West Antarctic Stations

As noted yesterday, the main enterprise of Steig et al is a re-interpretation of West Antarctica, saying that “no one really has paid much attention to” this area in the past and saying that their study improved over prior studies because previously:

“People were calculating with their heads instead of actually doing the math,” said lead author of the study Eric Steig, of the University of Washington in Seattle. “What we did is interpolate carefully instead of just using the back of an envelope.”

I observed yesterday that NASA GISS used the same source data for surface and AWS stations as Steig (READER) but were unable to calculate trends in West Antarctica (blank on the trend map -see yesterday’s post), a “defect” that Steig attributed above to Hansen’s group “calculating with their heads instead of actually doing the math” and “just using the back of an envelope”. Hansen’s bulldogs, Gavin and Tamino, who usually defend Hansen against the smallest slight or perceived slight, are strangely silent to Steig’s damning accusations against Hansen.

Today I’m going to look at the station data to see whether Steig’s accusations against Hansen are fair or not.

Figure 1 below is a location map of stations listed in Steig Table S2. The green boundary is more or less the outline of the “West Antarctica” hot spot in Steig’s graphics. There are 5 dots in the GISS-empty portion of this area, which I’ll discuss after you have a chance to locate the dots: 3 blue, one green and one red.


Figure 1. Location of Steig Table S2 Stations

Four of the stations are AWS stations shown in the bottom of Steig Table 2, excerpted below: MT Siple, Siple, Byrd and Harry. There are two surface stations: Byrd (in the same spot as Byrd AWS which operates a number of years after the surface station) and Russkaya (on the coast). I had a little trouble with my original plot of station locations because I used the information from Steig’s SI shown below.


Excerpt from Steig Table 2. First two columns are lat and long.

In Steig’s location table, Mt Siple and Harry AWS stations have identical latitude and longitude (first two columns). In fact, they have different locations: Mt Siple is shown with its correct location (a blue dot on the coast), but Harry’s coordinates are wrong. Harry is the green dot in this map. Are these coordinates incorrect only in the Nature SI? Dunno. Would it even “matter” in Mannian methodology if they were incorrect? Dunno. It hasn’t “mattered” in any other Mannian effusion and, in this case, it might not matter if Harry were on Broadway or in Antarctica. It’s hard to say. You’d think that they’d be on the look-out for geographic mislocations given Mann’s prior rain in Maine/rain in Spain problems, but I guess not.

Below is some information (including looked up GISS numbers) on these six West Antarctic stations, that appear to underpin the Steig reinterpretation for their “AWS” reconstruction. I looked up these 6 data sets in GISS dset0, dset1 and dset2. Hansen only qualified one of the six stations as meeting his dset2 standards (Byrd 1957-1975.) Thus, in Hansen’s view, the data quality problems (coverage, length of record etc) with Russkaya surface and Mt Siple, Siple, Byrd and Harry AWS stations all disqualified them from dset2 status. In addition, the coverage period for Byrd surface dset2 annual average was only from 1957-1970, thus not qualifying for a 1979-2003 trend. That’s why West Antarctica is blank in the GISS trend map.

name lat long id start end gissid
8 Byrd -80.0 -120.0 89125 1957.000 1975.000 70089125000
35 Russkaya -74.8 -136.9 89132 1980.250 1990.083 70089132000
43 Mt_Siple_AWS -73.2 -127.1 89327 1992.167 2006.000 70089327000
44 Siple_AWS -75.9 -84.0 89284 1982.000 1992.167 70089284000
45 Byrd_AWS -80.0 -120.0 89324 1980.167 2008.750 70089324000
46 Harry_AWS -73.2 -127.1 21355 1987.000 2002.917 70021355000

Let’s now turn to Steig’s allegation that Hansen (like all of Steig’s predecessors) “calculated with his head instead of actually doing the math” and that he “just used the back of an envelope.” While I may seem like an unlikely defender of Hansen and GISS, fair’s fair. I see no evidence to support Steig’s allegation. On the contrary, there is clear evidence that Hansen put the information for all six stations into the GISTEMP algorithm. Whether the GISTEMP algorithm is any good is a different question, but Steig’s characterization of the process, appealing as it might be to Hansen critics, is surely uncalled for.

Here’s a plot of the READER version of the six stations. Given the importance that Steig placed on re-interpreting West Antarctica, it would have been nice if he had included a plot of the data, as even RegEM covariance matrices ultimately are related to data. At a first glance, this doesn’t seem like a whole lot of data to work with.


Figure 3. Station Data: 6 West Antarctica Stations

Steig’s recon_aws reconstruction offers anomaly reconstructions for 63 AWS sites, including the 4 West Antarctic AWS sites, which are plotted below, with their slopes over the 1979-2003 (used in the color trend maps yesterday. One of the four sites (Mt Siple) has a slightly negative trend; two of three (Siple, Byrd AWS) have modestly positive trends; while one site (Harry) – the one with the incorrect location – has a very strong trend of 0.81 deg C/decade.


Figure 4. Steig reon_aws reconstructions

Stay tuned for some interesting news about Harry.


88 Comments

  1. Allen63
    Posted Feb 1, 2009 at 3:55 PM | Permalink

    You’ve got my attention — on Super Sunday yet. Interested to see the explanation for 0.81C/decade — and how important that was to the entire Steig conclusion.

  2. Dishman
    Posted Feb 1, 2009 at 4:00 PM | Permalink

    According to this link, AWS Harry was installed in November 1994.

    Eyeballing it, there’s a curious discontinuity and pattern change in the data at about that date.

  3. Steve McIntyre
    Posted Feb 1, 2009 at 4:01 PM | Permalink

    Dunno how “important” it is. It may not be possible to figure out exactly what they did without code. But it looks like there’s trouble with Harry.

  4. Steve McIntyre
    Posted Feb 1, 2009 at 4:03 PM | Permalink

    #2. Yep. But it’s going to be even better than just that. BTW if anyone can figure out the provenance of Harry data http://www.antarctica.ac.uk/met/READER/aws/Harry.All.temperature.html from 1987-1994 – prior to its reported installation in Nov 1994 http://uwamrc.ssec.wisc.edu/aws/harrymain.html – I’d be very interested.

  5. Joel
    Posted Feb 1, 2009 at 4:07 PM | Permalink

    A little humorous insight for Super Sunday. This episode reminds me of this scene from “Rocket Man” where the geeky Fred Randle shows the NASA higher-ups that they would get better results if they performed the calculations “the right way”.

  6. Urederra
    Posted Feb 1, 2009 at 4:10 PM | Permalink

    I have just made a movie title search on IMDb and I selected these titles:

    Deconstructing harry.

    Harry un ami qui vous veut du bien aka with a friend like Harry…

    The trouble with Harry.

    Dirty Harry.

    Which one shoud Steve use as the title for his next blog post about Harry station?

    • Dave Dardinger
      Posted Feb 1, 2009 at 4:40 PM | Permalink

      Re: Urederra (#6),

      Which one shoud Steve use as the title for his next blog post about Harry station?

      Personally, what came to my mind immediately was the old song, “I’m just wild about Harry.”

  7. jason
    Posted Feb 1, 2009 at 4:13 PM | Permalink

    Come on, Steve, cut the “damning accusations against Hansen” – snip

    Steve: You’re being misleading here. I made no “damning accusations against Hansen” – I defended GISS. It ws Steig who made the following accusation against Hansen and other developers of Antarctic temperature data :”People were calculating with their heads instead of actually doing the math,” said lead author of the study Eric Steig, of the University of Washington in Seattle. “What we did is interpolate carefully instead of just using the back of an envelope.” I think that this accusation against his predecessors can be fairly described as “damning”. Surely you will agree that the claim that GISS wasn’t “doing the math” is ‘damning.’ Perhaps you can suggest another term. Imagine the furor if I accused Hansen of “just using the back of an envelope”. Why shouldn’t Steig be accountable for making these claims against Hansen? Here I am – providing bulldog services for Hansen and you’re giving me a hard time.

  8. pete m
    Posted Feb 1, 2009 at 4:13 PM | Permalink

    There can only be one movie title (or franchise) which fits:

    “Back to the Harry”.

    Harry was there before he was there.

  9. Dishman
    Posted Feb 1, 2009 at 4:19 PM | Permalink

    The link I gave lists Harry as having an Argos ID of 8900…
    I believe that’s also the Argos ID of Siple.

    It looks like there’s a problem with the metadata, though resolving it is beyond me.
    It appears that the Harry data above resolved the metadata problems by splicing Argos 8900 data prior to 11/94 with Harry data after that.

  10. Posted Feb 1, 2009 at 4:20 PM | Permalink

    I was thinking “Harry And The Hendersons”, with Big Foot and all.

  11. janama
    Posted Feb 1, 2009 at 4:34 PM | Permalink

    here’s the correct location of Harry Steve.


    Steve: That’s what I showed in my location map -see the green dot.

  12. janama
    Posted Feb 1, 2009 at 4:39 PM | Permalink

    and here are the people who look after it.

    http://www.polartrec.com/files/beckendorf_Webinar_011508_opt.pdf

  13. ianl
    Posted Feb 1, 2009 at 4:51 PM | Permalink

    Figure 1 needs a clear scale bar to allow readers to aquire a feel for the areal influence of each data point.

    Not different from calculating the areal influence of drillhole data points. So I would definitely not buy mining shares from a prospectus written by Steig et al.

  14. Steve McIntyre
    Posted Feb 1, 2009 at 4:54 PM | Permalink

    #14. I presume that they have “teleconnections” in RegEM so area of influence might not be as straightforward as all that. It’s been a huge job in other Mannian studies trying to figure out the weight of individual data points and I’m nowhere near being in a position to do that here.

  15. Steve McIntyre
    Posted Feb 1, 2009 at 4:55 PM | Permalink

    #9. Argos ID number 8900 seems to move around.

    test=list(NA,30)
    for (i in 1:30) {
    year=1980+i
    loc=paste(“ftp://ice.ssec.wisc.edu/pub/aws/climate/”,year,”/climo8900_”,year,”.dat”,sep=””)
    test[[i]]=try(readLines(loc))
    }
    sapply(test, function(A) A[1] )

    [1] “Error in file(con, \”r\”) : cannot open the connection\n”
    [2] “Station Argos ID: 8900 Station Name: D-10 Year: 1982″
    [3] “Station Argos ID: 8900 Station Name: D-80 Year: 1983″
    [4] “Station Argos ID: 8900 Station Name: D-80 Year: 1984″
    [5] “Station Argos ID: 8900 Station Name: D-80 Year: 1985″
    [6] “Station Argos ID: 8900 Station Name: ALLISON Year: 1986″
    [7] “Station Argos ID: 8900 Station Name: MARTHA 2 Year: 1987″
    [8] “Station Argos ID: 8900 Station Name: MARTHA 2 Year: 1988″
    [9] “Station Argos ID: 8900 Station Name: MARTHA 2 Year: 1989″
    [10] “Station Argos ID: 8900 Station Name: MARTHA 2 Year: 1990″
    [11] “Station Argos ID: 8900 Station Name: MARTHA 2 Year: 1991″
    [12] “Station Argos ID: 8900 Station Name: MARTHA 2 Year: 1992″
    [13] “Station Argos ID: 8900 Station Name: ELAINE Year: 1993″
    [14] “Station Argos ID: 8900 Station Name: Elaine Year: 1994″
    [15] “Station Argos ID: 8900 Station Name: Elaine Year: 1995″
    [16] “Station Argos ID: 8900 Station Name: Elaine Year: 1996″
    [17] “Station Argos ID: 8900 Station Name: Siple Dome Year: 1997″
    [18] “Station Argos ID: 8900 Station Name: Siple Dome Year: 1998″
    [19] “Station Argos ID: 8900 Station Name: Harry Year: 1999″
    [20] “Station Argos ID: 8900 Station Name: Harry Year: 2000″
    [21] “Station Argos ID: 8900 Station Name: Harry Year: 2001″
    [22] “Error in file(con, \”r\”) : cannot open the connection\n”
    [23] “Error in file(con, \”r\”) : cannot open the connection\n”
    [24] “Error in file(con, \”r\”) : cannot open the connection\n”
    [25] “Error in file(con, \”r\”) : cannot open the connection\n”
    [26] “Error in file(con, \”r\”) : cannot open the connection\n”
    [27] “Error in file(con, \”r\”) : cannot open the connection\n”
    [28] “Error in file(con, \”r\”) : cannot open the connection\n”
    [29] “Error in file(con, \”r\”) : cannot open the connection\n”
    [30] “Error in file(con, \”r\”) : cannot open the connection\n”
    >

  16. Dishman
    Posted Feb 1, 2009 at 5:12 PM | Permalink

    Here shows:
    Harry 21355 83.003oS 121.393oW 945 Nov 94
    Siple Dome* 8900 81.656oS 148.773oW 620 Jan 97 89345
    * New locations for 1997 CRS, 25 March 1997

    Here shows:
    Harry 8900 83.003oS 121.393oW 945 Nov-94
    Siple Dome 8938 81.656oS 148.773oW @668 Jan-97 89345

    Both are from amrc.ssec.wisc.edu

    • AnonyMoose
      Posted Feb 1, 2009 at 9:35 PM | Permalink

      Re: Dishman (#17), in his first link (and a few comments earlier) he observed the ID change before I did. His first URL is the same field report in which I independently discovered the note of the movement of AWS 8900 from Elaine to the Siple Dome site. Just wanted to give Dishman credit for noticing ID oddities.

  17. Pedro S
    Posted Feb 1, 2009 at 5:16 PM | Permalink

    Why didn’t anyone mention “When Harry met Sally?” ;-)

  18. Hu McCulloch
    Posted Feb 1, 2009 at 5:38 PM | Permalink

    The U Wisc webpage found by Dishman above has photos of the AWS before and after it was dug out of the snow and raised to the new surface level on 1/12/06.

    The obvious title for this mystery is, “The Trouble with Harry, The Station That Wouldn’t Stay Buried.”

    • BarryW
      Posted Feb 1, 2009 at 10:28 PM | Permalink

      Re: Hu McCulloch (#19),

      Buried and dug up, oh it’s got to be called “The Trouble with Harry”, since that’s part of the movie’s plot. But who killed Harry?

  19. jarhead
    Posted Feb 1, 2009 at 5:42 PM | Permalink

    “The Trouble With Harry” Alfred Hitchcock 1995

  20. AnonyMoose
    Posted Feb 1, 2009 at 5:50 PM | Permalink

    Steve: Yup, the same ARGOS ID 8900 has been used for several stations. I noticed MARTHA 2 (Martha II) because this data directory has MARTHA 2 data as 8900. The AWS stations page lists Martha II and Harry, but I don’t see Sally.

  21. Geoff Sherrington
    Posted Feb 1, 2009 at 5:52 PM | Permalink

    Why are some many of the data points pefore 1990 exactly the same at Byrd and Harry? Who composed the “different” ones?

  22. AnonyMoose
    Posted Feb 1, 2009 at 5:54 PM | Permalink

    Odd. this 1996-1997 field report (.DOC) mentions “AWS 8900 was replaced with AWS 8915.” at the Elaine site, and a few paragraphs later says AWS8900 installed at a site related to Siple Dome.

    • AnonyMoose
      Posted Feb 1, 2009 at 8:54 PM | Permalink

      Re: AnonyMoose (#23), links to a 1996-1997 field report which says “Harry 21355 83.003oS 121.393oW 945 Nov 94″ where the second column, 21355 is “ARGOS ID”. So ID 1355 is only part of the ARGOS transmitter ID. That particular report does not mention any work done on Harry. The 1999-2000 field report does not mention specific work done on Harry, but searching for Harry shows that the ID changed to 8900.

  23. AnonyMoose
    Posted Feb 1, 2009 at 6:03 PM | Permalink

    Oh. ARGOS is a satellite data service. The ARGOS IDs are probably transmitter IDs. The transmitters have been moved to several sites over the years.

  24. Posted Feb 1, 2009 at 6:11 PM | Permalink

    Here’s the AR4 GCM indication (1958-2002) for Harry –

    He looks like one of the boys, nothing extraordinary, per the 11-model composite.

  25. PaddikJ
    Posted Feb 1, 2009 at 6:51 PM | Permalink

    Let’s see if I have this right: The only station to show unequivocal warming is the one in a duplicate & incorrect location? And even in its correct location, it’s been buried until recently?

    Oops.

    I’d say send Mann back to the drawing board, except there’s little to indicate he was ever there to begin with (yes, I realize Steig was the lead author, but it looks like Mann who’s making all the noise; interesting that in the Live Science article, Mann is quoted 2-3 times before we learn that Steig was the lead author).

  26. Jason
    Posted Feb 1, 2009 at 7:07 PM | Permalink

    I think its very premature to assume that:

    1. The Steig et al results are substantially dependent on data from a single station or
    2. Steig et al made some sort of gross error by including this station or were otherwise unaware of the station’s history.

    Obviously Steve isn’t saying this, but some of the posters in this forum are clearly jumping the gun.

    • Terry
      Posted Feb 1, 2009 at 7:24 PM | Permalink

      Re: Jason (#27),

      What’s the real story, Jason? With data this time, please. Thanks!

      • Jason
        Posted Feb 1, 2009 at 7:59 PM | Permalink

        Re: Terry (#30),

        I wish that climate scientists would learn to admit when there isn’t enough data to draw a conclusion.

        Perhaps you could use the same lesson?

        • Terry
          Posted Feb 1, 2009 at 8:05 PM | Permalink

          Re: Jason (#31),

          We’re of the same opinion. And this is Steig’s quote, not mine, or anyone else I’ve read on this site:

          “My advice is stop reading about the “debate” because there isn’t one.”

          Sounds familiar, if not very scientific, eh?

  27. Jason
    Posted Feb 1, 2009 at 7:15 PM | Permalink

    Can you imagine being Steig? According to his web page, he is stuck on James Ross Island (just off the coast of the Peninsula) until March 19th. His limited internet access prevents him from receiving email larger than 30kb.

    If I had a major paper published while stuck on Antarctica, and I saw the discussion on this site, I think I would go a little crazy too.

    He should be forgiven for his initial reaction if, a month after he comes home, he opens access to all of his methods and data.

  28. DaveCF
    Posted Feb 1, 2009 at 7:20 PM | Permalink

    The three tunnels dug in the Great Escape from Stalag Luft III were Tom, Dick and Harry. It does appear that someone did some rather enthusiastic digging to get Harry to Antarctica. Maybe there is a teleconnection to Sagan, Germany?

  29. AnonyMoose
    Posted Feb 1, 2009 at 8:05 PM | Permalink

    Summary of Harry:
    * Steig Table 2 listed the wrong location for Harry.
    * Harry was installed November 29, 1994.
    * Harry is identified as AWS 8900.
    * U of Wisconsin READER data is stored based on what is variously called “UW ID” or “AWS ID”.
    * AWS 8900 has been used by several stations.
    * Data is transmitted through the ARGOS service, so “AWS 8900″ is probably a data transmitter label.
    * READER 8900 data needs to be filtered based on what site is using AWS 8900.
    * Thus, it is erroneous in Figure 3 to use READER 8900 data before 1994-11-29 for Harry.
    * Thus, there is an error in Figure 4 because it has Harry data before 1994-11-29.
    * If Steig’s recon_aws is all of the west Antarctica stations which he used for calibrating the satellite data in the region, the study has lost its strongest-warming station in the region.
    * All of the released data for the study should be archived, for comparison with the next study.

  30. Bruce
    Posted Feb 1, 2009 at 8:10 PM | Permalink

    Interesting. While Harry’s 3 hour data ends in 2002, his 10 minute interval data continues at aleast until October 2008.

    http://amrc.ssec.wisc.edu/aws/archive/Harryarchive.html

  31. AnonyMoose
    Posted Feb 1, 2009 at 8:28 PM | Permalink

    As Steve noticed above, 8900 data contains Harry data starting in 1999. Station notes for Harry in 1999 begin “A new station was installed on 26 January.” The AWS 8900 transmitter was probably moved from Siple Dome to Harry then. Archived Harry data switches from prefix 1355 in 1996 (note data gap) to 8900 in 1999. Yup, “Station Argos ID: 1355 Station Name: Harry Year: 1996″ is the header on ftp://ice.ssec.wisc.edu/pub/aws/climate/1996/READER1355_1996.dat so Harry was ID 1355 until it stopped in 1996.

    Steve linked to http://www.antarctica.ac.uk/met/READER/aws/Harry.All.temperature.html (note READER in the name) which has some incorrect time periods for Harry, so someone (who probably is not at the data-gathering group at U of Wisconsin) apparently grabbed all the READER data by ID.

  32. MattE
    Posted Feb 1, 2009 at 8:36 PM | Permalink

    Steve, Have you seen this presentation from U Wisc. online: http://www.mmm.ucar.edu/events/antarctic06/presentations/weidner_aws.ppt#17?
    I found it and it looks like Harry’s been mostly burried at some point. I don’t know how often these sites get visited, or if the temp sensors remained above the snow or not (I’m no Anthony Watts). That could be an explanation for why data looks “insulated” after ~2000 (highs not as high, lows not as low). This and another site indicate Harry was “raised” in 2006. Some problems seem well nailed by other posters, but there may be even more to it.

    Matt

  33. MattN
    Posted Feb 1, 2009 at 8:37 PM | Permalink

    This is an unusual situation for Hansen because his “accounting” of temps at Antarctica is called into question, but he likes the results. So why would he really argue with it?

  34. Jeff Alberts
    Posted Feb 1, 2009 at 8:41 PM | Permalink

    “When Harry Met Siple”?

  35. Jeff Alberts
    Posted Feb 1, 2009 at 8:42 PM | Permalink

    Or more like “When Harry Became Sybil”

  36. AnonyMoose
    Posted Feb 1, 2009 at 9:17 PM | Permalink

    At the start of this article our discussion leader mentions yesterday’s article “Steig versus Hansen”. On the issue of Harry, Hansen wins. Yesterday’s article says that GISS mentions READER data, but the parent page of the mentioned UK Met pages links to background with data standards which Harry does not meet. Indeed, the UK Met data does not include Harry data, so GISS has not been contaminated by confusion between Harry and ID 8900. So Hansen won by proxy, due to the UK Met READER data omitting Harry data.

  37. AnonyMoose
    Posted Feb 1, 2009 at 9:45 PM | Permalink

    Apologies for referring to the British Antarctic Survey links as “UK Met”. No known connection between Harry’s data and the people running the UK’s Met Office. They’re both promoting AGW, but I don’t know of the Met Office being involved with this data processing.

  38. Geoff Sherrington
    Posted Feb 1, 2009 at 10:19 PM | Permalink

    Elsewhere I have mentioned a PC comparator process that is like putting a photo negative over its positive print. All being equal, you get black. In this computer case you get mid-grey. Any similar overlaps merge with the grey and drop out of sight.

    Using Steve’s graphs above, I overlaid Byrd (white) on Harry (black). TOP GRAPH. There are 4 time intervals. 1. Start to 1980, the same data seem to have been used, with an “adjustment” so Harry goes colder. 2. 1980-2003, not much drops out, so the data probably came from different ideas or places. 3. The last part of 2, 2000-2003, where dirty Harry goes from strongly cool to strongly warm, thus influencing the slope. 4. Period 2003-2007, the same data seem to have been used for both stations, with one set given a bi of noise here and there.

    Bottom graph. Byrd nudged one pixel sideways so that both graphs can be seen and any (few) dropouts are a coincidence.

    Using my trusty old atlas, I could not find coordinates that agree with some station locations either. Ditto Google Earth.

    These data are most dubious and hardly a central plank in a global warming claim.

    p.s. I have trouble getting the Image facility to work, so just use the URL if no image comes up.

  39. Geoff Sherrington
    Posted Feb 1, 2009 at 10:26 PM | Permalink

    Try image again

  40. Hu McCulloch
    Posted Feb 1, 2009 at 10:28 PM | Permalink

    RE MattE, #37,

    Here are the photos of Harry from the U. Wisc. AWS site before and after it was dug out of the snow and raised on 1/12/06:

    I don’t know which device is the thermometer.

    • D. Patterson
      Posted Feb 2, 2009 at 3:30 AM | Permalink

      Re: Hu McCulloch (#48),

      I don’t know which device is the thermometer.

      Re: Neil Crafter (#52),

      I’d be guessing that the thermometer package was one of the instruments buried under the snow in the first photo.

      No, the temperature sensor, is a PRT Bridge (Platinum Resistance Thermometer Bridge instrument), located inside one of the beer can looking radiation shields mounted on the underside of the crossarm in the left side of the image.

      It should be noted that the humidity instrument, the Vaisala HMP35, has had some known accuracy issues in the past.

      The temperature returned by the HMP45D Pt100 temperature sensor was
      within the Bureau specification for humidity probes (correction of less than
      ±0.2oC), but a significant number of the probes fell outside the criteria for
      AWS dry-bulb temperature measurement (correction less than ±0.08oC) and
      therefore the probes cannot be considered for AWS “dry bulb” measurements (Gorman, WMO ITR 661, 2002).

      Likewise, the PRT Bridge temperature sensors may be subject to long-term drift and hysteresis, particularly in adverse environments. Extreme sub-freezing temperatures, high humidity, ice fog, and other challenges to the specifications of the instruments justify extraordinary caution in accepting and evaluating measurement results. Special attention to calibration methods and maintenance with respect to accuracy of the measurements is required. Instrument testing in cold environments have used ice water baths to provide a basis for calibration. Presumably, the instruments used in the Antarctic environment would require specialized calibration methods suitable to meet the extraordinary challenges encountered at the Antarctic AWS sites.

  41. Steve H.
    Posted Feb 1, 2009 at 11:06 PM | Permalink

    Hansen will feel he’s being harried over Harry.

    har⋅ry   
    verb, -ried, -ry⋅ing.
    –verb (used with object) 1. to harass, annoy, or prove a nuisance to by or as if by repeated attacks; worry: He was harried by constant doubts.

    • Dave Dardinger
      Posted Feb 1, 2009 at 11:13 PM | Permalink

      Re: Steve H. (#50),

      Harry or Hanson,
      Lord of the world’s bottom end;
      you can make us warm.

  42. Posted Feb 1, 2009 at 11:37 PM | Permalink

    Is it not a little ironic that the weather station that provides the largest warming signal in W Antarctica has been pretty much buried under the snow and had to be raised from the dead?

    Perhaps an apt title might be “The Resurrection of the Harry”.

    Thanks to Hu Mc for the photos. I’d be guessing that the thermometer package was one of the instruments buried under the snow in the first photo.

  43. Steve McIntyre
    Posted Feb 1, 2009 at 11:42 PM | Permalink

    If you look at the info on other stations e.g. ice sheet stations, you’ll find both that they get buried and that they move considerable distances. Anthony will have a field day.

    But the Harry thing is different yet. It’s more of a classic Mann thing, tho not, in this case, true Mannian mead.

    • Posted Feb 2, 2009 at 12:56 AM | Permalink

      Re: Steve McIntyre (#53),

      See my comments re station history of the AWS station “Butler Island” on the RegEM thread (http://www.climateaudit.org/?p=4914-comment number 4). This station also shows one of the higher warming trends but the data is incredibly patchy to say the least and as you sat AWatts would have a field day…that is if he could find some snow shoes to fit.

    • Posted Feb 2, 2009 at 3:54 AM | Permalink

      Re: Steve McIntyre (#53),
      Steve, you may be interested in Eric Steig’s reply to a question posted on RC:

      39. I’d be interested on your comments on the quality of the AWS records. Take for example the station at Butler Island that you report has a warming trend of 0.45 degrees C per decade. GISS data shows net cooling over the record period (WTF!). But given the paucity of data points any trend appears optimistic. If you examine the station history through the University of Wisconson website there appears to be a long record of malfunctions and changes in site including changes in elevation of over 100m. The station appears to have been buried by snow a number of times. Station Elaine on the Ross Ice shelf is similarly affected.
      How do the problems with the AWS network affect your results? How much time was spent vetting the reliability of the AWS records? Did you accoutn for this in your analysis? Given their unreliability I am wondering why you bothered using AWS at all.

      [Response: The AWS records useful because they provide a totally independent estimate of the spatial covariance pattern in the temperature field (which we primarily get from satellites). I agree with you that in general the data quality at many of the AWS sites is suspect. However, the trends in our results (when we use the AWS) don’t depend significantly on trends in the AWS data (in fact, the result changes little if you detrend all the AWS data before doing the analysis. –eric]

      http://www.realclimate.org/?comments_popup=625

      • James Lane
        Posted Feb 2, 2009 at 5:03 AM | Permalink

        Re: MarcH (#63),

        Yeah, I think I’d quoted this response from Eric on the earlier thread. The AWS records are apparently “useful” but “don’t matter”.

        I’d like to suggest that everyone waits a bit until we better understand what might be the trouble with Harry. That said, as I understand it, Harry might be very important in the RegEM West Antarctic reconstruction due to its “central” location.

      • Stan Palmer
        Posted Feb 2, 2009 at 10:55 AM | Permalink

        Re: MarcH (#63),

        An Eric Steig Reply on RealClimate about sensitivity to AWS errors

        [Response: The AWS records useful because they provide a totally independent estimate of the spatial covariance pattern in the temperature field (which we primarily get from satellites). I agree with you that in general the data quality at many of the AWS sites is suspect. However, the trends in our results (when we use the AWS) don’t depend significantly on trends in the AWS data (in fact, the result changes little if you detrend all the AWS data before doing the analysis. –eric]

        It has been reported on this blog that Mann and otehr cliate researchers have performed sensitivity analyses for similar purposes on, fro example, the Mann 08 results and other findings. When these sensitivity tests are examined it is found that they are not sufficiently robust to exclude all sensitivity effects.

        Has Dr. Steig indicated if this sensitivity analysis has been described sufficiently enough for replication?

  44. Stephen Singer
    Posted Feb 1, 2009 at 11:46 PM | Permalink

    At this web site: http://ice.ssec.wisc.edu/stations.html

    The Siple Dome location had station id 8900 until Jan ’99.
    Note in 1999 there are two entries for Jan. first is id:8900 second is id:8938

    Maybe this is a clue to the puzzle about the Harry site ARGOS ID.

  45. Steve McIntyre
    Posted Feb 1, 2009 at 11:51 PM | Permalink

    #43. The problem is a little different than you think. I’ll post up what I’ve worked out tomorrow.

    • AnonyMoose
      Posted Feb 2, 2009 at 11:39 AM | Permalink

      Re: Steve McIntyre (#55),
      Yes, I see that ID 8900 does not explain all of the problems with the British Antarctic Survey data from “Harry”. But the BAS has data for several time intervals when Harry did not exist, so they are showing bad data where GISS is collecting data. I don’t know whether GISS uses all the BAS sites, so they might not be processing Harry…and whatever other bad data may exist.

  46. Steve McIntyre
    Posted Feb 2, 2009 at 12:08 AM | Permalink

    #47. Geoff, you’re looking at the “reconstructions”. There’s no actual HArry info in the early segment; Byrd is about the only station. At the end of the day, RegEM is just a complicated way of assigning weights. So Byrd is going to be used to “reconstruct” Harry and that’s probably why the early portion of recon HArry looks like Byrd.

  47. Phillip Bratby
    Posted Feb 2, 2009 at 1:21 AM | Permalink

    How many authors and peer reviwers does it take to miss an error in Table S2? It only takes one diligent independent unpaid auditor to spot the error. This says something to me about the quality of “professional” climate scientists. I use “professional” in the sense of “getting amply rewarded”, not in the other dictionary definition which is “showing the skill, or standard of conduct appropriate in a member of a profession”.

  48. J.Hansford.
    Posted Feb 2, 2009 at 1:34 AM | Permalink

    Maybe a Harry, but definitely no Houdini. There will be no getting outta this… ; )

  49. Paul S
    Posted Feb 2, 2009 at 1:50 AM | Permalink

    My title suggestion is a play on the name of the Cameron Diaz/Ben Stiller movie:
    “There’s Something About Harry”.

  50. ianl
    Posted Feb 2, 2009 at 3:49 AM | Permalink

    Figure 1 needs a clear scale bar to allow readers to aquire a feel for the areal influence of each data point.

    Not different from calculating the areal influence of drillhole data points. So I would definitely not buy mining shares from a prospectus written by Steig et al.Re: Steve McIntyre (#15),

    No, I wasn’t suggesting you do it, Steve

    Typically of point-located data, there are clusters in limited geographical areas and very much sparser data points over much larger distances. I’ve seen such data distributions on every mining project I’ve ever been involved with or even looked at. Distributing weighted values needs clear methodology, which I gather we haven’t uncovered yet.

    Again by coincidence, in January I was actually in the area (West Peninsula). Say what we may about the 19th century whalers, they made excellent navigation maps. The glaciers on the Peninsula have been retreating for over 200 years.

  51. Peter
    Posted Feb 2, 2009 at 6:01 AM | Permalink

    Hopefully RegEM did’t weight Harry too heavily……before he got there.

  52. Tim C
    Posted Feb 2, 2009 at 7:07 AM | Permalink

    Harry has IDENTICAL data relative to Gill for the period Jan 1987 through June 1989.

    http://www.antarctica.ac.uk/met/READER/aws/Gill.All.temperature.txt

    1987 -7.1 -16.5 -35.4 -38.4 -38.5 -31.4 -34.3 -40.2 -38.6 -27.3 -20.4 -6.7
    1987 -7.1 -16.5 -35.4 -38.4 -38.5 -31.4 -34.3 -40.2 -38.6 -27.3 -20.4 -6.7

  53. Steve McIntyre
    Posted Feb 2, 2009 at 7:14 AM | Permalink

    #66. Yep. And not just for 1987-89. Right up 1994. The ramifications (and how I came to spot that) will be today’s post: When Harry Met Gill.

  54. Craig H.
    Posted Feb 2, 2009 at 10:10 AM | Permalink

    Perhaps the appropriate movie would be “Dirty Harry” :

    “I know what you’re thinking- “Was Harry around for 50 years or only five?” Well, to tell you the truth, in all this excitement, I’ve kinda lost track myself. But, being this is RegEM, the most powerful climate analysis package in the world and would blow your head clean off, you’ve got to ask yourself one question: “Do I feel lucky?” Well, do ya, punk?”

  55. George M
    Posted Feb 2, 2009 at 10:21 AM | Permalink

    Slightly OT:
    Is there a good description of the physical installation of these Antarctic AWS stations anywhere? I see numerous things in the two photos which Hu lifted from the U of Mich site which I would like to get clarified. Like what is on the bottom end of those chains? And does the snow just keep getting deeper ad infinitum? So, just add another tower section every 10 years? etc.

    • D. Patterson
      Posted Feb 2, 2009 at 10:41 AM | Permalink

      Re: George M (#69),

      See the Website for the Automatic Weather Stations Project and Antarctic Meteorological Research Center. They have diagrams of a typical AWS, photographs of the equipment, maps, and lots more.

  56. Gerald Machnee
    Posted Feb 2, 2009 at 10:22 AM | Permalink

    I thought I saw Harry handing the ball to Reg. Or am I still dreaming football? Maybe it was Harry and Gill went up the glacier…

  57. George M
    Posted Feb 2, 2009 at 10:24 AM | Permalink

    OOPS, sorry, U of Wisconsin.

  58. Tim C
    Posted Feb 2, 2009 at 10:32 AM | Permalink

    Here are several things.

    1. a quick plot showing either the station intruments were moved or the database is corrupt or ?

    The altimeter shows umm… altitude, ok, barometer. From that it might be possible to deduce the location, will give a height and the pressure pattern at the location might be matchable against historic isographs.

    2. Mistakes happen.

    Seems the problem was spotted very quickly, some hours ago this appeared on the data web page

    “Note! The surface aws data are currently not available as they are being re-proccessed after an error was reported in the values at Harry station (2/2/09)”

    http://www.antarctica.ac.uk/met/READER/data.html

    • Ron Cram
      Posted Feb 2, 2009 at 10:53 AM | Permalink

      Re: Tim C (#72),

      That’s interesting! It’s nice to know the Met Office reads Climate Audit. Too bad they did not give Steve credit.

      • AnonyMoose
        Posted Feb 2, 2009 at 11:48 AM | Permalink

        Re: Ron Cram (#74),
        The British Antarctic Survey might not be reacting to CA directly. They probably got some email. I assume Steve has sent them some inquiries – I didn’t bother sending them any because we’re still discovering the available oddities so it would be an incomplete report/inquiry. And if Steve didn’t send any inquiries yet, it looks like there will be enough questions that he would send one soon. If they’re professionals one email is sufficient. Other parts of the BAS body can be tickled later if necessary.

    • David Jay
      Posted Feb 2, 2009 at 10:55 AM | Permalink

      Re: Tim C (#72), Well I guess they do read CA!

  59. Hank
    Posted Feb 2, 2009 at 11:03 AM | Permalink

    You know, these fellows have a really great idea that should save tons of money. All we have to do is put up weather stations at Seattle, San Diego, Houston, Miami, Charleston, and Boston and the models will tell us what the temps were in Phoenix, Chicago and Minneapolis.

  60. Posted Feb 2, 2009 at 11:07 AM | Permalink

    Verify and Validate >em>before release.

    That shouldn’t be too hard to carry to successful completion.

  61. Dishman
    Posted Feb 2, 2009 at 11:19 AM | Permalink

    “Note! The surface aws data are currently not available as they are being re-proccessed after an error was reported in the values at Harry station (2/2/09)”

    I appreciate and respect that professionalism.

    • Gerald Machnee
      Posted Feb 2, 2009 at 12:19 PM | Permalink

      Re: Dishman (#80),

      “Note! The surface aws data are currently not available as they are being re-proccessed after an error was reported in the values at Harry station (2/2/09)”

      Expect to see:
      “But it does not affect the calculations in the study.”
      Great work, Steve and crew!!

    • Tim C
      Posted Feb 2, 2009 at 12:32 PM | Permalink

      Re: Dishman (#80),

      I totally agree, thought, but didn’t quite know how to put it.

      Everyone makes mistakes, me included but when blame game appears and hiding starts it is a bad slope.

  62. Steve McIntyre
    Posted Feb 2, 2009 at 12:01 PM | Permalink

    I just noticed that they’d changed the data when I tried to re-run a script. I guess they’re monitoring CA. IT’s like last June when GISS data kept changing under our eyes as we were discussing. And needless to say, never providing credit to their source. I guess it’s Team policy. You can still get the screwed up series at GISS.

    I’ll finish off my post on Harry in a few minutes.

    However, I’ve got a copy of the data before they erased it … and will post it up later today.

    And the bad news for the Team is that while it’s easy for British Antarctic Survey to erase their data and replace it with a fixed version, it’s not so easy for the authors of Steig et al 2009, who were unfortunate enough to rely on it.

  63. Tim C
    Posted Feb 2, 2009 at 12:12 PM | Permalink

    Update: –

    The data has been changed. Now makes sense.

    (refresh any caches)

    I’ve eyeballed all the pressure datasets looking for any other obvious wobbles, saw nothing.

  64. Bob Koss
    Posted Feb 2, 2009 at 12:19 PM | Permalink

    Wow!
    GISS interpolating the annual temperature in 1991 with that missing D-J-F quarter while all the monthly values are the same sure made a difference between the stations. 0.93 degrees.

    I see the 1997-98 monthly values are also the same.

  65. Steve McIntyre
    Posted Feb 2, 2009 at 12:36 PM | Permalink

    #80. Well, how professional is it not to credit Climate Audit?

  66. Steve McIntyre
    Posted Feb 2, 2009 at 12:43 PM | Permalink

    Please continue discussion on new thread.

4 Trackbacks

  1. […] code.     “But it looks like there’s trouble with Harry… (Steve McIntyre) . West Antarctic Stations Climate Audit . ________________________________ 01 February, 2009 . “Sanity” Klaus […]

  2. […] network used.  Steve McIntyre (of Climate Audit) discovered that it was seriously defective, and posted about the error at 4:41 pm EST, on February 1 (Super Bowl Sunday). The British Arctic Survey (BAS) corrected the […]

  3. […] Yesterday, I noted that Steig had criticised previous developers of Antarctic gridded temperature data for not having “paid much attention” to West Antarctica (e.g. the NASA GISS trend map left the area blank due to lack of data meeting their quality standards) and reproached his predecessors (including, it seems, even Hansen) for “calculating with their heads”, rather than “doing the math”, which, in his case, seems to be inseparable from the use of Mannian algorithms. […]

  4. By Gavin’s Mystery Man « Climate Audit on Dec 11, 2010 at 7:24 AM

    […] Sunday, Feb 1 at 4:41 pm Eastern (3:41 blog time), I published a post describing West Antarctic stations. In that post, I observed that there was a very limited amount of station data (6 stations) in the […]

Follow

Get every new post delivered to your Inbox.

Join 3,423 other followers

%d bloggers like this: