How not to measure temperature – part 30

Russ Steele is out on vacation and doing several surveys while traveling. This one below is from St. George, UT. Here we see an MMTS measuring the temperature near the surface of an elevated parking lot. The effect of the asphalt and vehicles that park near it, engine forward, probably dwarfs the effect of the nearby a/c unit. The shading may help daytime temps some, but the asphalt likely biases Tmin the most. The complete photo survey is available on surfacestations.org

St George_south.JPG

St George_east.JPG

Advertisements

13 thoughts on “How not to measure temperature – part 30

  1. I suppose Gavin Schmidt would say this site doesn’t matter, or that the data can be accurately adjusted. The real question is how such a station could be added in the first place.

    Our tax dollars at work, collecting garbage.

  2. Ok.

    you have so many frigged up things here. Its a mini lesson in UHI

    1. Shading from insolation.
    2. Canyon effects ( multipath) when you do have insolation.
    3. Sky view impairment.
    4. evapotranspiration impairment.
    5. Wind shelter.
    6. Artifical heating.

    When the sun goes down that concrete slab and stucco walls start releasing the Joules…

    The most important thing.
    Upon VISUAL INSPECTION you could not guess whether these impairments were a net positive or net negative..

    2 Approaches.

    1. The Gavin approach. We assume that all impairments are net net neutral, UNTIL you prove otherwise. Garbage accepted.

    2. Mosh pit. Ahhh this site is Good FOR NOTHING. junk its data and the horse it rode in on.

  3. Later today, we’ll be breaking into groups to consider what key recommendations to put forward from what we’ve learned from his conference.

    Knowing what I know about the state of USHCN, suggesting a culling of stations that don’t meet climate monitoring standards and then a renewed analytical focus on those that do, will be my recommendation.

    I don’t know if Gavin was originally talking about 60 stations worldwide or 60 in the USA, but I think we can find 60 in the USA if that was the case. Not sure about ROW…had lunch with a guy studying weather stations in Tibet…no metadata at all. And Mcintyre has found only 10 in Africa with a data record that might qualify them.

    Mosh do you recall whether Gavin was speaking of USA or ROW for 60 stations??

  4. Now, this is just weird:

    I accessed the trend for this station through:
    http://www.co2science.org/scripts/CO2ScienceB2C/data/ushcn/ushcn.jsp

    It’s fairly flat until about 1979, then becomes a yo-yo (I’m guessing that the parking structure was built in the late 70’s).

    Even more weird, the trend for the last 50 years is DOWN.

    *** NOTE FROM ANTHONY Steve, carefule with Co2 Science’s graphs, I’ve seen cases where they aren’t fully representative on trends becuase some of the pre1900 data was measured without shelters, giving higher than normal annual means then. That can cause the modern slope to appear negative.

  5. Weird.

    Looked at the USHCN trend for that site. Pretty flat until about 1978, then turns into a yo-yo. Hard to believe any place has swings like that.

    Surprisingly, the trend is down!

  6. Andy,

    Are there any sites in the No. Va area or Metro DC area that you need someone to take pictures of?

    I live and work in the No. Va area.

  7. “the data can be accurately adjusted”: in my experience, undergraduates are franker. They refer to “fudging” the data.

  8. I just looked at that Saint George trend. From what little I know of the area, the huge annual temperature swings don’t really surprise me much. The whole area is a deep hole, entirely enclosed by mountains. (The irony of the sensor itself being in a similar hole is amusing) It’s temperature would depend much more on wind patterns, and inversion layers, than on traditional heating and cooling methods. Most intermountain west stations have the same problem, and show similar swings. This leads me to two questions:

    Is it reasonable to use such an obvious microclimate in creating a world average?

    Is it reasonable to use ANY data set with such huge variations, in creating a world average?

    To determine a trend, a higher annual variance requires a much longer measurement period. Clearly the period for Saint George is nowhere near long enough to coax a trend out of it. A single year, and not even an unusual one, can entirely reverse the 127 year trend. But because it’s isolated, it’s fairly rural, and it’s changed fairly little in the last 100 years, (compared to other cities) it’s probably given MORE significance than most other sites. And it’s probably used to adjust the trends of Las Vegas, Salt Lake, Phoenix, etc, because big city measurements aren’t very stable.

    Even if the Saint George data is taken as just a single point in the 1200 or so stations, it’s importance is surprisingly big. The SEVENTEEN DEGREE DROP from 1999-2000 was enough to change the entire country’s average by .013 degrees. I’d be willing to bet that the US average is entirely dominated by stations between the Front Range, and the Sierra Nevada. The typical 3-4 degree variations in Poughkeepsie, or Chapel Hill are comparatively insignificant. But those, at least, are measurements of temperature, instead of wind patterns over mountains.

  9. After looking through some of the records on:

    http://www.co2science.org/scripts/CO2ScienceB2C/data/ushcn/ushcn.jsp

    (thank you, Anthony) I’m noticing a rather disturbing pattern. At least for the western states, it seems like the 60’s and 70’s were unusually stable compared to other decades. In and of itself, that doesn’t make much difference. But most of the “temperature anomaly” graphs I’ve seen were zeroed on the average from 1960-1990, plus or minus a year or two. Has anyone done any statistical analyses on that period? If the 60’s and 70’s really are anomalous, then using them as the baseline poses some serious problems.

  10. Anthony,
    Thanks for the warning, and sorry for the duplicate post.
    Hope the conference continues to go well.

Comments are closed.