Statistics

BEST And The Wall Street Journal’s Numbers Guy

Carl Bialik, the Wall Street Journal‘s Numbers Guy, has two columns about BEST’s temperature reconstruction model: “Global Temperatures: All Over the Map” and “A New Trove of Global Temperature Numbers”). Read them. BEST temperature

Of those statistically modeled temperature reconstructions, Bialik wisely sought out a quote from a Peer1:

The Berkeley scientists “have a very complicated model,” says William Briggs, a member of the probability and statistics committee of the American Meteorology Society. “They reported on the setting of one of those dials” in the model. “That is not the actual temperature.”

Briggs is right. Sort of. The BEST model, like all statistical models, has a large number of dials which need to be set before it will produce results. How to set them all to the proper place is an interesting question, but one which is not of primary interest. Why one would want to report on the settings and not on the actual thing modeled, i.e. the yearly global average temperature, is what concerns us.

The conflation of settings of dials and the actual thing modeled is, as regular readers will know, an enormous and persistent error many users of statistical methods make. The reason it is an error is that our certainty of the dial setting does not translate into equality certainty of the thing modeled.

Let me explain. We can write down some very complicated statistical model of yearly global average temperature (GAT) as a function of all sorts of things, like location, yearly change, time, effects other locations have on this location, and on and on. Each of these things will have associated with it a dial, or parameter, which must be set to just the right position. The BEST model, for example, has a very large number of these dials.

Various formulas will let you set each of the dials, with different formulas producing different sets of configurations. Which is the right formula? Shhh. Don’t ask. Just pick a formula and set the dials. The formula will tell you the value on all the dials, but you can report on just the dial associated with, say, yearly change. This is where the problem begins.

Nobody believes, and it isn’t true, that the setting of the dials is perfect, without error, exact. There is some uncertainty. Assuming the model of GAT and the formula used to set the dials are perfection themselves allows us to produce a guess of the uncertainty of the dial settings. If you are feeling especially vigorous, you can report that uncertainty along with the dial settings. This signals to your audience your honest and tells them that you don’t quite know what the exact dial settings are.

But in the end all you are left with is information about a bunch of dials and their values. Who cares about those? Well, statisticians, maybe, but not the civilians who remember that the original goal was to produce a guess of the yearly GAT through time. How do we get that?

What we can do is to set each dial at its best guess and then use the model to give us a guess of the yearly GAT. This works. But since nobody believes that the settings of the dials is perfect, the guess of the yearly GAT will be too certain.

To fix that, we run through all possible settings of the dials, and for each setting we produce a new guess of the yearly GAT. Not all settings of the dials are equally likely, so we give more weight to the settings that are most likely, and give small weight to the settings that are not so likely. We do this for all dials simultaneously, not just one or two.

The end result will be a guess of yearly GAT that accounts for the uncertainty we have in the dial settings. We do not report on dial settings at all. Necessarily, this guess of GAT will be more uncertain (relatively speaking) than the certainty we have in any particular dial setting. This is what BEST should have done but did not.

Careful readers will have remembered that we still assumed the model and formulas are without error. How can we know that? There’s only one sure method. Bialik again quoted from an expert:

“The only way to prove these models are any good is to demonstrate that they skillfully predict new data,” said William Briggs…”None have done this. There is reason to suspect all models.”

Briggs is right again.

—————————————————————————————–

1As in one who reviews.

Categories: Statistics

31 replies »

  1. Bialik’s headlines refer to “Global Temperatures.” The BEST numbers are so far land only.

    And aren’t we really looking for changes in energy using temperature as a proxy? More dials?

  2. To repeat an important critical point: The BEST numbers do not join on to the satellite data over land during the satellite era 1979+. The satellite trends are signficantly lower than BEST (and the NOAA/NCDC and NASA/GISS trends). The difference is huge — nearly 1 1/2 degrees/century. This is prima facie evidence that whatever factors are contributing to the difference between the American surface data and the satellites have not been identified or eliminated by BEST.

    I should add that the British CRU data are in much better agreement with the satellites, though still trending higher.

  3. @Noblesse Oblige

    OR whatever factors are contributing to the difference between the American surface data and the satellites have not been identified or eliminated by the staff analyzing the satellite data. Or both, actually. And by the way, BEST data is not just American surface data.

    Sorry, but the satellite series are not by definition precise and certain. You might recall the error, uncorrected by Spencer and Christy for nearly a decade, that understated the warming trend by as much as 40%. Before you start drawing false conclusions about the divergence between the surface data and the satellite data, read this:

    http://www.skepticalscience.com/Primer-Tropospheric-temperature-measurement-Satellite.html
    http://www.skepticalscience.com/news.php?n=1081

  4. Orbital variatiions were accounted for and corrected decades ago. It is really old information. We have two independent groups, one academic (UAH) and one commercial (RSS), who take the same basic satellite data and apply different algorithms to compute temperature as a function of position and height in the atmosphere. The two outfits get slightly different results but the difference between them is far smaller than the difference between the NOAA/NASA trends and both UAH and RSS. Again, the CRU surface data lie between the satellites and NOAA/NASA/BEST.

    Another interesting factoid is that the difference between the satellites and surface OCEAN temperature trends is negligible, so that the overall GLOBAL trend differences between surface and satellite are actually small. But the fact that the satellites and surface measurements agree over water but differ greatly over land is more circumstantial evidence that the surface LAND measurements are contaminated by human activity that is not removed by BEST.

  5. @Noblesse Oblige

    You haven’t read the links. The Spencer/Christy error (which fueled the denialist myth “but the satellites don’t show warming” for more than a decade) was related to the diurnal cycle, not orbital variations. You’re probably talking about Wielicki et. al 2002 which didn’t correct their results for orbital decay, giving poor Lindzen hope of proving low climate sensitivity, for this precious little moment. But that’s another story.

    Be a sweetie and read the referenced articles. We could talk about facts then, not products of your imagination. There are at least three satellite temperature series, and three methods of producing them other than the ones used in RSS/UAH. Satellite data are by definition contaminated by a cooling signal from the stratosphere, and recently used algorithms a trend of 0.13 C per decade in the middle troposphere (http://www.agu.org/journals/ABS/2006/2005JD006798.shtml). Most probably for the TLT series it will around 0.18 C / decade — quite higher than in RSS/UAH.

  6. @ GS
    The shift of measurement time due to diurnal drift has also been dealt with. I don’t know why you continue to bring up issues that have been solved years ago and use them to try to throw doubt on the validity of the satellite measurements.

    Check that. I do know why. The parade of busswords is a giveaway.

  7. Leaving the past behind for a moment, a rational person would ask why, if global temperature (whatever that is) is such an important metric, haven’t scientists invented, agreed on and implemented a system to measure global temperature (whatever that is) from now forward?

    Man is carried naked and unconscious into an emergency room. Doctor orders some standard tests including body temp. Three doctors, two nurses and a pharmacist then have a meeting to decide how to measure body temp. Under the tongue? Rectal? Armpit? Ear? Glass/alcohol/mercury thermometer? Electronic?

    Nearby hospitals are called and asked how they measure body temp.

    The American Association of Body Temperature published a special edition calling for more research and recommending that mercury/glass thermometers be used because that’s the way it’s always been done.

    Med Student walks over to the patient, puts his hand on the patient’s forehead and says, “Feels cool to me. I think he’s dead.”

    Christy paper: What Do Observational Datasets Say about Modeled Tropospheric Temperature Trends since 1979?
    http://www.mdpi.com/2072-4292/2/9/2148/pdf

  8. @Noblesse Oblige

    Why don’t you just read up before you speak up?

    a) Fu et al. 2004, 2005:
    http://www.ncdc.noaa.gov/temp-and-precip/msu/nature02524-UW-MSU.pdf

    b) Vinnikov & Grody et al. 2005:
    http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.143.4699&rep=rep1&type=pdf

    c) Zou et al. 2010:
    http://www.agu.org/journals/ABS/2006/2005JD006798.shtml

    I’m sorry I’m “throwing doubt” at your religious beliefs. It may be a shock to you, but in science there are no holy grounds where no-one dares to tread. And satellite temperature series are anything but holy, as you should be aware, since you apparantly have heard about errors introduced by incorrect handling of the diurnal cycle and orbital decay. You’re contradicting yourself here a bit.

  9. @Speed

    The fixation on the measurement of a “global temperature” is a sure sign of confusion. This is a technical quantity. What matters is the trend, that’s why temperature series are built on the basis of average deltas from the base period, not a delta of a “global average temperature”. Your analogy is simply false. It’s more like: “We’re seeing that the patient is getting hotter, we suspect an infection”. “Oh, no, you can’t say that just because you have a series of oral measurements indicating a temperature rise, after all the mouth is just a tiny fraction of the human body, and we are seeing a divergence of 2 degrees from your measurements in comparison with our parallel rectal readings, and you haven’t properly quantified the uncertainty, so the whole series is a shambles — until we have a completely redesigned system of temperature measuring and a set of parallel thermometers placed all over the patient, and until we have used them for a month or two (after all, what are just three days in comparison with his lifetime of forty years?) we refuse to take radical steps that could incur considerable cost on the side of the hospital. For all we see, he might be equally well getting colder. Or it’s a cycle”.

    And thanks for referring to the Christy et al. article — it’s nice to see how suddenly data sets restricted to the tropics are representative of global trends, when at the same time BEST is being criticised for publishing results on land surface only.

  10. This may be completely OT but over the years I’ve come to notice that parrots don’t comprehend the words they use and, it you listen very carefully, they aren’t really saying anything that makes sense. Conversation is impossible.

  11. @DAV

    And the worst part is that those stupid parrots just keep repeating references to research results, paying no attention whatsoever to what’s traditionally been politically correct and acceptable on a particular forum. That’s sooo antisocial.

  12. As a general rule, I don’t normally read the comments section of a blog. I have made an exception for this blog since the comments are generally respectful and thoughtful, and I often learn things, or follow a commenter’s link to an interesting article. For my purposes, at least, Grzegorz Staniak seems to be intent on sabotaging the comments section whenever the subject turns to AGW. I find many of Grzegorz’s comments to be disdainful and disrespectful, and the links generally point at long, dense articles. A very ineffective (and, well, just plain lazy, if you ask me) way to teach something, which leads me to believe that that teaching is not on Grzegorz agenda.

    So I haven’t been reading Grzegorz comments.

    I don’t suppose that those folks responding to Grzegorz could clearly mark their comments as such (@Grzegorz, for example) so I could skip over them, too?

  13. Milton Hathaway, DAV, all,

    Sorry, all. But I’m still overseas and a bit tied up. I too haven’t been reading that person’s comments. I haven’t had any time to do anything about it, but hope to soon. Meanwhile, the strategy of indifference cannot go wrong.

  14. Mr. Briggs,

    You’re taking a certain risk by classifying some of your posts under the “Climatology” tag. Namely, you might be viewed as someone whose inention is to present their views on this subject — and by doing so, open them to criticism. It’s clear that you’re not even trying to be impartial in your treatment of climate science here, on occassions you simply spread gross misinformation, like in the case of the CLOUD project, and you haven’t answered a single question of those that I’ve asked you — some of them effectively proving you wrong, too. I guess it must be part of this “strategy of indifference” that prevails on this forum: let’s ignore anything that doesn’t agree with our preconceived opinions. That’s actually your problem (I mean most of you guys). Noblesse Oblige doesn’t know what he’s talking about as far satellite temps are concerned, but he simplt doesn’t want to know — it’s below him to read a couple of popular artciles so that he realizes that he’s point about RSS/UAH is not an “important critical point”, but a delusion. Milton considers it “sabotage” for me to answer Noblesse and show (or try to show) him what satellite series are really about. Well, guys, all you need is a bit of patience. I’m sure as soon as Mr. Briggs comes home he’ll do a bit of censorship and restore peace, calm, and the blessed uniformity of opinions in your little mutual admiration society.

  15. Milton Hathaway
    I have read a few of those long, dense articles Staniak has linked to. Usually they don’t prove what he says they do and quite often prove him wrong. Either he hasn’t read them or didn’t understand them.

  16. Grzegorz Staniak said, “The fixation on the measurement of a “global temperature” is a sure sign of confusion. This is a technical quantity.”

    Temperature is not a quantity, technical or otherwise. The balance of that paragraph is gibberish — many words, no point.

    Grzegorz Staniak said, “BEST is being criticised for publishing results on land surface only.”

    If you are referring to my comment above, “Bialik’s headlines refer to “Global Temperatures.” The BEST numbers are so far land only.” my criticism is for Bialik.

    If you (Grzegorz Staniak) wish to debate the validity of satellite measurements, I suggest that you chime in at Climateaudit at the link I posted above.

  17. G.S. As you’ve no doubt noticed by now your remarks, even the abusive remarks (e.g. “Be a sweetie and read the referenced articles. We could talk about facts then, not products of your imagination.”) . A civil response from you seems beyond your capability?

    This abusive pattern of yours, coupled with “flawed” analyses naturally prompt some thought as to what, if any, your qualifications might be. As you’ve provided in your inputs some contact info (clicking on your image you’ve provided with your comments) we find that you’re a computer Systems Administrator & you offer no evidence of a degree or other qualification indicating any background in or qualification for real analytical capacity. That doesn’t mean you’re lacking in this regard, maybe you just want to conceal it. Your vitae certainly do not support the know-it-all attitude you convey … though the lowly lack of apparent qualifications does support the belligerant abusive attitude one would expect from an outsider would-be wannabe. If you have any real qualifications perhaps you could share them? Or, perhaps just stick to the facts & conclusions–understanding that reasonable people can disagree/reach different conclusions based on the same facts.

    Or, you can go on making a fool of yourself. Your choice.

    Grzegorz Staniak FROM: http://pl.linkedin.com/pub/grzegorz-staniak/1/6b8/808
    Systems Administrator at LubMan UMCS sp. z o. o. (no degree/education listed)
    Lublin Area, Poland
    Computer Networking

    From the link posted here: http://en.gravatar.com/gstaniak at: gstaniak@gmail.com

  18. @Jim Breeding

    I’m sure you’ve read the articles and found I’m wrong about them, just as I’m sure that you won’t give an example. It’s simply easier to just declare things, like you guys do here all the time, than to prove them.

    @Speed

    If on the other hand you (Speed) wish to debate the validity of satellite measurements, I suggest you at least read these:
    http://www.skepticalscience.com/Primer-Tropospheric-temperature-measurement-Satellite.html
    http://www.skepticalscience.com/news.php?n=1081
    so that you know what you’re talking about.

    @Human Person Junior, Jr.

    And of course you can elaborate on that and show how SkS shuns skepticism. Unless of course you meant the skepticism that’s so popular on this blog: “I don’t, but I’m pretty sure, so I don’t need to read or discuss”.

    @Ken

    LOL, Ken, you’re the man. 🙂 “Maybe you just want to conceal it”. Sure, Ken, sure. That’s why I use my real name on the net instead of pseudonyms like yours, use my real photo on the Gravatar and use Gravatar on LinkedIn, and use the same photo on Facebook (http://www.facebook.com/profile.php?id=100002799025400) — to hide and conceal. As opposed to all you Kens, Speeds, Noblesses and Human Persons, who are sooo open about yourself that it makes me ashamed of my secrecy. :))

    Way to go, Ken. A splendid example of logical thinking.

  19. Once upon a time the was a TCO who at least would actually converse and state (though rarely attempt to prove) his position but sometimes his reasoning would show. George, unfortunately, lacks all of TCO’s redeeming qualities. TCO’s bad side was button pushing. George tries hard but can’t seem to master even that. He consistently avoids articulating any of reasons he believes the things he links. This likely indicates his lack of understanding them. IOW: he’s a parrot.

    Well, George, how about it? You came here under your own steam. You weren’t dragged here. What for? How about actually telling us why you think all of these links are so pertinent? Are you really content with just saying as little as possible. This is your big chance to show you should not be completely disregarded. Can you show us you really do think for yourself?

    Bet you can’t do it, George. C’mon! Are you a man or a parrot? My money says you will post something tangential and without substance (like a childish “You first” or its equivalent). Prove you aren’t just a silly childish parrot.

  20. Grzegorz Staniak said, “If on the other hand you (Speed) wish to debate the validity of satellite measurements … ”

    I do not.

  21. Speed,

    You should. Satellite measurements are far from perfect and are, you might not be surprised to learn, the result of a statistical model. Actually, many models. The “inverse problem” is not light nor simple (how do you tell whether the photon that just hit the sensor represents temperatures from 10 M above the surface or 10,000 M?, etc., etc.). Mistakes and approximations are made. Satellite measurements come attached with uncertainty bands—these bands are usually stripped off the moment somebody wants to use the measurements. Result? Over-certainty.

  22. Staniak,

    Either begin acting like a gentleman or I delete your comments. I will not debate this point.

  23. This remark has been placed under HTML comments for being ungentlemanly. If you wish to see it, please view the page source: Control-u in Firefox; right click “view source” in IE.

  24. Mr. Briggs,

    Please elaborate on what you mean by “acting like a gentleman”. I’ll grant you I’m using an unpleasant tone and I’m generally being a bit of an obnoxious,

    This remainder of this remark has been placed under HTML comments for being ungentlemanly and for its use of foul language. If you wish to see it, please view the page source: Control-u in Firefox; right click “view source” in IE.

  25. Show me that yourself you’re not just a parrot squawking politically correct nonsene”

    WOW! Did I call it or what? It’s the equivalent to the childish “You first” just as I predicted.

    (*sigh*) It’s nice to be proven right but at the same time this is sad. You have a chance to state your purpose in posting here and can’t bring yourself to do it. The same old same old. So Very Sad and Foolish.

  26. Mr. Briggs,

    OK, you win. I’m not gonna play whack-a-mole with your censorship. You gentlemen have a good time, and bask in the sunshine of mutual admiration, undisturbed by the attacks of reality.

  27. Briggs, thanks for the suggestion.

    Many years ago I was involved in cardiac electrophysiology studies. The state of the art required inserting an electrode tipped catheter into the heart and recording electrical potentials to locate non-normal pathways. There were groups trying to map surface potentials (from 100 or so skin electrodes) back to specific locations in the heart — the inverse problem. Every year it was, “We are making good progress and should have a solution in the next year.” We’re still waiting. It’s a hard problem.

    I fail to understand why a satellite is trying to measure air temperature when the real variable of interest is the rate at which energy is being radiated away from the earth. “Global warming” is the difference between the goes-intas and the goes-outtas.

  28. If one reads Hoffer on the nature of the true believer then you realize thier passion merely reflects the cracks in their own mind.

    A point worth consideration when dealing with any internet fool.

Leave a Reply

Your email address will not be published. Required fields are marked *