How not to measure temperature, part 19

In traveling around California and Nevada to look at NOAA USHCN climate
monitoring weather stations I’ve seen some odd things. I’ve seen temperature
sensors near asphalt and concrete, sensors placed within feet of buildings and
cars, sensors placed near air conditioner exhausts, and sensors that had
barbeque grills in the vicinity.

Last Friday June 6th, I traveled to Santa Rosa, CA to the Press Democrat
Newspaper, a wholly owned subsidiary of the New York Times, which according to
NOAA, has the climate station of record for Santa Rosa.

SantaRosa_Press_Democrat_Building.JPG

I figured it would be a fairly straightforward survey, and that I’d probably
find the temperature sensor near the back of the building by the parking lot, as
I’ve found many others placed. I figured it would look a lot like our local
newspaper, the Chico Enterprise Record (owned by Media News Group), weather station. It is in the rear, in a bit of
disrepair, missing some slats and part of its roof assembly, but otherwise ok.
It is seen below:

ER-wxstation.JPG

But nothing prepared me for what I was about to find at the Santa Rosa Press
Democrat.

When I arrived, I couldn’t locate the NOAA MMTS sensor anywhere around the building, but I
did see a tower on the roof of the building, and in the rear of the building
they had a Davis Vantage Pro2 weather station on a pole. I knew that wasn’t
the official climate temperature sensor provided by NOAA. So, after doing a
perimeter search twice, I went inside to inquire within. Everything in the lobby said
"go away". I guess it was the bullet proof glass, and the cameras, and the
security guard. After getting a name of the person resonsible for their weather
page from the front desk, I called on my cell
phone, no answer.

Undeterred, I decided to try looking outside again. It was then I noticed the
5 level parking garage about a block north.

From the top of the parking garage a quick scan with my binoculars located the NOAA MMTS
temperature sensor. It was there, about 8-10 feet above the roof, surrounded by
a sea of air conditioners and exhaust vents!

Here is what I saw from my binoculars:

SantaRosa_Press_Democrat_MMTS.JPG

And here is a panorama of the entire rooftop of the Press Democrat. Click on
the image to get a full sized view with panning functions:

SantaRosa_PressDemocrat_rooftop_panorama.JPG

A complete photo essay is available on my surfacestations.org website.

An independent check of aerial photos that I had access to confirmed the
placement of the sensor in the middle of several air conditioning units.

 SantaRosa_Press_Democrat_aerialview.JPG

 I’ve seen some poorly thought out places to measure temperature, but
this one takes the cake. Not only do we have the sensor above a sea of air
conditioners with warm air exhausts, there are two rooftop building exhausts,
plus the roof and building itself, and then lets not forget that the Press
Democrat itself is in a sea of buildings in downtown Santa Rosa, all of which to
contribute to the Urban Heat Island (UHI) the thermometer is exposed to.

Yes folks, this is an official USHCN Climate station of record. The data from
this station goes into the national climatic database. But given the absurd and
irresponsible placement of this NOAA MMTS thermometer, is it any wonder at all
that the graph of temperature at Santa Rosa looks like it does?

SantaRosa_GISS_station_plot.gif

This isn’t a case of gradual encroachment by localized site changes that happened around the thermometer, like what happened in Marysville. This is a deliberate placement of an official thermometer in the worst possible measurement scenario. Somebody had to choose this location, the building and air conditioners did not grow up around it.

About these ads

27 thoughts on “How not to measure temperature, part 19

  1. Leaving aside the effect on the temperature record for a moment,how do the locals get an accurate temperature report?

    No wonder the poor old weather man gets such a bad name!

  2. If this trend in measurement unreliability continues, one may well start to wonder if the current long term global warming cycle we are supposedly in isn’t just a lot of hot air.

  3. How does the fact that the paper is a “wholly owned subsidiary of the New York Times” play into this observation? Other than the fact that the New York Times is considered the voice of the far left, I am struggling to find a rationale for including that information in this posting. Was there one? And why not mention that the ER is owned by Media News Group? It seems as if political bias is creeping into the “science”.

  4. Sean: If you’ll look at the flags, the blue one, you’ll see that it says “New York Times Company”.

    You can read it in the panorama #2 visible here at the highest resolution.

    http://gallery.surfacestations.org/main.php?g2_itemId=11269&g2_imageViewsIndex=4

    Since they were flying the flag of the NYT right next to the USHCN climate station, it seemed worthy of note. Otherwise I probably would never have noticed. The ER, is in fact owned by Media News Group, but they don’t have such flags or banners next to their weather stations but you are right, to be balanced I should have said that, so I’ve added the wording to the post.

    The point is, organizations, government or private, big and small, no matter who owns them aren’t paying close enough attention to how temperature is measured and keeping the measurement environment as bias free as possible.

    Given the importance of the measurement, at least a modicum of due diligence is expected. Pride in reporting news by flying a flag proclaiming such pride should also apply to pride in doing such an important job, measuring temperature for a climatic database that serves the public trust, correctly and bias free.

    I don’t know if they are aware of just how bad a job they are doing, as I’ve seen many other places public and private that are equally oblivious to the measurement environment under their care.

  5. 2 questions:
    Where should the measuring devices be placed for Santa Rosa? Is the NYT getting paid to be the host for the devices?

  6. Regarding Sean’s gripe: “How does the fact that the paper is a “wholly owned subsidiary of the New York Times” play into this observation?” and “It seems as if political bias is creeping into the “science”.

    How does mentioning who owns the ER or the Press Democrat change the fact there is a complete disregard for surface station standards at the Press Democrat site and a reasonable compliance at the ER?

    Sean’s complaint of “political bias” is misplaced, smacks of obfuscation, and is wholly beside the point. I suspect that the possible political bias of a site owner has little correlation to how a site is set up, but it might make for an interesting study.

    Maybe Sean should start his own blog and weather station survey which adheres to his high standards of political correctness, looks for political correlations to station implementation, and leaves the question of scientific correctness to others.

    In any case, Anthony has taken Sean’s gripe to heart and has amended his blog to mention the corporate owners (or renters) of both sites.

  7. David:

    If anything is beside the point it’s that the building belongs to a paper that is owned by the New York Times.

    The tone of Anthony’s posts is that science is being done with the surfacestations.org work. I don’t like to see politics enter into the sphere of science and my question was an indirect way of asking that we avoid mixing the two.

    Now, with regard to the science that is being done I have just one question: where is it? Anthony states that the sensor is in the “worst possible” collection placement, and yet I’d hypothesize that placing the sensor directly on top of one of those AC vents would be worse. Of course with only a picture and no temperature readings we have no way of knowing whether the site is in fact the best or the worst.

    Further, Anthony states that a better site could have been selected, and then does not offer any evidence that this is the case. That site has more than 100 years of data associated with it, so moving it 2 miles away is not possible. So, rather than just state that a better site is available why not tell us exactly where that would be and provide the numerical analysis to back it up. Otherwise it’s just an opinion and that’s worthless in a scientific debate.

    Of course you’re right about one thing: this is Anthony’s personal blog and he is certainly entitled to post whatever he likes. Perhaps I’ll do just as you advise and create my own. I’ll keep you posted.

  8. I’m interested in knowing how we should interpret the graphs that oftentimes accompany photos of the reporting stations. Quite often we see the generally upward sloping pattern of temperature readings. Let’s say that the compromised weather station has been in place on the roof, to pick one example, for 10-15 years. If so, might not the annual effects of A/C, roof material, etc. be constant or at least relatively constant? I wouldn’t think that the use of A/C, for example, would vary very much year to year unless there was a temperature/humidity increase that would cause the A/C units to run more and more over time. This seems a little far fetched, however, unless there is some other information to support or refute the patterns that are reported. Hope someone could shed some light on this situation. Thanks.

  9. Sean wrote: “Now, with regard to the science that is being done I have just one question: where is it?”

    It is in process, once a more even national distribution of surveyed USHCN sites exists, then a correlation and ranking based on the NOAA CRN site rating criteria will begin and stay current through completion of all 1221 sites.

    Right now we are in data gathering mode and will stay that way for quite awhile.

    We’ll be using this rating scheme, described by Michel Leroy (1998). It is being used by Meteo-France to classify their network of approximately 550 stations. The classification ranges from 1 to 5 for each measured parameter. The errors for the different classes are estimated values provided by NOAA.

    Class 1 – Flat and horizontal ground surrounded by a clear surface with a slope below 1/3 (3 degrees.

    Class 2 – Same as Class 1 with the following differences. Surrounding Vegetation 5deg.

    Class 3 (error 1C) – Same as Class 2, except no artificial heating sources within 10 meters.

    Class 4 (error >= 2C) – Artificial heating sources = 5C) – Temperature sensor located next to/above an artificial heating source, such a building, roof top, parking lot, or concrete surface.”

    One of the goals of the http://www.surfacestations.org project, is to apply this site classification standard set forth by NOAA for the USCRN to the existing network of surface stations so that the problems associated with some sites can be quantified, and the good sites can be clearly defined as well. This will help scientists whom use the surface temperature data do more accurate analyses.

  10. John Rieman asked about “where should the sensor be?” That answer can be found in the classification system posted above. NOAA used this rating system already to define placement of the newly created “Climate Reference Network” (CRN) which can be seen here:

    http://www.ncdc.noaa.gov/oa/climate/uscrn/

    The CRN is not yet complete, and it won’t solve the bias problems that have crept in to the historical near surface temperature data.

    As for NYT being “paid” as far as I know, the answer is no. The NWS Cooperative Observer Program relies solely on volunteers, just like my project to document the stations does.

  11. It has been a few years since I visited Santa Rosa, but I vaguely recalled noticing some greenery in the vicinity of downtown. So, I pulled up one of the satellite images, and I see a number of small park and recreational and simply unused green areas within 6 – 8 blocks of the newspaper. Any of which, it seems to me, would be a more appropriate location for temperature sensors. Moving 2 miles would be unnecessary, although if you review some of the other station histories, that seems to not be a problem. Encinal, TX for one example I recall.

    Now, about the mention of the NYT. Forget politics. I would take that to mean the SR paper has deep pockets, and access to top scientists who could have given proper advice on (re-)location of that station. And they could afford to move it, and maintain it properly.

    As far as stability of the error due to surroundings, a new roof is usually required about every 15 to 20 years while the color (reflectivity) changes continually, air conditioners are serviced annually, and often the &%*# covers are replaced incorrectly. Also, they degrade in performance (efficiency), requiring replacement about as often as a roof, and so on. Nothing stable there.

  12. good article and as always interesting. I live 2 blocks away from the PD building and I think the building has been there about 30 to 35 years….if that info is of any use to anyone

  13. How does that all correlate with local develpment over the same time? True, the building and exhaust vents didn’t “grow around” the heat sensor as you say, but I wouldn’t be surprised if local development made some contribution?

  14. “Right now we are in data gathering mode and will stay that way for quite awhile.” — Mike Young

    If I understand Anthony’s comment about this, then that could be restated as, “Right now they are in the biased-baseline setting mode, and we’ll stay that for a long while to be sure to lock the errors in place for the forseable future.”

    And that bias will never be erased as long as it is in the database. Eventually it may be mitigated but that could take a long time, depending on how fast they normalize data collection to as close to neutral input as possible. But to do that, they should probably do pilot-studies to determine the best way to do that and which, as sloppy is this all seems, have doubtless not been done.

    The real problem is that allegedly “scientific” conclusioins and policy initiatives based on them are RIGHT NOW being hatched based on faulty data: data which anyone with an ounce of common sense can see are faulty.

    This is NOT science, and pretending it’s o.k., just because we are “only collecting data” (the most critical aspect of the scientific process) makes no sense whatever, especially given the potentially devastating economic consequenses these errors could have.

  15. Response to Sean Greer

    Re: “If anything is beside the point it’s that the building belongs to a paper that is owned by the New York Times.”

    Perhaps I haven’t been clear. Unless some privacy issue is at stake why not mention who is responsible for a site? Sean Greer insists that to so mention represents a form of political bias creeping into the science.

    Let me be absolutely “to the point” — the notion that mentioning ownership of a particular site is evidence of political bias is absurd. It is complete, utter nonsense and has all the the earmarks of political desperation grasping at straws for the purpose of diversion and calumny.

    Where is the science? Good question! Where is the adherence to NOAA standards any reasonable researcher should expect?

    Mr. Greer’s notion of what is worthless and what is not in the scientific debate is likewise nonsense and does not merit a response other than to note it is nonsense.

    With regard to what makes a suitable site —

    Why should Sean Greer, or anyone for that matter, need Anthony Watts to hold their hand and tell them what comprises good station implementation? Anyone can read the standards set by NOAA for themselves.

    Nevertheless Anthony has already done as much but some, evidently, do not choose to listen. I guess political theater is more important to some folks.

    With regard the specifics on the effects of site characteristics, my understanding is evaluations are in the works. Nevertheless, NOAA set standards, they had a reason to set standards, and the standards are not being enforced nor are stations even inspected for compliance.

    Perhaps Mr. Greer should ask NOAA how they decided what were to be the standards, why such standards should be established, and why they even bothered to make a set of standards since they do not, evidently, expect them to be followed.

  16. “A station’s potential for heat island bias over time should be low.” — from the ncdc ushcn website.

    The key phrase there is “should be.”

    It doesn’t look like they are doing that. It also doesn’t look like they are in start-up and debugging mode, or they have any intention of tossing out any compromised data (or even have any intention of determining to what extent there are problems).

    But I haven’t been looking at this stuff for long, so if someone can direct me to where they EXPLICITLY state any of that AND how they are applying it, I would appreciate it.

    So far it just doesn’t look like they are adhering strictly to their protocol.

    And, if it were up to me, I would have 2 sets of stations. Those in heat islands, and those remote from them: the better to quantitate the effect. But then I’m just picky that way.

  17. Late4Lunch,

    I didn’t say that we are in the “data gathering stage.” Anthony did. I just posed a few questions about the stability or instability of the environment in which the stations exist compared to the historical record.

  18. Shouldn’t the temperature increase be somewhat relative? Yes, this is obviously a terrible place to measure temperature but it seems like the temp readings would have been equally bad 25 years ago. Has the Santa Rosa area grown enough to cause this much change in the urban heat island effect? Say the temps in the 70’s were reading 7F too high in 1975 because of being on a rooftop, shouldn’t it be reading the same 7F too high today and graph a straight line over the past 30 years? Why this increase if the thermometer hasn’t moved? Just curious, and keep up the good work.

    -Matt

  19. I am a scientist (physicist) and I wrote my first “report” about global warming when I was in the 4th grade. I have been involved in this study all my life and I agree with the owner of this site that the temperature data around the world is of compromised value. The example posted here is typical of what is found world wide. Fact is if you take temperature sensors that were moved decades ago out of heat islands and still remian unaffected by urban sprawl, the data from those sensors does not show dramatic temperature increases. Most often the poor placement of data sensors is not intentional, but back in the early 1900’s scientists and weathermen recognized heat island effects and moved almost all sensors to remote airports. Most early airports were dirt runways or had minimal hardscape (paved runways) and most then were concrete not asphault. Over time most of those locations due to urban expansion have now become part of the heat island once again. Thus the data is not an accurate representation of the actual temperatures but far more a reflection of urbanization across the US and the world.

    Sean Greer’s post criticizing these conclusions seems far more a reflection of this info not fitting his personal view of global warming. It is not necessary in scientific debate to recommend alternatives, a simple statement of fact that points out a flaw in data is all that is necessary as long as that statement was based on widely accepted information. Heat islands are a fact, Air Conditioning is a heat transfer system and rejects heat to the outside. These are facts with no known dispute. A good example is that I would not have to point out alternatives to the cold fusion failure to prove it does not provide output data as published.

  20. Pingback: Jennifer Marohasy » Beware Thermometer Temperature Data

  21. Great spotting Anthony. I wonder too who is responsible for placing the NOAA MMTS
    temperature sensor in such a ridiculous location. Hmmm….someone’s obviously not at all concerned by the daily exaggerated daytime temperature spikes.

Comments are closed.