4 More USHCN Stations

Anthony Watts writes in about 4 more USHCN stations:

I visited 4 USHCN stations in California last Friday, Davis, Napa, Santa Rosa, and Healdsburg. The sites were 4 for 4 with air conditioner units within the viewshed of the thermometer.

Of particular interest is Santa Rosa’s USHCN station, which is on the roof of the Press Democrat and surrounded by a sea of air conditioner units. Here is a picture:

Santa Rosa's USHCN station

More here and here

Then there’s Napa State Hospital, which has an air conditioner about 10 feet away along with close-by parking:
Napa State Hospital

More on Napa here


  1. jae
    Posted Jul 9, 2007 at 12:16 PM | Permalink

    Sheesh, all those air conditioner condensers right by thermometers. Is there no common sense anywhere, anymore?

  2. jae
    Posted Jul 9, 2007 at 12:18 PM | Permalink

    Looks like the Team has a new type of correction to make to the surface tempertaure data, the ACA (air conditioner adjustment).

  3. Dave Dardinger
    Posted Jul 9, 2007 at 12:25 PM | Permalink


    The interesting thing is that since air-conditioner usage should be some sort of proxy of high temperatures, the presence of air conditioners in the vicinity of official temperature recorders should be an amplifier of a pre-existing signal. Thus USHCN or GISS should claim ala Microsoft, that this is a feature, not a bug in their systems.

  4. BarryW
    Posted Jul 9, 2007 at 1:02 PM | Permalink

    I’ve been thinking about siting biases and how they affect the measurement. Consider the ACBE (airconditioner bias effect). This would not affect the measurements in a uniform way. If the site was in a temperate or cold cimate you would expect no effect in the winter months (unless there was considerable equipment in the building like a server farm) and you would expect to see positive bias in temperature in the mid summer months. The effect should be more pronounced in the maximum temp since that is when you would assume the ac is operating the hardest and little or no effect on the minimum summer temps, assuming it’s at site that is only operational during the day, since the ac would be running less or not at all after hours. So it would seem that the maximum highs for the year would be biased upward but the lows should not be affected, at least by ACBE. It would be interesting to monitor the ac usage vs temp at a site to see what actually happens with the ac on or off and compare that with a reference some distance away.

  5. MarkW
    Posted Jul 9, 2007 at 1:13 PM | Permalink

    Most businesses have to run their AC units all day, for much of the year. It’s not just server farms that put off a lot of heat.
    You squeeze a couple of hundred people, a couple of hundred PC’s, and all those lights (even flourescent) into a building, and that building is going to need to dump a lot of heat. Most modern buildings are very well insulated, so the only way to dump that heat, except when it is very, very, cold outside, is via the AC units.

  6. David Smith
    Posted Jul 9, 2007 at 1:38 PM | Permalink

    As mentioned, the MMTS sensors come with a cable which has to be run to an indoor readout/recording device. From that comes the temptation to locate these devices closer and closer to buildings. The old boxes had no cable run, so they could be located anywhere.

    One cooperative observer told me how ironic he found it that, after years of walking in bad weather to a remote box to take readings, the NWS installed the convenient indoor readout but, lo and behold, placed the new sensor very close to the building. If that new, close siting was OK, then why didn’t they do that years ago and save him all thosee trips to the distant box?

    I wonder if, when the MMTS sensors came along a decade or two ago, no one told the NWS installers about the importance of siting.

  7. Posted Jul 9, 2007 at 1:40 PM | Permalink

    The one here in Morrisville, Vermont is surrounded by blacktop and adjacent to (about 15 feet away) one of the busiest highways in the state. Good thing we have Bernie Sanders in Washington most of the time or there would be even more hot air around.

  8. L Nettles
    Posted Jul 9, 2007 at 2:16 PM | Permalink

    I found a cool one. I visited the Conway South Carolina station (381997) on July 4. It is not near a air conditioner. It is in fact in the shade of a large Live Oak tree in front of the fire station. As a southerner I believe this microclimate to be significantly cooler than an open meadow. My photos did not come out as I was using an unfamiliar camera. I will have to go back to get good pictures.

  9. bernie
    Posted Jul 9, 2007 at 2:53 PM | Permalink

    Based on Anthony et al’s work in this new branch of climate science, they have brought to light the previously unrecognized fact that the use of AC units globally since the Middle Ages is HS-shaped!! This HS relationship I have no trouble believing.

  10. Steve Sadlov
    Posted Jul 9, 2007 at 2:57 PM | Permalink

    RE: In many parts of the Sun Belt, A/C, even in homes, is used 8 or more months out of each year. The exceptions would be the immediate California coast (due to the natural A/C of the cold ocean) as well as upland areas. As MarkW noted, in commercial locations, it may be run even more than that.

  11. Murray Duffin
    Posted Jul 9, 2007 at 2:58 PM | Permalink

    There seems to be a discontinuity in the temp curve just before 1980. Is this when the site was moved here, or when the a/c was upgraded?


  12. BarryW
    Posted Jul 9, 2007 at 3:37 PM | Permalink

    RE #5 good points. The server farm comment was just because I was trying to think of a worse case. It really depends on the siting. One size does not fit all. The one that was in my thoughts was the Napa Hospital site with the window ac unit. Sites like Santa Rosa could be expected to run their ac 24/7, but a window unit may not be run all the time. When I get a chance I want to see if there is any noticable signature in the data due the the ac. It probably will be drown out by the other heat sources and noise though.

    #Re 11 the last move for Napa was on 9/22/97 according to the history file and 1/1/03 for Santa Rosa

  13. Roger Dueck
    Posted Jul 9, 2007 at 5:22 PM | Permalink

    I’m just wondering, why not have all of the sensors INDOORS, so they are out of the way of the harmful elements that might disturb the actual signal, and then adjust for the “known” differences to get the REALclimate data? After all, the teleconnection to the earths real climate would still exist.
    This also in response to Mann’s recent paper, closing with:

    We invite other researchers to download
    the source codes (and data) we have provided at http://
    http://www.meteo.psu.edu/mann/PseudoproxyJGR06, and to
    further explore CFR performance using either the CSM1.4
    simulation results, or other appropriately chosen climate
    model simulation results, using either surface temperatures
    or other fields (such as precipitation or sea level pressure),
    and using pseudoproxies such as those used in this study, or
    generalized to represent possible nonlocal teleconnected
    and/or nonlinear relationships between proxies and climate.

  14. novoburgo
    Posted Jul 9, 2007 at 6:26 PM | Permalink

    RE: #1 et.al. I would be more concerned with the effects of the extensive blacktop and the proximity of vehicles that can apparently pull right up to the sensor. You have to believe that winter temps will be greatly affected on sunny days when that asphalt is just sucking in heat only to be released at night. Warm winters = warm annual temperatures. Must have been sited by a bureaucrat.

  15. novoburgo
    Posted Jul 9, 2007 at 6:38 PM | Permalink

    RE NAPA View 20. If the AC is only 10 feet from the MMTS then the parking lot must begin within 5 ft. I can imagine a light 2-3kt northerly component wind transporting the shimmering heat right into the sensor!

  16. steven mosher
    Posted Jul 9, 2007 at 7:03 PM | Permalink

    We are measuring the efficiency of heat exchangers.

  17. Russ
    Posted Jul 9, 2007 at 7:57 PM | Permalink

    I have been helping Anthony with his survey. Curious I compared the temperature records of two sites with companion AC units, with two that lacked a companion AC. The AC sites showed warming. The non-AC sites did not have a discernible visual trend. You can see the details here. Is the “A” in AGW, really just short for AC. Should we be discussing ACGW?

  18. Jon
    Posted Jul 10, 2007 at 12:03 AM | Permalink

    MarkW writes:

    You squeeze a couple of hundred people, a couple of hundred PC’s, and all those lights (even flourescent) into a building, and that building is going to need to dump a lot of heat. Most modern buildings are very well insulated, so the only way to dump that heat, except when it is very, very, cold outside, is via the AC units

    Its amazing how true this statement is. Its sort of ironic actually. I once lived in fairly poorly insulated house (construction dating from the 1920s)–no insulation in the ceiling except one room, minimal insulation in the walls. In the summer this house was very comfortable without A/C. Sure it would be hot around 3pm inside and out, but it would cool-off rapidly. Convection currents tended to even make the middle of the day tolerable. Plus if it was really hot you could just work in the basement for a while. Neighbors with well-insulated modern homes would boil inside without the AC.

    Insulation is important for winter heating efficiency. It isn’t necessarily efficient.

  19. Anthony Watts
    Posted Jul 10, 2007 at 12:27 AM | Permalink

    Like Russ, I was curious about A/C units too, and I did some investigating for trend data and found some at the DOE.

    The numbers of A/C units is staggering, and on the rise, in part due to the availability of very inexpensive A/C units manufactured in China, Taiwan, and Korea, some of which can be had for about $100 US making it affordable for even the lowest income families.

    The rise of households using A/C over time looks vaguely familiar:

    Total US Households (Blue) versus Households with A/C units (Magenta)

  20. agn
    Posted Jul 10, 2007 at 3:37 AM | Permalink

    WRT MarkW’s comment in #5, you may be amused to read this (from The Daily WTF/Worse Than Failure, a techie site I sometimes visit): http://worsethanfailure.com/Articles/Im-Sure-You-Can-Deal.aspx

    It highlights the downside of over-conscious ecofriendliness…

  21. BarryW
    Posted Jul 10, 2007 at 8:47 AM | Permalink

    Re #19 So maybe we’re talking about AC Heat Islands. People used to go to the country or the beach in the summer to leave the sweltering cities if they could. Now they go for a week and use their ac units the rest of the time. In Washington, D.C. Congress used to disapear each summer now we can’t get rid of them at all since the invention of AC. So the UHIE correction should really be based on the population of AC units!

  22. Andrey Levin
    Posted Jul 10, 2007 at 9:09 AM | Permalink


    The best way is to have big thermal inertia mass inside well-insulated building to level-off day/night temperature differences, both in hot summer or cold winter.

    For example, in Israel standard floor/ceiling in highrises are feet thick concrete with floor surface from reconstituted stone tile. Works best in dry Jerusalem, where days are hot and nights are cold.

  23. Steve Sadlov
    Posted Jul 10, 2007 at 10:08 AM | Permalink

    RE: #19 – Certainly, during the latter half of the 1960s and most of the 1970s, A/C was no where near as well used as it was from the late 70s onward. As a child during that time, we lived in two places, a suburb of SF in the transition zone between coastal climates and inland hot areas, and, a place with warm to hot summers in the Midwest. In the SF burb, even though “normal” summers highs were typically above 80F, no A/C. Even in the Midwest, where summer would oscillate between the low 80s and up to 100, often with a good deal of humidity, we had a lone window unit for the dining room and kitchen. People have become real wimps about heat since those days. Now, most houses in the places I spent as a child have central A/C.

  24. Steve Sadlov
    Posted Jul 10, 2007 at 10:13 AM | Permalink

    I might add that the increasing obesity problem in the US must also take some of the blame – fat layers make one feel hotter and cuase one to exert more energy doing any and all tasks or simplly walking short distances at a normal pace. Ergo, the increasing “need” for A/C and increasing dependence on energy dissipative transpo and mobility systems.

  25. steven mosher
    Posted Jul 10, 2007 at 11:24 AM | Permalink

    Well, We all need to think about how AC exhaust will corrupt the record.

    which months? How much? will it corrupt TMAX or TMIN.

  26. steven mosher
    Posted Jul 10, 2007 at 11:34 AM | Permalink

    It is important to find good sites

  27. MarkW
    Posted Jul 10, 2007 at 11:38 AM | Permalink

    More importantly, will a slightly breezy day blow away the output of the AC units?

  28. Anthony Watts
    Posted Jul 10, 2007 at 11:53 AM | Permalink

    I have an idea, but I don’t have the time right now to pull it off. Maybe somebody else would like to do it?

    What is really needed is an idea of what a window mounted a/c unit heat plume looks like. This could be accomplished with an accurate digital thermometer and a notebook and measuring tape. Set up a grid around a window a/c unit and run it, get a shielded air temperature sample from several points on the grid, then graph the plume. Do it for calm days and windy daya.

    Obviously there are hundreds of a/c models and ratings and thousands of possible situations, but having a basic picture of just how the exhanust plume drops off with distance would be useful to know. Easy enough to do a simple experiment or two to get an idea. What I really want to know is how far away is the plume detectable given the known resolution of a mercury max/min thermometer and MMTS sensor.

  29. Steve Sadlov
    Posted Jul 10, 2007 at 12:41 PM | Permalink

    RE: #28 – Or it could be modelled …. 🙂 …… seriously, modelling it would be a good thing to do. I wonder if there is a software similar to Flotherm (or maybe even Flotherm itself could be used) to model it. Then, of course, field measurements to confirm.

  30. crosspatch
    Posted Jul 10, 2007 at 12:55 PM | Permalink

    Considering the change in energy requirements of our data center at my place of employment, I might be interested in seeing a plot from that temperature sensor against a plot of the electricity consumption of that building over time.

    Newer computers (particularly servers) have been increased in electricity requirements as they have become faster and while we have fewer computers where I work, we are at the limit WRT power available in the data center. And judging from the chatter in various professional groups of which I am a member, that isn’t an uncommon problem. And it is a double whammy because more power consumption by the computers requires more power consumption from the air handlers designed to rid the building of the generated heat.

  31. Sam Urbinto
    Posted Jul 10, 2007 at 12:56 PM | Permalink

    The grids should be comprised only of accurate data, that is, from stations of known accuracy. Every dubious measurement should be removed and the data reassesed. I don’t have a good feeling about trusting the temp numbers any more. I used to at least somewhat, but now seeing the stations and thinking about the margin of error from then to now, I’m feeling it’s the measurement environment and/or the quality of the measurements. I mean, it’s never exceeded 1 degree C (up or down) since 1880 and has basicially hung out in the +/-.5 range the entire time!

    Yes, Yes, the trend is all, I know…. But if you plot out the temp variations on the same scale as the CO2 rise and compare them, it seems we should be seeing far higher temperatures based upon the huge huge huge upswing in CO2 at Mauna Loa compared to it. Yeah, I know, it’s the aersols etc, keeping it cooler, but that co2 a big big curve compared to a blip, and that co2 curve and trend since the 1940’s!

    But doesn’t it seem more likely it’s the readings and the advent of digital thermometers and then them getting less expensive and more widely used? Even if it’s not the A/C units and tarmac/concrete/vehicles etc. That would tend to explain the trends we’re seeing in the temps pretty well and we wouldn’t need to make models or wait for volcanoes to explain it, and wouldn’t have to adjust the data to make it fit.

  32. Douglas Hoyt
    Posted Jul 10, 2007 at 1:18 PM | Permalink

    Anthony (re #28) – I think you can rent a thermal imaging camera and look at these plumes. Also the local utility company might also have some of these cameras which they would use to look for energy leaks in buildings. Maybe you could get them to take some photos for you. Or they might already have photos.

  33. John G
    Posted Jul 10, 2007 at 3:05 PM | Permalink

    The developers / manufacturers of the AC units might have done some thermal flow analysis and measurements on the units to determine appropriate clearances for proper heat dissipation. If they did this analysis, though, it probably would take a miracle to get them to release it.

  34. MarkW
    Posted Jul 10, 2007 at 3:18 PM | Permalink


    I really don’t want to hear any more talk about models. Unless her name is Christy, Tyra, …

  35. steven mosher
    Posted Jul 10, 2007 at 4:17 PM | Permalink

    re 29.

    This issue keeps coming up. I’ll take a look at flow therm. Last time you mentioned it
    I took a quick peek. My sense was that it could handle half of the problem.

    My sense is that if you take a pristine site, you can pollute it but at some point
    you hit a threshold.
    ( logistic curve response)

    It’s also clear that the impairment might hit TMIN more than TMAX.

  36. Sean Houlihane
    Posted Jul 11, 2007 at 3:11 PM | Permalink

    I’ve not been following the temperature saga too closely, but my impression is that the key parameter which is used is the average of two temperature readings, Max and Min. I was observing that in the UK, temperatures have been low recently, but the hadobs charts only indicate a small difference compared with the baseline. Is the charted daily average value representative of surface thermal energy for that day, and how do max and min relate to this?
    There is a set of high frequency temperature readings covering a 10 year period at Cambridge Computer Lab which might provide interesting for analysis. Days with higher max also seem to have lower min temperatures, the average doesn’t seem to particularly correlate with max or min values – so a cloudy day appears to have a similar average temperature to a sunny one with a clear night sky. Clearly some proxies will also respond to max temperature rather than mean temperature as well – just wondering if the parameter which is being measured is really related to what the models are trying to predict…

    Note that this site makes no claim to be ‘ideal’, it is on the top of a building, has moved, etc. The intention is to provide real-time data rather than a historical record.

  37. Geoff Sherrington
    Posted Jul 11, 2007 at 8:41 PM | Permalink

    Re # 36

    Ever since I started writing on this Board, I have been saying that max min temps or temps at midday or 6 am or whatever are a poor proxy for daily heat flow. Have tried to get some original data where temp is measured at short intervals, no luck. We need to look at various correlations between integrated temps and point temps. Good for you Sean if you can do it. Love to see the results. Then would look at Parker’s data on wind effects and UHI with an integrated daily record and see if Puff the Magic Dragon disappears. Geoff.

  38. BarryW
    Posted Jul 12, 2007 at 8:58 AM | Permalink

    ASOS data has 1 and 5 min interval data

    Try http://www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwdi~ASOSPhotos and click on the station id

  39. BarryW
    Posted Jul 12, 2007 at 9:07 AM | Permalink

    Sorry, wrong reference.

    http://www.ncdc.noaa.gov/oa/climate/climatedata.html#surface has a link to the ftp site for 1 and 5 min ASOS data

  40. D. Patterson
    Posted Jul 12, 2007 at 2:07 PM | Permalink

    Re: #10 While living in Los Angeles County at the western end of the San Gabriel Valley nearby the Santa Anita racetrack during the decade of the 1970s, we typically used the air conditioner part time or full time only for a few months from about June through September. December temperatures often threatened possible frosts which could damage the citrus trees in the backyard. It was ususally necessary to wear a jacket or coat from December through March. This changed rather dramatically after about 1979. By around 1982 the winters became too warm to wear a coat unless you drove into the adjacent mountains and snow. Consequently, it reached a point where the coats stayed in the closet unused the year around. Use of the air conditioner increased from 2 to 3 months per year to 11 to 12 months per year. Instead of experiencing temperatures of 35F on December 10th, the air conditioner was running full blast to counteract the effects of 90+F temperatures. Garden plants which required a winter chill to properly germinate failed to grow.

    Likewise in Southern Illinois and Puget Sound in Washington there have been notable changes as the warmer temperatures have encouraged the use of air conditioning in older homes which had not needed them very much before to remain cool. Now, however, there does seem to be another change taking place in Southern California, Southern Illinois, and Western Washington. The latest winters have been more harsh and longer lasting than has been seen since the decades of the Sixties and Seventies. It now appears as if the cycle is poised to reverse and return to the colder weather patterns I experienced 30 to fifty years ago. If this results in a Dalton Minima or Maunder Minima in Solar Cycles 24 or 25, there is going to be a generation of people who have no experience to prepare them for the consequences of such continuing cold weather.

  41. Steve Sadlov
    Posted Jul 12, 2007 at 6:23 PM | Permalink

    RE: #40 – In the SF Bay Area, the 70s had a wider swing in annual temps but rather dry winters for the most part. The El Nino of ’78 coincided with the flip to positive PDO and upended it all. From then and throughout most of the 80s, there was notably wetter and slightly warmer winter weather while summers were relatively warm. From ’87 to ’91, the drought which had begun in SoCal around ’84 spread to the north and affected us as well, but not badly. When that ended, we got into a very El Nino dominated set of years – even neutral and La Nina years were warm and wet, this lasted until late ’97. Something new started in the mid 90s, namely, crummy and weak Springs, where late winter like weather would persist until it was nearly the Solstice. By ’98, the signs of the demise of the positve PDO were there, and we got into what appeared to be ringing at the end of the cycle. Crummy springs continued and even summer heat became more sporadic (but these sproadic heat events became seemingly more intense, in an almost early Fall-like, Santa Ana-esque way). Now I’m calling it pure play negative PDO, all the signs are there, what with the very dry winter we had along with record cold and yet another crummy spring. Summer’s a dud thus far, for the most part as well, in contrast to the SoCal where it’s tended to be hotter than normal inland.

  42. beng
    Posted Jul 12, 2007 at 8:06 PM | Permalink

    FYI, I can feel my central A/C exhaust at least 30 ft away. It’s still in a surprisingly narrow plume, but by then well-mixed w/adjacent air. Might try a test next time it’s needed — don’t right now w/nights in the low 60s.

    Highest summer temp here in the Appalachians (900 ft elev) has been 93F (34.5C) & only a couple days of 90+ so far. Unless the surrounding forest drastically dries, these are typically the highest temps that can occur (hot air mass w/sun all day & little wind) due to evaporative cooling, outside of urban areas. Severe forest drying greatly increases the potential highest temps, like the 1930’s & 1960’s droughts.

  43. beng
    Posted Jul 12, 2007 at 8:10 PM | Permalink

    Whoops, I rechecked & there was a 94F day – so correct to my #43.

  44. Posted Jul 16, 2007 at 1:50 AM | Permalink

    When I was young, I worked for the US Forest Service in California… My first year of work was at the Region 5 Forest Fire Laboratory in Riverside.. There we were taught how to take weather measurements, because your crew was used in major project fires to record weather information..

    The next fire season I set up an un-official station at the heli-tack base where I worked… Two fire seasons later, I set up an official station that is still being used today after more than 30 years(Oak Opening, CA) .. They (the USFS) has a manual for setting up weather stations.. It gives the height of the station enclosure, distance from other taller trees and other stuff that might interfere.. The USFS is not the only fed gov’tagency with manuals… Hey, they all have manuals, all written by the same people.

    Violating their manuals is common, apparently from reading this website. No one cares once it is set up, and a AC condensor installed 10 feet from the station means nothing? Holy cow, this is insane..

    My suggestion is to figure out the agency, and ask to see their manual for installing and maintaining weather stations..


  45. Earle Williams
    Posted Jul 16, 2007 at 8:07 AM | Permalink

    Re #44

    Christophger Hill,

    This is apropos to nothing, but I just want to say that I hope to see more of your postings if only for your signoff.

    I’m fully intending to go document the Milwaukee, WI, station at Mount Mary College. I lack a GPS but the station is where it appears on the Google Map; it already has accurate coordinates. Nicely shaded garden spot, surrounded by a parking lot. No AC units in the vicinity, just plenty of cars.

  46. MarkW
    Posted Jul 16, 2007 at 8:27 AM | Permalink

    That there has been a change is beyond question. That the change is due to warmer summers is not.
    Could it not be possible that the fact that airconditioning units, have gotten cheaper and more efficient, making them less expensive to own and operature, have something to do with the fact that air conditioning units have gotten more plentiful?
    Could the fact that air conditioning is now considered by most to be a neccesity, rather than a luxury, also be playing a role?

    As to fewer coats in the winter. A couple of possibilities.
    People are more obese than they were a generation ago. Fat is a good insulator.
    It takes time for people to become acclimated to changes in climate. The fact that people no longer expose themselves to the extremes of summer heat, could make it due to the omnipresence of AC units, makes it possible for them to adapt to winter temperatures more quickly, thus they don’t “feel” as cold, and hence don’t feel the need to encumber themselves with coats?

  47. Steve Sadlov
    Posted Jul 16, 2007 at 10:22 AM | Permalink

    RE: #46 – RE: Air conditioners. Also, consider the obesity factor here as well. Furthermore, whereas, 50 years ago, not having A/C meant sweat, today, in many settings, especially businesses and some homes, it may mean greatly exceeding the operating temperature specs of electronic equipment sometimes leading to incendiary results!

    RE: Coats. Another thing those of us in the US have no doubt witnessed are the strange aspects of the now-much-emulated gang banger look whereby one overdresses in the summer and underdresses in the winter. Go figure …. 😉

  48. Christopher Hill
    Posted Jul 20, 2007 at 2:31 AM | Permalink

    Dear Earle Williams:

    You will not see my sign-off.. and your statement that my post is apropos to nothing makes absolutely no sense whateeover…[snip]

    My point that is obvious to all except you, is that you may challenge station location, elevation of measuring station, proximity to parking lots or air conditioners, by simply asking to read that agency’s manual on setting them up… but alas, that will pass you by I know.. Oops, did you hear that noise? That was the real world passing by again…

    Christopher Hill

%d bloggers like this: