Tucson Update

You may recall the post earlier this year where the USHCN official climate station with the largest positive trend in the USA turned out to be located in a parking lot at the University of Arizona in Tucson. See below:

Click picture to see image gallery at surfacestations.org

The GISS surface temperature plot for Tucson:

The Uof A Tucson station is still there in the parking lot, but as www.surfacestations.org volunteer Bob Thompson finds out while visiting family this Thanksgiving, it appears they are in the process of dismantling the station. The wooden portion of the Stevenson Screen has been removed.

Click picture for full sized image

It seems odd for them to remove the shelter, yet leave the precision aspirated thermometer, which is the long tube with the “T” on the left side spanning the old shelter mount.


  1. steve mosher
    Posted Nov 23, 2007 at 6:52 PM | Permalink

    Anthony, I wonder if they had measurements from both systems?

  2. Kenneth Fritsch
    Posted Nov 23, 2007 at 7:25 PM | Permalink

    In the plot of the Tucson station temperature trend I see similarities to the plots that resulted when I compared the CRN123 to the CRN45 station trends using USCHN Urban data sets for the time period 1920-2005. What I see is large trend from 1950-1980 with lesser or no trends in the time periods prior to and after this time period. After considering Steve Mosher’s suggestion to look at change points in these plots, I am becoming more and more convinced that such an analysis would be most applicable here and may provide some insights into the timing of the quality changes that are currently being measured for these stations.

    Ideally this should be performed with the advice of a qualified statistician. In my limited knowledge and reading on change point analysis I have found several methods that are available, but not all the articles I have read caution against some misapplications that can occur from not removing or considering the presence and effects of auto-correlations. I would consider it a shame not to submit all these data and quality measurements to some well informed statistical analysis. As a layperson, with some effort, I could attempt something of this nature but I would not have the full confidence that I did not overlook something and I would understand my work not being taken seriously.

    Do any of you posters here have any thoughts on such an analysis? In a former life I was often able to persuade people I worked with or supervised to get involved in these types of analyses by throwing out my theories in such a manner as to encourage them to prove me wrong. I have no confidence that that ploy would ever work here or that with my advanced years I could pull it off.

  3. Anthony Watts
    Posted Nov 23, 2007 at 7:53 PM | Permalink

    Mosh, the surveyor of the station for the original picture, Warren Meyer, indicated that the min-max mercury thermometers had been removed from the shelter. I’m not sure how he determined that, since the gate for the enclosure is locked.

  4. Posted Nov 23, 2007 at 8:25 PM | Permalink

    I took the original pictures and did the original survey. At that time, there was nothing in the Stephenson screen. This may be as simple as removing it because it is no longer necessary and might interfere with measurement by blocking wind or sun. But I would think those effects would be small compared to the sea of asphalt around the station ;=)

  5. steve mosher
    Posted Nov 24, 2007 at 9:26 AM | Permalink

    re 2. AGreed. Somebody with some experience needs to step up.

  6. Fred
    Posted Nov 24, 2007 at 11:21 AM | Permalink

    Wonder what the Team would be saying if all these stations currently awash in seas of black pavement were instead positioned on large pads painted white, surrounded by acres of green lawns that were auto sprinkled with cool water every few hours.

    Or if all those air conditioners were turned around so they blew cool air instead of hot towards the instrumentation area??

    Not trying to pick on the Teamsters, just curious.

  7. Kenneth Fritsch
    Posted Nov 24, 2007 at 12:03 PM | Permalink

    Re: #6

    Wonder what the Team would be saying if all these stations currently awash in seas of black pavement were instead positioned on large pads painted white, surrounded by acres of green lawns that were auto sprinkled with cool water every few hours.

    What you say is not facetious and has been discussed by Christy, as I recall, as a danger in making adjustments to temperatures. If one feels that temperatures are trending up sharply one tends to look more closely and diligently for the need for those adjustments that adjust temperature up (legitimate as far as it goes) but neglect the adjustments that go against their inclinations..

  8. BarryW
    Posted Nov 24, 2007 at 1:09 PM | Permalink

    Re 7

    Works the other way too. Stations that were moved to airports were assumed to be “cooler” than in city locations and corrected upward.

    [snip- please lay off the editorials. Everyone. ]

  9. Alan Bates
    Posted Nov 24, 2007 at 4:22 PM | Permalink

    The temperature plot (Tucson5.ipg) does not display for me …

  10. Posted Nov 24, 2007 at 8:37 PM | Permalink

    Anthony, what happened circa 1955 (transitory -2 dC change) and c. 1998 (permanent
    -2 dC change) to cause the big blips in temperature?

    In any event, according to MMS, Tucson switched to MMTS as its primary
    thermometer on 7/9/86, with min/max as backup only. But then
    on 6/25/96 a “sixes” thermometer was also introduced. What is this?

    — Hu

  11. Anthony Watts
    Posted Nov 24, 2007 at 11:52 PM | Permalink

    RE10 Hu,

    A “sixes” thermometer is part of an evapotranspiration pan. Are you sure you are looking at the right station? That sounds like the UofA Ag station.

    The way they label these Tucson stations is easy to confuse. This is station 028815

  12. BradH
    Posted Nov 25, 2007 at 8:35 AM | Permalink

    Given the unquestionable familiarity with temperature measurement requirements and the dedication to accurate science required of each and every faculty at this esteemed institution, I presume that they have temporarily suspended providing temperature data to the network, given the station’s present, patently unserviceable status?

  13. Alan Bates
    Posted Nov 25, 2007 at 10:41 AM | Permalink

    Re #9

    Thanks for fixing it!

  14. Al
    Posted Nov 25, 2007 at 5:20 PM | Permalink

    If we’ve learned one thing from the Surface Stations push, it is that there’s an insane number of potentially influencing factors.

    The recommendation that seems key to a truly functional Climate Network would be a minimum two-year overlap during any station change. At least that way all of the static factors can be reasonably estimated.

  15. Paolo
    Posted Nov 26, 2007 at 2:07 AM | Permalink

    I’ve been looking at rural surface station data in the US hunting for the decadal warming trend. I chose rural stations for the initial analyses, figuring other stations might be biased by virtue of being located in densely populated areas. I reckon this story nails it. I am amazed anyone would collect and report temperature data from a hot parking lot!

  16. David Smith
    Posted May 10, 2008 at 12:06 PM | Permalink

    Here’s an experiment I’m trying. The site is shown below:

    I have a temperature recorder at point A and point B and will be adding a third at C. These devices record to the nearest degree F, so sensitivity is limited, but I think they’re OK for gauging magnitude. They record every 30 minutes.

    “A” is in a baseball field at least 200 feet from anything other than mown grass. It approximates a CRN 1 or CRN 2, I believe.

    “B” is at the edge of a 200 ft by 350 ft weathered asphalt parking lot. To the north is grass for at least 200 feet. To the south is the parking lot. The prevailing wind is southeasterly, from the parking lot across the sensor.

    “C” is at the adge of the field, near trees. It is similar (in terms of tree encroachment) to several USHCN sites I surveyed.

    I’ll run this for about a month and then report the temperature profiles.

    I’ll also do other combinations. I’m especially interested in the behavior of temperature near trees, as I suspect trees do affect the temperature profile and that the effect can be hard to detect as trees slowly encroach on a sensor. I particularly wonder whether there’s a peak effect near the south dripline of a tree, where a sensor would get full insolation but face limitations on IR radiation.

  17. Kenneth Fritsch
    Posted May 10, 2008 at 2:51 PM | Permalink

    Re: #16

    David, will you be recording wind, cloud, and other pertinent weather conditions that could effect the temperature differences at these locations?

  18. David Smith
    Posted May 10, 2008 at 3:28 PM | Permalink

    Ken, yes indeed. There is also a high-quality private station about two miles south which keeps records and which will be the backup.

    I also plan to take IR photos of the area. I’m not sure what, if anything, those will reveal but it can’t hurt.

  19. Sam Urbinto
    Posted May 10, 2008 at 8:20 PM | Permalink

    This is great! Fits in well with the idea that getting min/max to get mean to be “indicative” of the area is rather dependent upon where the sensor is more than anything else.

    Since you’ll have the 30 minute data, are you planning on comparing min/max to get the mean to using a weighted amount of time at each level to get the mean?

  20. David Smith
    Posted May 11, 2008 at 6:37 AM | Permalink

    Re #19 Sam, yes I do. Also, I’ll post the raw data in an Excel spreadsheet for anyone who wants to slice-and-dice it.

  21. steven mosher
    Posted May 11, 2008 at 6:50 AM | Permalink

    re 20. check “geiger climate near the ground” pgs 263 ff. a whole chapter on
    trees and forests and clearings etc

%d bloggers like this: