look positive to you? graphs

Of course I expect it to continue warming about 0.6C per century as the long term trend, but the obvious point is that the climate model projections are too high.

Gosh I am tired of reading that idiocy. Are you under the bizarre impression that anyone thinks that temperature and CO2 should rise and fall in a 1:1 ratio? That there are no other factors? There are other factors. These are the same factors that denialist pretend to understand and pretend that climatologists don't know about.

For example, we are just getting out of the bottom of the 11 year solar cycle. not to mention that 1998 was a flukey spike in temperature, so using that as a starting point is extremely convenient for you, isnt it?

Your spouting off that repeatedly debunked nonsense is like when creationists say that evolution disobeys the 2nd law of thermodynamics, showing that they neither understand evolution nor thermodynamics.

Or do you also think that evolution disobeys the 2nd law of thermodynamics?

Parallel, for someone who has been following the AGW debate closely you must be a very inattentive reader.
Like many dogmatic AGW believers, you only want to look at and believe reports that confirm your beliefs. If you had been following the subject closely, or even just read the linked piece in my post, you would have seen that the facts do not support your contention.

Do the slopes of the temperatures, from the two main data bases shown on the linked graphs, really look positive to you?

Well, it looks more to me like trying to fit eight years of climate data to a linear model is pretty foolish, and proclaiming that this manufactured slope is negative and therefore global warming has ceased is just contrived nonsense. The second graph of your own source seems to make it fairly clear that global temperature is fluctuating but the overall trend is one of warming. The fact that we have been in a down-turn of that fluctuation doesn't change that.

Well, it looks more to me like trying to fit eight years of climate data to a linear model is pretty foolish,

How else do you suppose one calculates the rate of change? As it happens, it is pretty linear in this case.
Do you think the IPCC should now abandon showing a projected rate of change, because that would be foolish? Or possibly others might compare the actual measured rate with their projected one?

proclaiming that this manufactured slope is negative and therefore global warming has ceased is just contrived nonsense.
The slope shown is the result of a standard calculation. I assume you can see that it is negative.

The second graph of your own source seems to make it fairly clear that global temperature is fluctuating but the overall trend is one of warming. The fact that we have been in a down-turn of that fluctuation doesn't change that.

As I have been saying for some time, the rate of change, shown in that second graph, and for a much longer period here
Akasofu does indeed show that the world is warming. It also shows that the rate of warming is not too far off a straight line. The problem here is that the rate of warming is so far below that forecast (projected) by the climate models that these are on the point of being falsified to the confidence level stated by their authors.

Maybe you don’t think that matters.

As I have been saying for some time, the rate of change, shown in that second graph, and for a much longer period here
Akasofu does indeed show that the world is warming.

Well, then we're in agreement. Your assertion that global warming has ceased is false.

I see the Akasofu link doesn't work correctly.
Try downloading the .pdf file from here:
http://people.iarc.uaf.edu/~sakasofu/little_ice_age.php

Well, then we're in agreement. Your assertion that global warming has ceased is false.

I’m glad you think we are in agreement. I’m not so sure.
The world has stopped warming for the period shown in that graph, and there has been no obvious reason for that, like a volcanic eruption.

This is only important in that it is a strong indication that the global climate models are missing some important factors, and/or that the forcing factor used for CO2 is too high. This matters, because if our government acts on this false information, it will drop the GDP for no good reason.

I don’t know of any scientifically inclined skeptic who doubts that the world is warming (possibly some Russian scientists do) the question is, by how much? CO2 alone would only warm the world by ~0.4C/century (? from faulty memory) unless some forcing factor is applied.

Have you ever googled "galactic positioning" and "ice age"? Maybe if we hit the right spot in our lil ol part of the Milky Way, and global warming is real, it'll all balance out.

In the mean time, I've started to wonder about 9.2 billion of us by 2050, and how all that heavy breathing will affect the planet.

Thanks for your blog, from a critical-thinking comedian, a questioner as opposed to a skeptic. When solid empirical evidence disproves a belief I have, then I simply change my mind. Until then, I believe my own experience, and any solid proof I can find, which seems rare.

Parallel wrote: Like many dogmatic AGW believers, you only want to look at and believe reports that confirm your beliefs.

Oh the irony!

Parallel,

Please look up the fallacy "Cherry Picking" and then ask yourself why its convenient to start your sample window in 1998.

Jill,

The 9 billion number is not worrisome because of all the extra breathing. Its worrisome because of the extra resources required for support and because if we do not change how we get our energy, GHGs production will only rise, not fall.

As a demonstration, think of the amount of CO2 you produce by breathing as opposed to driving a car. Or the fact that for every killowatt-hour you consume from coal, you produce 2.3 pounds of CO2.

Breathing is hardly the issue.

That's one reason I don't buy the clean coal stuff. Even if some scientist finds a miraculous process to burn coal without leaving dangerous waste behind, it's still taking carbon that was safely sequestered underground and putting it in the atmosphere.

Breathing is just taking bio-available carbon and shuffling it around. It'll end up in some plant, which will end up in something that eats the plant, which'll get expelled in an exhale, and so on. Same thing with cow methane.

Please look up the fallacy "Cherry Picking" and then ask yourself why its convenient to start your sample window in 1998.

Obviously the graphs I linked did not start in 1998. The first started at the publication date of the forecast, the other in ~1981 Lucia tends to do her comparisons of forecasts with actual results from the date of publication of the IPCC report. She has explained the significance of the starting date many times.

Suggest you read Prof Akasofu’s paper (linked above,) covering a wide range of time periods, as you seem to have some difficulty understanding problem.

I have yet to see you, or anyone else, answer the factual points made earlier with counter facts, instead just using unsupported generalities or opinion.

Wait a second, your point is not that you cherry picked from 1998, but an even shorter timespan? LOL.

I have yet to see you, or anyone else, answer the factual points made earlier with counter facts, instead just using unsupported generalities or opinion.

What? Perhaps you can bullet point your "facts" here, because as far as I can tell you are trying to build a strawman (i.e. that CO2 and temp should rise and fall in a together), which I have already addressed.

can I quote your link here?

As you can see, the 5 year running mean has taken dip. Does this mean global warming has ended? Nope. Smoothed graphs aren’t particularly good splendid ways to figure out what’s happened to trends.

So can we now agree that global warming has not ended since you are even linking to sources that agree with skemono?

Wait a second, your point is not that you cherry picked from 1998, but an even shorter timespan? LOL.
What do you see when you look at the graph here http://www.drroyspencer.com/latest-global-temperatures/ ">global temps ? The blade of the hockey stick?
What? Perhaps you can bullet point your "facts" here, because as far as I can tell you are trying to build a strawman (i.e. that CO2 and temp should rise and fall in a together), which I have already addressed.
Obviously I do not think temperatures are strongly related to CO2 levels. But that is what the AGW folk claim is it not? I listed the unanswered questions in my first post on this thread. Care to answer any of them?
So can we now agree that global warming has not ended since you are even linking to sources that agree with skemono?
Please don’t misquote me. See Prof Akasofu’s paper for a good explanation.

Hi TechSkeptic,

Thanks for your input- that was my dry sarcasm being somewhat obscure. Heavy breathing was an indirect reference to breeding, and I'm in complete agreement about the concern of the over consumption of resources. Clean water alone is scarce and difficult to obtain for many people already. Over population is responsible for a number of problems we are facing today, which is why I brought it up. I've thought about it everyday for at least the last 20 years.

I don't want to continue off topic too much. I will say I'd rather not risk the continuation of emissions that, at the very least, have caused asthma rates to skyrocket in major cities. Can scientists really predict with 100% certainty how current and growing emissions will affect weather patterns in the next 100 years? In the next 20 years? I don't think so. I say take the safe road, or at least start looking for it. Human beings are so stuck on combustion, and it's time to move on. Unfortunately, I'm an artist, not an inventor, though I did have an interesting dream about an electrical amplification device.

I do take these matters seriously, but occasionally need to decompress with some humor.

Unfortunately, I'm an artist, not an inventor, though I did have an interesting dream about an electrical amplification device

As did Shockley, Bardeen and Brattain, the creators of perhaps the most important invention of the twentieth-century: The transistor.

LOL Jill,

Sorry. I've spent too much time over at that Joanna Nova site, apparently and humor flies over my head.

"Explanation of why global warming has ceased for almost ten years despite the increase in CO2 levels. See http://rankexploits.com/musings/2009/giss-feb-reported-trend-since-jan-2001-still-negative/ and also note the second graph from that link. The previous response was to doubt Lucia’s ability. If that is the case, care to point out the mathematical error?"

Lucia assumes that the standard confidence interval of the slope of a temperature series is what tells you the statistical range of the current temperature trend.

That's not exactly true. An analogy I've used is this. What if you tried to determine the long term temperature trend by calculating the confidence interval of the temperature slope of the last 10 days (global averages)? OK, if you want more data points to get a tighter confidence interval, don't use the last 10 days, use the last 1000 hours.

There are more appropriate techniques. Tamino, for example, would tell you to consider auto-correlation. But even this doesn't work for really short ranges.

What I've suggested is looking at 11-year ranges in the historical record, and determine where 95% of the slopes fall relative to the trend.

It turns out that 11-year slopes are (+/-) 2.7C/century from the long term trend 95% of the time. Last I checked, the 1998-2008 slope was only -1.53C/century relative to the expected trend, which is 2C/century.

I wrote about that here:

http://residualanalysis.blogspot.com/2008/08/why-1998-2008-temperature-trend-doesnt.html

2008 was an unusually cold year, and 1998 was an unusually warm year. We'll see what happens in 2009.

I’ve seen these statistics calculated in dozens of different ways. Lucia’s current post shows what happens when “..(you) pretend to have selected the start year at random, but then do it for every year between 1960 –.. “

No matter how you slice it, the temperatures have not gone up in the way projected by the GCMs, so something is wrong with them.

My larger concern is the poor quality of the temperature proxies used to show earlier temperatures. Steve McIntyre showed graphs of all the proxies used (hundreds of them) in www.climateaudit.org from which it was clear you could get any trend you wanted by selection of which proxies to use. Talk about cherry picking.

Couple that with the abysmal siting of the modern surface temperature stations, such that most must show some degree of “urban heating” and I view any claimed certainty in all these graphs with great skepticism.

It would be logical to repeat the proxy sampling, and site the temperature stations properly – something that would cost a tiny fraction the AGW research budget- but this is harder than sitting behind a desk in an office & shuffling paper….

OK, I was just going to ingore this entire thread, but this one in particular is too good to miss...

What do you see when you look at the graph here global temps ? The blade of the hockey stick?

I'll tell you what I see: I see the words "The smooth curve in the graph is a fourth-order polynomial fit to the data, which smooths out the large amount of monthly variability in the data and helps reveal the underlying ‘trends’" [my emphasis] and I see someone who either (a) doesn't understand what a polynomial fit is and why it is completely wrong for analysing climate data, or (b) someone being deliberately deceptive towards people who don't know what a polynomial fit is and etc... I see someone whose been tinkering in Excel to find a trendline that produces the result he wanted.

I'm pretty sure that Roy Spencer does actually know what a polynomial is and why it's completely the wrong choice for this data, so the only reasonable conclusion is that he's playing fools like you for suckers. And you fell for it - hook, line, sinker, and copy of the "Angling Times". How does it feel, sucker?

Here's a tip - any time you see someone using a polynomial fit on data where the y value is not a function of the x value, you're dealing with either a fool or a charlatan. Or both.

I'm certainly not going to be taking anything you ever say about statistics seriously after this howler...

Awesome!!

Mann, Bradley and Hughes published a quantitative study (MBH98) using a new climate field methodology, showing temperatures as a hockey stick shape, and eliminated the Medieval Warm Period, flattening the fluctuations in global temperatures over most of the past millennium.

Looking at the possibilities shown in the graphs, I don’t think Mann had to worry too much about finding a hockey stick somewhere.

I see the words "The smooth curve in the graph is a fourth-order polynomial fit to the data, which smooths out the large amount of monthly variability in the data and helps reveal the underlying ‘trends’" [my emphasis] and I see someone who either (a) doesn't understand what a polynomial fit is and why it is completely wrong for analysing climate data,

Well surprise, surprise, the method used shows the temperature trend you say. Is that not down or level at best? I always thought the hockey stick blade pointed up. Do you have it inverted on your computer?

It would be interesting to see you produce a graph, by any method except Mannumatics, that reproduces temperatures in the way shown by Mann. Oh I forgot: you can’t take anything I say seriously. Particularly if it shows you to be in error.

You have no idea why it's completely retarded to use a polynomial fit on this data, do you? Do you even know what a polynomial is?

Do you think that global average temperature is dependant on the number given to the calendar year in the Gregorian calendar? I would hope the answer to that is "No". Well, using a polynomial only makes any sense if it is.

Of course, if he'd chosen a third-order or fifth-order poly, the "trend" would appear to go in the opposite direction - which is why he's chosen fourth-order. A second-order poly wouldn't look anything like a trendline, even to you (as it would diverge wildly from the actual data almost immediately), and Excel can only go up to sixth-order.

You're right, I can't take anything you say seriously. Why should I? You clearly haven't got the faintest clue.

I'm going to do something weird and side with parallel for the moment (until skeptico gets off his ass and writes something new...LOL!)

Not on the polynomial fit stuff, but on his link to graphs. I think the point he was badly trying to make was that using proxy data from the last 2000 years to make a trend can lead to different conclusions about whether there is a rising trend or not.

The claim at the site that he linked to was that Mann didn't use the full set of data points (400 of 1200 or so) to create his fits and that the author could choose a different set of points, get better R values, and see a trend downward.

OK, now i'm done siding with him because the whole exercise is stupid. I have no idea what they think Mann was doing, but you don't use a polynomial to show a statistical trend. You use a moving average, with a decent window (10 or 20 years). As Dunc mentioned perfectly, you would talk about a fit to apolynomial if you thought that the Y axis was a function of the X axis.

So that whole post is pretty idiotic (and if mann was in fact doing as they claimed, that was equally idiotic).

a moving average of proxy and measured data shows the hockey stick. Get over it.

What if... the "Hockey Stick" Were Wrong?

Of course, parallel won't believe a word from RC.

Climate myths: The 'hockey stick' graph has been proven wrong

He probably won't believe NS either...

Personally, I don't have the statistical expertise to really get into the nitty-gritty of the precise methods used to derive these multi-proxy reconstructions. However, I think it's pretty clear by now that parallel doesn't have the statistical expertise to tell a valid argument from a hole in the ground.

However, I do know that the MBH results are pretty robust with respect to the selection of proxy datasets. You can knock out (IIRC) any three of the datasets used without changing the overall outcome. And I also don't believe for an instant that the various objections raised haven't been considered by the reviewers.

Now, while it's true that you can mess around with your selection of datasets and weightings to produce different results if you try, it is also true that there may be perfectly valid reasons for a particular selection of datasets or weightings. The fact that someone acting in bad faith can produce results to order from red noise does not in itself prove that Mann et al. did so. It's an accusation of outright fraud, and so far I'm not seeing any real evidence to back it up. It also implies a conspiracy amongst reviewers and other authors to cover-up and perpetuate that fraud. Again, no evidence.

I'm also not about to take the word of someone who says "I failry am new [sic] to this science so you can follow my own learning curve as I figured out the scope of the problem." over that of real peer-reviewed scientists.

Oh yes, we've got a full-blown conspiracy theorist:

"The Mann08 paper which showed flat historic temperatures was in my opinion intentionally created to trick people into thinking recent temperatures are unprecedented. The very fact that this passed peer review is to me proof of peer bias toward the global warming cause... or increased funding."

[My emphasis]

Also:

"Nearly all of the remedies proposed by the very same scientists who predict global warming are directed toward cost and taxation."

"The only consensus that exists is between a small number of primarily government funded scientists."

Finally, for a real laugh:

"One of the purposes behind starting this blog was to figure out if AGW was correct. I’m 7 ish months in, I’ve read dozens of papers, done the math on several others and guess what. I have no answers."

You've been at it a whole seven months, read dozens of papers, and you've done the maths on several papers other than the ones you've read? No wonder you have no answers.

From Ten Reasons to be a Global Warming Skeptic. The other's aren't much better.

Arggghh! Grocer's apostrophe! I hate myself.

Dunc,

when you are done here, maybe you wanna head over to that JoNova site to have a full lobotomy. I got sucked in and then realized that site if filled with creationists. AGW is hardly the place to discuss evidence, theory and modeling when they havent gotten past evolution, and perhaps not gravity.

TechSkeptic,
Lay off the insults, they just make you sound childish.
I invited you to post a graph in rebuttal, but so far you have declined to do so.

Dunc,
Should you really be interested in delving a bit deeper into the hockey stick, you might try Steve McIntyre’s most recent http://www.people.cornell.edu/pages/asu1/do-we-know-that-1990s-warmest.pdf ">presentation. The upward pointing blade is mainly the result of the dubious bristlecone series that Mann used.

Sorry, go read the thread. Its the truth. there is simply no reason to try to discuss science with people who haven't gotten past evolution yet.

If they aren't willing to understand 150 years of overwhelming evidence, you can bet that global warming is going to be simply arguments from authority also.

There is no graph to rebut. Its ridiculous to rebut a fit like that. Show me a physics based model that fits the last 100 ears of dat as well as presently used models, but without the AGHGs in it (keeping modeling coefficients within reasonable ranges).

Deniers have had 30 years to work out a model like that...why aren't there any? Instead they rely on polynomial fits as if year is an input to the function. And tiny data sets. Its kind of a joke.

The thing is, even if you completely disregard the entire field of paleoclimate reconstruction, we still have multiple independent lines of evidence for AGW, starting with the basic physics of radiative energy transfer - basic physics which no-one has ever seriously challenged.

Frankly, I don't find McIntyre's powerpoint that convincing. The precise warmth of the MWP isn't really that important in terms of the overall science.

Another classic crank symptom - a laser-like focus on one not-especially-important matter to the complete exclusion of all else.

Even if it should turn out that past climate is more naturally variable than we thought, that doesn't mean we can screw with the system with impunity - quite the contrary in fact. The rougher the seas, the more important it is not to rock the boat unnecessarily.

when you are done here, maybe you wanna head over to that JoNova site to have a full lobotomy.

Thanks, TS, but I'm really trying to reduce my exposure to corrosive idiocy. It's not good for my mental health, or faith in humanity.

You have to turn your brain on first. Nothing produces "Garbage in = Garbage out" quite like statistical analysis.

I’ve seen the analysis done countless ways and prefer to actually see the program with the results. For some reason the AGW authors seem to want to keep their workings secret. You have to wonder why if they have nothing to hide.

Nothing produces "Garbage in = Garbage out" quite like statistical analysis.

That's pretty rich coming from a guy who doesn't know what's wrong with using a polynomial fit on a chart of temperature by year. You also appear to be completely ignoring all of our other points, and not acknowledging when your prior arguments have been comprehensively demolished - preferring instead to just move on to yet another stupid evidence-free claim.

I see no point in continuing this debate any further. I've had more reasonable discussions with creationists.

My previous post was deleted. Let’s see if this does any better.

Not only has the lower atmosphere been cooling this century, but so have the oceans.

Where do you suppose all the extra heat trapped by the increased level of CO2 is hiding?

Just a guess: between your ears?

That's pretty rich coming from a guy who doesn't know what's wrong with using a polynomial fit on a chart of temperature by year. You also appear to be completely ignoring all of our other points, and not acknowledging when your prior arguments have been comprehensively demolished

Let’s see if I understand your inferences.
1. You say I don’t know about polynomial fits (and a dozen other ways) when I do.
2. You don’t like the kind of smoothing used by Dr. Roy Spencer, a senior scientist who received NASA’s Exceptional Scientific Achievement Medal for their global temperature monitoring.
3. Because Dr. Spencer smooths the global temperature in a different way than one would get with Mannumatics (or smoke & mirrors) this proves that a downwards slope is really pointing up.
4. You don’t have to provide any actual facts to support your point of view, just you saying something is proof enough.

A polynomial fit, from what I understand (which isn't much, I'm only able to figure it out based on my current enrollment in Algebra II and the context of the discussion), would basically work anywhere where Y=something involving an X. Why would there be a function of year=temperature^2+5 times temperature+7 or something like that?

Explain to me how that makes sense, please. Or why I don't understand it, either way works.

Wait, I forget, is year usually the y or the x of the graph? If I got it wrong, just pretend I wrote it the right way.

The year's going to be the x of the graph, and I'm reasonably certain you're correct about the polynomial fit (it certainly falls in line with my math knowledge and what Tech and Dunc have been saying in this thread). There'd be no reason to use such a fit on any graph of [Anything] vs. Time, unless [Anything] was specifically a function of time.

It seems it is Tech and Dunc who don’t understand. The gratuitous insults suggest they are students or raw graduates.

Look here should you be interested in finding out more about polynomials

You can try some examples for yourself
here

See also the neat demonstration for curves

As I wrote earlier “Nothing produces "Garbage in = Garbage out" quite like statistical analysis.”

Awww, the blog ate my comment!

I guess I'll redo it.

I was, apparently, right about the temperature=year^2+5year+25 (or whatever function it's actually using) thing, based on your link. Therefore, you have 2 questions to answer.

A) What happens in, for example, the year 48972? Do you understand why the function would be inaccurate?

B) Why does it work if years are X on the AD scale? If it was Y, then all you'd have to do is adjust the final X-less number to fit. But since years are X, it doesn't make sense that they'd work based on our arbitrary AD scale.

It seems it is Tech and Dunc who don’t understand. The gratuitous insults suggest they are students or raw graduates.

Look here should you be interested in finding out more about polynomials

Ha ha! EPIC FAIL!

If you're going to link to something to refute an opposing argument, you should make sure you understand it first. Your link proves my point, you moron.

Dunc,

I'm not sure if i missed your explanation somewhere above, but I don't really understand why it is wrong to fit the temperature data with a higher-order polynomial, as long as you're not trying to extrapolate beyond the last data point (i.e. trying to predict the future, which he specifically denies trying to do).

Perhaps there's a clue in what Tom Foss said:

There'd be no reason to use such a fit on any graph of [Anything] vs. Time, unless [Anything] was specifically a function of time."

Could someone explain this a bit more in detail?

It's perfectly simple. A polynomial only makes any kind of sense if the y-axis value is a function of the x-axis value (specifically, a polynomial function). If that is not the case, then using a polynomial isn't actually a trendline - it's just a completely spurious function which may look like a trendline over a specific range (if you're careful enough) - but it's not actually showing any trend in the data. Frequently, if you "fit" a polynomial to data which isn't actually suited to it, your "trendline" will diverge wildly from your actual data, even within the range between the outermost inflection points - because it's not actually a trendline.

Then there is the question of what order of polynomial to use. Normally, if you know that the data you're working with is polynomial, there will be a good reason to pick a particular order. However, if the data isn't actually polynomial, there is no possible justification for choosing a particular order. In this case, Spencer has chosen a fourth-order poly because he wants the "trendline" to go down at the right-hand end. There is no possible justification for using a forth-order poly rather than any other order, as the data isn't polynomial in the first place. If he had chosen an odd order poly instead of even, the "trend" would be rising at the end.

Then there is the fact that polynomials always exhibit symmetry, the precise type of symmetry depending on whether it's even or odd ordered. Surely the fact that a polynomial must always be symmetric would give away the fact that it's a completely stupid choice for climate data? Even if you're not trying to predict anything (and what is the point of a trendline if it has no predictive power - that's the whole frakkin' point of a trendline) it's completely wrong. The best options for smoothing this data would be either moving average or local regression.

This is basic high-school mathematics.

Hang on a minute - I may well be wrong about the symmetry bit... It's been a long time. However, all my other points still stand.

Hi Dunc,

I'm not sure choosing the right curve fitting technique for climate data is part of basic high-school mathematics, but the part of your comment that does perhaps belong in high-school;

If he had chosen an odd order poly instead of even, the "trend" would be rising at the end.
is wrong.

The first paragraph of your comment was useful, so thanks for that.

Okay, if anything, this thread has really made me follow through on what I've been planning for a year or so now--reopening my Calculus books and refreshing my memory of sorta-advanced math.

Anyway, here's what I think Tech and Dunc have been getting at, in one way or another. The graph of y=x^3 is rising--in other words, y-values to the left of the origin are negative, and y-values to the right of the origin are positive. This is true too of y=x, y=x^5, y=x^13429, and so forth. All odd functions of x are rising from negative to the left of the origin to positive on the right of the origin.

With the exception, of course, of odd functions where the leading term is negative--for instance, y=-x^3. That's the opposite function, so y-values on the left of the origin are positive, and y-values on the right of the origin are negative.

There's more to this, of course. I say "origin," but that's not entirely accurate; the graph of y=(x-1)^3 would have its turning point at (1,0) instead of (0,0). Depending on what you add or subtract, and where you do it, you can change the polynomial to have any number of shapes.

Anyway, to finish off the odd/even stuff, odd functions are generally rising or falling, depending on whether or not the leading term has a negative coefficient. Even functions, on the other hand, are concave. The graph of y=x^2 is a parabola that is concave upward, while the graph of y=-x^2 is concave downward. The sign of the coefficient just determines which way the parabola (or similar figure for higher powers) is pointing.

The shape of the polynomial depends mostly on what the leading term is (this determines the general shape), what other terms are involved (these determine where the turning points and intercepts are--and a function x^n will never have more than n-1 turning points), and how far you're zoomed in or out. Once you zoom out far enough, the graph of every third-degree graph looks pretty much the same as every other third-degree graph, every fourth-degree graph looks pretty much the same as every other fourth-degree graph, and so on. This is, I think, key to the discussion.

If you're looking at the graph of y=(x-1)^3-x+2, and your range is from x=0 to x=1, you're going to see a very pronounced downward curvature to the line. If, however, you zoom out to a range of x=-5 to x=5, you'll see that the downward portion from 0 to 1 is a small bit of the overall rising trend of the graph. Zoom out far enough, and you'll see it going from x=-∞ to x=∞.

This is the point, I think, where the rest of Dunc's post is going. Zoom out far enough from any odd-degree polynomial fit, and you'll find -∞ at one end and ∞ at the other. That alone makes it pretty well ridiculous as a way of predicting the future or retrodicting the past just from a set of data points where the dependent variable isn't a function of the independent variable.

It is (as is often the case with denialists and woos) a problem of mistaking description for prescription. Yes, you can plot a set of data points, and yes, you can finagle a polynomial to run in or near most of those data points, but all that does is provide a line that, for a certain range, describes the arrangement of data points you've entered.

This is not the same thing as starting with some function y=f(x), where you can plug in any x-value and get the one and only corresponding y-value. In this latter case, the function f(x) is prescriptive: y is defined as what you get when you apply the function f(x) to x. The former case is descriptive. In the data set, the y-values may bear no relation to the x-values; changing the x-value may not necessarily change the y-value, and plugging an x-value into the function f(x) will not necessarily result in a y-value that corresponds to the real-world situation that you measured to produce the data points you provided.

In other words, I can take the data point for my height at age 10 (4'7", as I recall) and the data point for my height at age 25 (6', give or take), and I can come up with a fit line that goes through both points. From that fit line, I can predict that when I'm 50, I'll be 8' 4 1/3".

In other words, it's silly to treat descriptive fit-lines as though they're prescriptive functions. You might be able to get a ballpark estimate of what the next data point might be, based on an overall trend or a trend in the last few points, but the fit line won't necessarily give you any more accuracy than just the data set.

To Dunc: you're pretty much right about the symmetry bit. Odd functions are symmetrical about the origin and even functions are symmetrical about the y-axis; assuming there haven't been any changes to shift the graph one way or another, in which case it'd still be symmetrical, but the landmark has changed.

I hope that's been informative and accurate.

Thanks Tom, you clearly have far more patience than I do.

Hi Tom/Dunc,

Your intro to polynomials is well and good, but the point here, as you say, is that the temp data is not polynomial in nature.

Like you, I'm refreshing this as we go, so I'm not 100% confident about this, but these images should be illuminating:

Picture 1 shows a 6th order polynomial and 4 different order polynomial fits, to illustrate what you were saying, and also to show where I think Dunc's confusion came from (i.e odd point up, even pointing down).

Picture 2 is a random signal with 2 different (odd/even) polynomial fits both showing an downward trend.

Picture 3 is the same random signal, except I have altered two of the 101 entries (#96 and #100), to higher values, to prove my point. The same two poly fits are applied, this time showing a upward trend.

I think this illustrates that the order of the polynomial does not dictate the "trend" in the context we were speaking.

It also shows that a polynomial fit to random data is extremely sensitive to end-point values (e.g two fluke readings flipping your "trend" upside-down). A good reason not to use a polynomial`fit in the case of the UAH Global temperature data perhaps?

I seem unable to relate the talk of "zooming out" to infinity or the symmetry of functions to this discussion, but that might be a limitation on my part.

Ok (for the second time you comment-eating-piece-of-shit-software), that was not entirely accurate.

You could dictate the "trend" with the order of the poly fit too, of course, even for the random data. Which brings me to the point where I wonder what my point really was, and also to the point where I'm out time. Guess I'll have to re-read this later.

Sorry for this mess. I just realized what I was trying to say, here in non-verbose mode:

An odd (even) order polynomial fit always produces and upward (downward) "trend" if and only if the polynomial fit is applied to a polynomial.

Can someone run this claim through the truth-machine? Thanks.

Oh dear..

Here the correct picture 2.

:(

Martin: Zooming out means increasing how much of the function you look at. To try it, go to the Window function on your basic graphing calculator, and raise the x-max and y-max values and lower your x-min and y-min values. (You might also want to increase the x- and y-increments so that it doesn't look all cluttered.) That will show you what we mean by "zooming out".

The comments to this entry are closed.

Search site