Fruit Loops

We already exposed the graph of “Arctic Region Temperatures” used by Joe D’Aleo as not being so. Now Verity Jones has posted a “rebuttal” on WUWT.

D’Aleo posted a graph labelled “Arctic Region Temperatures” which most certaintly was not. I suggested that the stations used had been cherry-picked to give the false impression that the 1930s-1940s were as hot as present times in the “Arctic Region” (you know, the label on top of the graph).

It seems that neither D’Aleo nor Verity Jones can support it as being “Arctic Region Temperatures” or deny the stations were cherry-picked. In his “rebuttal” Jones says this:

Well excuse me but of course they were cherry-picked, but not for the reasons Tamino suggests.

Whoa! Is this the defense against an accusation of cherry-picking — to admit that the stations were cherry-picked? But, says Jones, it was not to represent Arctic Region Temperatures, oh no — it was to show a cyclic variation. He further says:

So no, the stations in the graph weren’t meant to represent the whole of the Arctic.

Let me sum up: Joe D’Aleo ends his post about Arctic sea ice with a graph labeled “Arctic Region Temperatures” which, it turns out, is based on a very limited set of stations, which his defender admits were cherry-picked and do not represent the “Arctic.”

Imagine that.

Verity Jones seems to have two strategies for saving face. One is to name as many fruits as he can think of, perhaps to divert attention away from cherries. I’m surprised he didn’t accuse me of being a mango.

The other is to suggest, using yet more “spaghetti graphs” of Arctic temperature records, that the 1930s-1940s may really have been as hot as the present time, at least somewhere in the Arctic. It’s sad that even though the original topic was the Arctic, the whole Arctic, and nothing but the Arctic, the one thing we can’t seem to get from Jones is: the Arctic. Instead we’re subjected to more graphs of hand-picked (shall we say “cherry-picked”?) stations which represent limited Arctic regions in an attempt to suggest that modern temperatures aren’t so hot compared to back then — at least somewhere. Perhaps the saddest part is that he does such a poor job of it.

For instance, here’s his spaghetti graph of 15 stations in northern Canada:

Spaghetti graphs have their uses, but if you want to know about regional temperature (that was the subject, right?) you should compute an actual regional average. Of course these spaghetti graphs are ideal if you want to create more confusion and doubt and give the wrong impression — don’t just leave in all the noise, make it look confusing to boot!

Let’s take the same 15 stations used by Jones and — hang on to your hats folks — actually align the station records and compute a regional average. This won’t be the “Arctic Region” since it’s limited to northern Canada, but at least it won’t be a pile of spaghetti either.

It’s a lot hotter nowadays in northern Canada than it was in the 1930s-1940s.

The other “regional” report Jones dumps on us is one from Ed Caryl, who takes station data from the far north and separates it into “urban” (with a possible human influence) and “isolated” (without). Caryl reports little or no trend in the “isolated” stations. Caryl gives us an example of self-delusion: essentially he takes the stations with the most trend and calls them “urban” while taking those he believes have little trend and calling them “isolated.” He actually classifies Eureka (in the NorthWest territories) as “urban”, reporting the massive population thus: “In winter it has always been 4 or 5 men. In summer, the population increases to as high as 20.” OMG!

But even if we do accept Ed Caryl’s cherry-picked set of Arctic stations, what do we get? Of course Verity Jones shows us his spaghetti graph, which looks like this:

Now if you simply average these data records you won’t get the right result, because Ed Caryl’s “normalization” process shows that he really doesn’t know how to align different station records. Let’s do it for him, and compute an average for these “Far North” stations:

Wow! The last decade is blazing hot compared to the 1930s-1940s. And remember, this isn’t all the stations, it’s the ones which were hand-picked — cherry-picked — by Ed Caryl to minimize Arctic warming. The ones that were heating up really fast, he chose to omit by calling them “urban.”

Apparently this is all we can get from deniers like Joe D’Aleo, Verity Jones, Ed Caryl, and their host Anthony Watts. Cherry picking and deliberate noise and confusion. When it’s exposed for what it is, their house of cards falls apart. Amusingly, we sometimes hear cherry-picking defended with a hearty cry of “Well excuse me but of course they were cherry-picked”, and misrepresentation of Arctic region temperature with “So no, the stations in the graph weren’t meant to represent the whole of the Arctic.”

But now I must correct an error of my own. We can get much more from D’Aleo and friends than just cherry-picking. In fact, shame on me for choosing a title for my follow-up post about D’Aleo’s Arctic misiniformation that implies further cherry-picking from Joe. It really doesn’t expose more cherry-picking at all — this time, it shows Joe D’Aleo just making stuff up.

Perhaps Verity Jones can explain that one to us.

45 responses to “Fruit Loops

  1. I suspect that we’re not far from the day when AGW septics just use glossolalia for argument.

    The point of the exercise from their point of view is simply to delay action. Why? I have no idea. They live on the same planet — the one they’re gleefully helping to render less habitable — as the rest of us. They aren’t going to go to some special right wing safe haven. So, why do it? I presume it’s simply to belong to the club.

    There was a study in ethics a couple of years ago that looked at the genetic basis for values. It turned up 2. Those for whom the chief value is universality. Like the Golden Rule. It applies to everyone. The other foundational value is inclusion. A gens, a race, a club, a football team, a political party, etc. With initiation rituals, like baptism or circumcision, and exclusive perks for membership. The Cheyenne called themselves The Human Beings. (The rest of us are something else.) Jews called themselves “The Chosen People”. Christians say baptism is necessary to be admitted to Heaven. So, right wingers have come to demand Loony Tunes. Who knows why? But you see it all the time. The goofier the assertion the better.And nobody even blinks. You wanna be in the club? you gotta dance the Chicken Dance on Main Street.

  2. Imagine that–ineptitude at WUWT!

    Another recent instance is the post on ‘a third peer-reviewed paper showing negative cloud feedback.’ (I was forced to read it when it was loudly trumpeted as yet another nail in the coffin on a newsite I comment upon.) Trouble is, the paper–which seems perfectly reasonable in and of itself to my eye–shows negative cloud *forcing* instead; the issue of cloud *feedback* is barely mentioned. (Net negative cloud forcing is no shocker; as the paper notes, it’s fully ‘consistent with past results’–a bit Watts ignored, of course.)

    I suppose that if one makes stuff up out of ignorance, one isn’t technically lying.

    Well, anyway, thanks again Tamino for whacking another mole; this one seems particularly hapless.

    [Response: In a very telling episode, even after both Roy Spencer and the paper’s author corrected him, and Watts had no choice but to change the title of his post, he still insists his interpretation is correct. Watts takes “delusional” to a new level.]

    • Y’all made me look, big mistake. The author actually corrects Watts *twice*. Watts and 99% of those posting there seem to think they understand the paper while the author doesn’t.

      Not the first time we’ve seen this level of insanity but geez, it’s hard to read through a thread filled with such a combination of ignorance and hubris …

      • Watts has “corrected” his post. After one strikeout, he offers this explanation: “While the -21wm2 and ~1.2 W/m2 values are correct, the comparison is wrong, and it is my mistake. The values are Top of Atmosphere and Surface, which aren’t the same.”

        Except that this isn’t a TOA/Surface confusion issue, it is the difference between a change of forcing (eg, +1.6 W/m2 from CO2) and the existing forcing from a substance (eg, -21 W/m2 for clouds, or whatever the total, large, W/m2 forcing from CO2 is). As well as the confusion between forcing + feedback.

      • Still, for Mr. Watts to write the words “my mistake” is something of a black swan, from what–or Watt–I’ve seen.

      • Wasn’t there a post a few years back where Watt admitted he was wrong about something? I remember a big debate about it between some posters here and tamino. He does seem to have occasional moments of lucidity.

      • I trust you on that, BPL. Still a pretty low bar he’s set, though.

      • was that the anomalies thing? where he wrote a series of posts based on a staggering misunderstanding of how temperature anomalies work, admitted he was wrong and would write another entry clarifying matter (which never materialised, of course), and now bans anyone who mentions it.

        either he’s too dishonest to correct his mistake, or still too much of an imbecile to understand as simple a concept as temperature anomalies.

  3. I read the WUWT post before I came here, and had trouble getting my head around it. It linked to a web page about deliberately looking for long term temperature records with a negative trend. It mentioned that strange urban/rural classification of which Anthony is so fond.

    On one post a while ago, he classified various Australian sites as urban or rural. One site classified as “urban” was Ceduna airport. Now I know its an airport, and those mighty jet engines could be heating things up, except that you’d be lucky to get more than a single-engine propellor plane there – a couple of times a day. Cars going down the road nearby would provide far more heat, because at least you’d get a few of them every hour. There are not many places on this earth more rural than Ceduna airport.

  4. Horatio Algeranon

    Verity Jones can’t tell a lie
    His name just won’t permit it
    “Of course the data were cherry-picked
    How else could we fit it?”

  5. Can you do a plot confidence intervals for the Arctic temp plot from the 15 stations? I think that part of Verity Jones’ point is that supposedly the stations which do extend far enough back for a good comparison do not show as much of a difference between the two periods… Yes I know her argument is useless but it would be interesting to see a graph of the combined longest records (say 70 years or greater).

  6. Based solely on the first name (which is also my daughter’s middle name) I vaguely imagine that Verity Jones warrants feminine pronouns.

  7. Haha.. Nice… I love to see anphoney- n-gang get exposed for the fools they are… Very entertaining..

  8. Lying liars lie.

    Film at Eleven.

  9. Something tells me Verity’s next response will be written in a profound state of denial. Either there will be no reply whatsoever or she’ll happily post more crap not worrying about the quality – because she feels she is right anyway.

    *getting pocorn*

  10. I read Verity’s ‘rebuttal’ at WUWT, and the comments. I doubt that very many of the commenters, if any, even read Tamino’s post here. If they did, and looked carefully at the graphs, knowing that Tamino is a very experienced statistician… they wouldn’t be spouting all that ideological drivel and instilling completely misplaced faith in all the anti-science that appears on WUWT, would they? Oh… wait.

    But that’s what I’ve found through experience over the past few years. These people won’t even read anything on a supposed ‘warmista’ site. They are uneducable, stuck forever in denier echo chambers like WUWT, Claimate Audit, Bishop Hill, Climate Etc., et. al.

    BTW, I did a Google image search for ‘verity jones’, and even though most of the results were for an actress, quite a few pictures of a guys did show up. I’m assuming Tamino did a bit of research into Verity’s background, or is familiar with him from previous WUWT posts.

    [Response: No, I don’t really know whether the Verity Jones who posted on WUWT is male or female. Really: does it matter?]

  11. How can you compute averages using series on different time windows? For example, from what I saw in the opponent’s post, Whitehorse data kicks in around 1940, but it’s temperature seems to be ten degrees warmer than the rest. That could explain the behaviour of the second graph, which goes up around 1 degree that year. If those averages are not simple averages, as I suspect, and there is some data manipulation involved to «normalize» the windows or whatever before computing them, how can I trust the graph? (Or btw any other graph posted by members of any of the two teams).

    [Response: It is possible to align station records in order to reveal the best estimate of the temperature change over time, even when their time spans are not coincident, by determining an offset which brings them into closest possible agreement during their period of overlap. See this.]

    • I guess I’ve got a few minutes to expand a bit on Tamino’s answer, although he has more expertise.

      If you want to argue that any method more complex than a simple average is suspect, then that’s a big claim, and invalidates large portions of statistics. And, more relevantly to the current discussion, it invalidates the instrumental temperature record. If we can’t align station records by some method, then we are dependent on complete station records. Which means can only produce a global temperature record going back to ~1950. If you want the whole of the 20thC, you can only produce local records for Europe and the US.

      However, there are a number of well established techniques for dealing with this problem (and they are not specific to climate science). The simplest is to use anomalies based on some common period. In the ITR, 1950-1980 works well. You pick a period which is present in most of the records (possibly excluding some stations which don’t cover that period), and align all the stations by adding an offset to set their mean on that period to zero. (Detail – you actually do it by month). That works well.

      A more elegant method which doesn’t involve dropping those stations which don’t cover the alignment period is to calculate an initial average over the stations, and then offset each station by a constant to best match the average over the period for which that station has data. Calculate a new average, and iterate until convergence.

      Better still is a mathematical method of doing the above method in a single step. Nick Stokes describes it nicely here:
      I think that is similar to the BEST method which Tamino now uses.

      Can you trust these methods? Well, there are a number of easy checks:

      1. They all give the same answers.

      2. You can take a set of complete station records and chop each station up into several non-overlapping fictional stations. Feed these stations into any of the above methods, and as long as there some subset of stations exist which overlap each of the cuts, you get the same results.

      3. If you take a region which is well sampled, and compare a regional temperature from a simple average over a set of complete stations with the temperature from a set of incomplete stations aligned with the above methods, you get the same results to within the level of error expected given the number of stations.

      In short, if you don’t trust the methods, test them. But this sort of method absolutely unremarkable and used across many fields.

  12. You say the normalization by Caryl is wrong but don’t explain why. So I went and read his explanation and I think I see the problem although I almost missed it. He ought to compute the anomaly for each year from the average for the individual station, but he is comparing it to the average of the averages of all the stations. (Also he uses annual averages instead of monthly, but I don’t think that’s the problem.)

    I’m having trouble sorting out what this does though.

  13. Just as Watts was complaining about Skeptical Science having two Penguins looking at a plant growing through ice on the site’s header as proof that SkS is innacurate it seems Denialists are moving into febrile fields of diversion, now with ludicrous over cooked spagetti graphs.

  14. This is a very interesting discussion and some how seems to indicate we actually know the history of the Artic. Im not a scientist but studied the Artic as part of my undergaduate degree at UBC. We read the records of explorers, settlers and church officials. The Inuit have no written record so much of the information was anecdotal but the British Navy has preserved logs from some of its vessels, specially those involved in searching for Franklin. I don’t think anyone can claim certainty of trends in the Artic[I’m talking Canadian] because we have so little scientific data to rely on.

  15. Yes, but the instrumental record started in the late 1940’s. Five stations in the Artic. A joint US/Canada venture. The graphs indicate temperature constucts from 1900’s. Far North, Northern Canada, Isolated Stations. Wouldn’t want to hang my hat on anything before 1950.

  16. And just how far do those stations go back and are they supposed to give an accuate reading of the Canadian Artic? If you read the literature, Canada ignored the Artic for at least half the 20 Century. There was no way to get to it, only ships during the summer months, technology wasn’t available, vast areas very sparsly populated etc. I live in British Columbia and official records of temperature didn’t start till 1948.

    • Is it too much trouble to look at the link? The dates of record are listed.

      The oldest appears to be Upernavik, with records back to 1880; Barrow appears to be next, starting in 1901. Several others date from the teens, twenties or thirties.

      Would we like more and better data? Of course.

      Are we completely in the dark about the Arctic prior to 1950? No.

  17. Pete Dunkelberg

    OT – Reason says climate change keeps you alive (paraphrase)

    Click to access deaths_from_extreme_weather_1900_2010.pdf

  18. Upernavik is in Greenland. !880-1980. All the posted graphs previous to the comments section are Canadian. Thats why I stuck with Canada. If the science rests on a few far flung outposts with no correlation or agreement on temperature protocals, then how do you compare with any certainty? Even reading the past reconstucts from NOAA on North America and Alaska, indicate the difficulty of relying on some sites. But back to Canada, this from the ASTIS. 1951. K.Hare and M. Montgomery. Ice, Open Water, and Winter Climate in the Eastern Artic of North America. “In December temperatures on Disko Island [West Greenland] are some 25 degrees F. warmer than Clyde River [Baffin Island] even though shore lines are just 375 mi apart”. Again, if you think the data is at all consisitant, there is a raft of reports stating how difficult it was to get constant temperature records.

    [Response: You seem to have completely missed the point. The graphs which show so few station records for northern Canada have been restricted to stations which were hand-picked — cherry-picked — to give a desired preconclusion. They do not represent the number of available stations in northern Canada, only the number chosen to give this misimpression. The GHCN alone has over 30 station records from northern Canada with data prior to 1945, and 20 stations with data as early as 1930.

    Furthermore, your reference to temperature differences at stations only 375 mi. apart is utterly meaningless. It’s not necessary for locations to show the same temperature in order for their trends to be related — it’s only necessary for their long-term temperature *changes* to be highly correlated. This has been proved, time and time again, to hold for stations as much as 1200 km distant. The number of station records in norther Canada is more than sufficient for us to know whether or not it was warmer there in the 1930s-1940s than now. The evidence indicates it was not — and this is true even if you only use stations which are cherry-picked by fake skeptics to give that false impression. That’s the point of this post.

    As for the actual temperature change in the Arctic, not restricted to Northern Canada, see this.]

  19. Hey guys, I’m trying to understand the science not take a position one way or another. No wonder ts difficult for lay people to be part of the conversation. Everyone gets their panties in a knot when some questions are asked. Then where do I go to understand what constitues a station? What is the definition of a station?

    [Response: You seem to be the one who decided to “gets their panties in a knot” — when your false claims and implications were confronted. Here are some of the things you said in earlier comments:

    “Franklin. I don’t think anyone can claim certainty of trends in the Artic[I’m talking Canadian] because we have so little scientific data to rely on.”

    “Yes, but the instrumental record started in the late 1940′s. Five stations in the Artic.”

    “And just how far do those stations go back and are they supposed to give an accuate reading of the Canadian Artic? If you read the literature, Canada ignored the Artic for at least half the 20 Century. There was no way to get to it, only ships during the summer months, technology wasn’t available, vast areas very sparsly populated etc. I live in British Columbia and official records of temperature didn’t start till 1948.”

    “If the science rests on a few far flung outposts with no correlation or agreement on temperature protocals, then how do you compare with any certainty?”

    Making false statements is a poor way to “not take a position one way or another,” and certainly no way to understand the science.]

  20. Making false statements is a poor way to “not take a position one way or another,” and certainly no way to understand the science

    Not to mention evidence that he’s been cutting and pasting from at least one denialist site. A presumption also supported by …

    But back to Canada, this from the ASTIS. 1951. K.Hare and M. Montgomery. Ice, Open Water, and Winter Climate in the Eastern Artic of North America.

    Randomly quoting one and only one scientific paper which we all know he didn’t find and read on his own …

  21. Allan wrote:

    I live in British Columbia and official records of temperature didn’t start till 1948.

    Could you please explain how or where you got this idea? I have no trouble finding data from prior to 1948 at Environment Canada’s National Climate Data and Information Archive.

  22. Have any of you actually been to the Arctic? I used to go up all the time and buy Inuit Art. If you’ve never been to Baffin Island, well worth the trip Fascinating area. Thats when I got interested in the history. As for making false statements, cut and paste, denialist etc, why would dhozaga impugn my motives? A rather infantile and puerile approach to discusssion. Again, where do I go for the information regarding stations? The Canadian historical records are obviously missing some rather pertinent information.

    [Response: Information on the latest version of the GHCN is here. Environment Canada has a National Climate Data and Information Archive here.]

  23. Philippe Chantreau

    “why would dhozaga impugn my motives?’
    Stating you can’t find records prior to 1948 while these are readily available, the lack of even most basic research that would have led you to the places linked by Tamino, and the statements highlighted earlier all seem to point in one direction.

  24. Philippe, I have looked at data from the archives at Enviornment Canada and yes there is data but a lot of blanks and not consistant. What I’m trying to understand how do you decide what constitutes a station and how you statistically agree on what is solid data. I took a statistics course many years ago and do not claim competency. When I used to spend time in the Arctic, 70’s, the only other southerners where government employees, usually with Aboriginal Affairs, Enviornment Canada or Canada Armed Forces. You couldn’t help but discuss what projects they were working on and why they were there. But they all complained of lack of information and data on the Canadian Arctic so again, why can we be certain of the temperature constucts prior to 1948? Is it simply that that we are doing the best with what we have? Since I’m not a scientist, humour me.

    • Allan, you are asking a big question, one not really amenable to answers that are simultaneously brief and satisfying.

      For a very brief version, let’s take the example that you raised, the comparison of Clyde River with Disko Island (which, by the way, was the first mention of Greenland—not my mention of Upernavik.) Your thought was that since the absolute temperatures of the two locations in December “differed by some 25 degrees F” (as Hare and Montgomery said), then nothing could be concluded about temperature from those records.

      But that is not what Hare and Montgomery concluded. In fact, they graphed the relationship between temps in the two locations over the course of the year. What that suggests is that there is a discernible, somewhat predictable relationship between those two records—far from their disparate temps being unrelated, if you know (say) the Disko Island temperature for a particular December, you will be able to make a good guess that Clyde River’s temperature that month was close to (Disko Island minus 25 F.)

      That means that it is possible to use the existing data to “fill in” spatial and temporal gaps in coverage with a reasonable degree of confidence. The crucial bit of this is termed “station alignment.”

      How it is done in practice can get complicated, but this is a great site to learn about it. Here are some related posts from Tamino. (Note that many look at wider issues than just Canadian or Arctic data, but they do get into the aspects of station alignment and data quality-checking and analysis.)

      (Review of past analyses)

      (GHCN verification)

      (How to align station records)

      (Critiquing/compensating data problems)

      (How NOT to align station records)

      (data links; also accessible from header tab)

      (BTW, I found these links by using the search box at the bottom right of the page.)

      Don’t know if the information below will be helpful in addressing your specific interest in the Canadian data, but I’ve listed Canadian GHCN stations north of the Arctic Circle. (I pulled the list a while back—a fake skeptic was claiming there was only one of them, or something like that. The station name is in the left column, the right column is the latitude—I used Excel to sort and extract these most Northerly Canadian stations from the larger list of Canadian GHCN stations.)

      As with the Environment Canada stations you looked at, not all are current, and coverage varies. But as stated above, that’s where the process of ‘aligning’ station records comes in.

      ALERT,N.W.T. 82.5
      EUREKA,N.W.T. 79.98
      ISACHSEN,NW 78.78
      MOULD BAY, N. 76.23
      REA POINT A,NW 75.37
      RESOLUTE,N.W. 74.72
      ARCTIC BAY,NW 73.03
      NANISIVIK A,NW 72.98
      POND INLET, N 72.7
      POND INLET,NW 72.68
      HOLMAN,NW 70.73
      CLYDE,N.W.T. 70.48
      CAPE PARRY,N. 70.17
      CLINTON POINT,NW 69.58
      KOMAKUK BEACH A,YT 69.58
      SPENCE BAY,NW 69.53
      TUKTOYAKTUK,NW 69.45
      IGLOOLIK,NW 69.38
      CAMBRIDGE BAY 69.1
      ROWLEY ISLAND,NW 69.08
      SHINGLE POINT 68.95
      CAPE YOUNG A,NW 68.93
      LONGSTAFF BLU 68.88
      SHEPHERD BAY, 68.82
      HALL BEACH,N. 68.78
      BYRON BAY A,NW 68.75
      GLADMAN POINT A,NW 68.67
      DEWAR LAKES,N 68.65
      LADY FRANKLIN 68.5
      CAPE HOOPER, 68.47
      PELLY BAY,NW 68.43
      MACKAR INLET,NW 68.3
      INUVIK,N.W.T. 68.3
      AKLAVIK A,NW 68.22
      COPPERMINE,N. 67.82
      OLD CROW A,YT 67.58
      CAPE DYER,N.W 66.58
      EAGLE PLAINS,YT 66.42

      Good luck in your efforts to understand this issue.

  25. OK, the latitudes didn’t display as a column. Could have saved the time spent physically aligning them in the message box, I guess. . . Gotta remember that for next time.

    Shouldn’t be much of a readability issue, though.

  26. Philippe Chantreau

    Not being a scientist isn’t grounds to be humored. Making statements such as “the instrumental record doesn’t start until the late 40’s” certainly is.

  27. Thanks Kevin, back to school I go.