Lights = 0, Air Conditioners = 22

There’s been some recent discussion about how only rural stations have been used in the GISS analysis, and those rural stations are qualified by looking at night time DOD satellite photos, and doing a count of visible streetlights within a radius to quantify UHI potential or lack thereof. The “best” stations are labeled “lights=0”

One of those stations is Happy Camp, California, population 2182, an old gold mining and logging town located in the rugged NW corner of the state, and about 100+ miles from any major city. MMS reports data back to 1931 with 3 small distance station moves, and no changes to equipment. GISS reports data back to 1914.

It looked like a good candidate to look at for a lights=0 survey. The weather station is located at the Ranger Station:
Happy Camp Ranger Station - USHCN climate station of record

But what you can get from satellite images and databases can’t really prepare you for what you may find. I “expected” to find an old classic Stevenson Screen, probably near the Ranger Station office. Check on that. But what I didn’t expect to find was a “rural” station swimming in a sea of exhaust from 22 air conditioning units within 100 feet of the Stevenson Screen. Ridiculous, you are making this up you say? Well that would be my first reaction too.

But here they are, count them, I’ve labeled the A/C units for your convenience:

Happy Camp Ranger Station looking West from Stevenson Screen
Happy Camp Ranger Station looking West from Stevenson Screen

Happy Camp Ranger Station, looking NE, Stevenson Screen visible
Happy Camp Ranger Station, looking NE, Stevenson Screen visible

Happy Camp Ranger Station, looking North towards courtyard
Happy Camp Ranger Station, looking North towards courtyard

Happy Camp Ranger Station, looking southwest inside courtyard
Happy Camp Ranger Station, looking southwest inside courtyard

To help you get bearings on this walking tour, an aerial photo is available here

And the complete collection of photos is available on

In addition to the 22 A/C units within 100 feet there are other biases too. Granted, not all 22 may be introducing a bias, but since Hansen counts lights near stations, to asess UHI magnitude, we can count A/C’s. Additionally. for other biases, positive and negative there’s the buildings, the shade trees, the wind sheltering, and the lawn sprinkler. There’s also the big parking lot to the southwest, and the Stevenson Screen is at the top of a slope and there’s a parking lot downslope.

When I mentioned to the site curator about the A/C units she said “hmm, I never thought about that” but then added, “But I can tell you that when we water the lawn, my high temps are lower”. I asked the curator what the prevailing wind direction was, and she said “south to southwest usually”.

Now there doesn’t appear to be much of a trend according to the NASA GISS plot, but there are some large amplitude swings and discontinuities:

Unedited NASA GISS raw data plot for Happy Camp RS
Unedited NASA GISS raw data plot for Happy Camp RS

So one has to wonder, with all the observed microsite biases, what is the data really showing? One also wonders what the plot might look like if this station was better sited.

And if a lights = 0 station like this one, far removed from urbanization, has so many such micro-site biases, could others have similar problems? It looks like more hands-on site surveys will have to be done to determine the true value of lights=0 USHCN sites.


  1. Russ
    Posted Aug 1, 2007 at 6:43 PM | Permalink


    How bad can it get? Just when you think you have seen the worst site, there is still one more just over the horizon. Good science is full of surprises.

  2. Steve McIntyre
    Posted Aug 1, 2007 at 6:47 PM | Permalink

    How many feet is the parking lot from the screen?

  3. Anthony Watts
    Posted Aug 1, 2007 at 6:49 PM | Permalink

    about 15 feet, see this picture:

  4. Anthony Watts
    Posted Aug 1, 2007 at 6:51 PM | Permalink

    RE3 BTW the dent in the screen front leg is from a riding lawnmower accident, according to the curator.

  5. Steve McIntyre
    Posted Aug 1, 2007 at 6:53 PM | Permalink

    For review, here are station siting policies:

    The EPA recommends the sensor be no closer than four times an obstruction’s height, at least 30 m from large paved areas, and located in an open level area that’s at least 9 m in diameter. The open areas should be covered by short grass, or where grass does not grow, the natural earth.

    Avoid these:
    ‘€¢ large industrial heat sources
    ‘€¢ rooftops
    ‘€¢ steep slopes
    ‘€¢ sheltered hollows
    ‘€¢ high vegetation
    ‘€¢ shaded areas
    ‘€¢ swamps
    ‘€¢ areas where snow drifts occur
    ‘€¢ low places holding standing water after rains

    This one has multiple problems – too close to pavement, shade, obstructions, air conditioners. Not all of them necessarily go in the same direction.

    While the “voluntary observer” excuse is regularly trotted out, this is done institutionally. I’m sure that the Forest SErvice wouldn’t have had any objection to a proper siting. So the problem is abysmal NOAA administration of the network on Karl’s watch.

  6. Posted Aug 1, 2007 at 7:07 PM | Permalink

    Note that this station isn’t really being used anymore… The GISTEMP data stops in 1995 for annual averages, and 2002 after a few years of very sporadic monthly data.

  7. TCO
    Posted Aug 1, 2007 at 7:22 PM | Permalink

    I think the burden of proof is on you to show that the throw of these ac’s is in the 100 foot range. It boggles my mind as an HVAC engineer that 100 foot away window units have an impact. In any case, burden of proof is on you. And the reporting and hoo-haaing about this kind of in process work is just silly.

  8. Stan Palmer
    Posted Aug 1, 2007 at 7:37 PM | Permalink

    re 7

    I think the burden of proof is on you to show that the throw of these ac’s is in the 100 foot range

    Check the second picture, the air conditioning unit is wwithin one trcuk lenght of the weather stattion. This is far less than 100 feet. Your credibility is strained and your cuase not helped when you make statements that are so easily refuted.

  9. BarryW
    Posted Aug 1, 2007 at 7:38 PM | Permalink

    Any idean when the ac units were installed? It looks like a temperature upswing started in the late 50’s early 60’s.

  10. mjrod
    Posted Aug 1, 2007 at 7:47 PM | Permalink

    RE: 7,

    The burden is not on Steve or Anthony to prove anything. The “burden” is on the NOAA to ensure their equipment meets the standards THEY (NOAA) set up.

  11. JJC
    Posted Aug 1, 2007 at 7:53 PM | Permalink

    “It boggles my mind as an HVAC engineer that 100 foot away window units have an impact.”

    An HVAC “engineer”? Is that anything like the culinary technician that took my order at Domino’s? Plugging numbers into a Manual J calculator is not engineering. You don’t have to be a DeVry graduate to know putting a temperature sensor near an AC exhaust is a bad idea.

  12. TCO
    Posted Aug 1, 2007 at 7:56 PM | Permalink

    5: I don’t see air conditioners listed in your list. Are you saying that this is a LARGE industrial heat source? And what is near? Is it 100 feet? 1000? 100000000?

  13. steven mosher
    Posted Aug 1, 2007 at 8:01 PM | Permalink

    RE 7.

    Actually the burden of proof is on any researcher who has used USHCN sites to PROVE that they are well
    sited before utilizing them in analysis.

    It goes like Like this.

    Researcher: I used the USHCN network. According to Karl it is a high quality network and the signals
    we get from it reflect the climate.

    Reviewer: did you look at the sites?

    Researcher: well, no; I’m Lazy. Prove my assumption wrong. I’m lazy squared!

  14. hoi polloi
    Posted Aug 1, 2007 at 8:06 PM | Permalink

    Can an HVAC engineer present some data to show at what distance between the sensor and the A/C boggling of the mind occurs?

  15. steven mosher
    Posted Aug 1, 2007 at 8:10 PM | Permalink

    TCO HVAC engineer

    Please read this for tommorrows test:

    Station metadata plays a critical role in the accurate assessment of climate data and eventually of climatic change, climate variability, and climate prediction. However, current procedures of metadata collection are insufficient for these purposes. This paper introduces the GeoProfile as a model for documenting and visualizing enhanced spatial metadata. In addition to traditional metadata archiving, GeoProfiles integrate meso-scale topography, slope, aspect, and land-use data from the vicinity of climate observing stations ( We describe how GeoProfiles are created using Geographical Information Systems (GIS) and demonstrate how they may be used to help identify measurement bias in climate observations due to undesired instrument exposures and the subsequent forcings of micro- and meso-environments. A study involving 12 COOP and US Historical Climate Network (USHCN) stations finds that undesirable instrument exposures associated with both anthropogenic and natural influences resulted in biased measurement of temperature. Differences in average monthly maximum and minimum temperatures between proximate stations are as large as 1.6 and 3.8 °C, respectively. In addition, it is found that the difference in average extreme monthly minimum temperatures can be as high as 3.6 °C between nearby stations, largely owing to the differences in instrument exposures. Likewise, the difference in monthly extreme maximum temperatures between neighboring stations are as large as 2.4 °C. This investigation finds similar differences in the diurnal temperature range (DTR). GeoProfiles helped us to identify meso-scale forcing, e.g. instruments on a south-facing slope and topography, in addition to forcing of micro-scale setting. Copyright ⧠2006 Royal Meteorological Society.

  16. SteveSadlov
    Posted Aug 1, 2007 at 8:25 PM | Permalink

    I am not sure what is meant by lights=0. Certainly, Happy Camp is not Eureka or Redding. But with 2K people, there are not zero lights (or zero “urban” effects). Happy Camp is a typical strip development along a major highway connecting the coast with the I-5 corridor. So, there is densification albeit strung out along the main drag. It’s got to be hotter than the surrounding wild lands. It’s got to be.

  17. Anthony Watts
    Posted Aug 1, 2007 at 8:35 PM | Permalink

    I don’t mind answering questions about the site.

    Prove 100 feet? Easy, just look at the 100′ scale (lower left) on the aerial photo of the Ranger Station, which you can see here:

    Apply it as a radius from the Stevenson screen location.

    22 A/C’s of about 1500 to 2500 BTU each (common for window units, lets call it 2000 BTU) x 22 = 44000 BTU of potential waste heat vented into the area. That’s an industrial range. Given that the Stevenson Screen is in a wind sheltered area, and the courtyard area is wind sheltered except for one small south opening pointing directly at the North side of the screen, the potential exists for harboring waste heat near the screen.

    Additionally, since the curator says that prevailing winds are “south to southwest”. When that happens, laws of fluid flow say that heat is transported from the parking lots and from A/C unit venting, flowing along the southwest wall of the building and towards the screen. In the case of an east wind, there’s up-slope from the lower parking lot (see picture). For a west to north wind, the screen is sheltered, though some of the courtyard A/C will likely be transported towards the screen.

    And regarding posting this data online: I’d rather be accused of being too transparent than be accused of withholding data as we’ve seen some people do. I’ve said from the start that the site surveys will be publicly posted as a work in progress. If you don’t want to look, as they say in TV, “change the channel”.

    The most important point of all this is: We wouldn’t be having this conversation if NOAA and Karl et al had done their jobs and monitored/repaired these sites at the measurement/siting level. Trying to disentangle trustworthy data from the many variable +/- bias noises should not have to be the job of the researcher.

  18. TCO
    Posted Aug 1, 2007 at 8:42 PM | Permalink

    Antony, so what is the delta T impact from one AC 100 feet away? (I’m not trying to say that it is small with this rhetorical question, I’m traying to say that you DON’T know it.) You don’t know the mixing rate. The throw. You have some BTU numbers there to impress the polloi. But the key issue is mixing (and in the out doors!)

  19. steven mosher
    Posted Aug 1, 2007 at 8:51 PM | Permalink

    I’m not so concerned about the AC units. If they were installed by our local
    HVAC expert I’m sure they are non functional

    The wind shelter bothers me more. This from a study of the CRN. The CRN, where they actually
    take PHOTOS of sites and pay attention to siting details and assume the burden of proving the siting
    does not reflect non climatic signals. Just a tidbit.

    mean absolute difference averaged from all seven sets
    is 1.00C for Tmin, about 2.5 times larger than for Tmax.
    The mean absolute difference averaged from paired
    stations within 10 km is 0.71C for Tmin, about 4 times
    larger than for Tmax. The weaker effect of siting on Tmax
    is the consequence of daytime heating and the
    enhanced mixing of the lower atmospheric boundary
    layer. However, the effect of siting is stronger for Tmin
    since it is highly sensitive to surrounding environment
    and local terrain. For example, the USCRN site at the
    Arboretum near Asheville, NC is located in a shallow
    topographic bowl and is surrounded by trees 25 to 50
    meters away. These siting characteristics allow for low
    wind speeds, reduced mixing of the boundary layer, and
    the drainage of cooler air towards the site. Contrasted to
    this is the USCRN site at the Horticultural Center only
    10 km to the South-Southeast. It is open to the wind and
    sits on a small hill that allows any locally cooled air to
    drain away from the site. Thus the mean difference of
    Tmin is 1.11C cooler at the Arboretum. Paired stations
    at Stillwater OK are both located at fairly open sites,
    however, one station is located on a slight hillside while
    the terrain surrounding the other station is quite flat. The
    cold air drainage effect at the first site thus brings a Tmin
    difference of 0.85C between these two sites. A
    dramatic effect of cold air drainage is better seen at
    Gilmore Creek, AK, where the COOP station is only 0.8
    km away from the USCRN station but lower in elevation
    by 260 meters. This different exposure leads to a Tmin
    difference of 3.74C annually.

    Not that things like the wind breaks within 25 meters have any effect.

    The other day It was hot. I saw one of the hoi polloi under a tree in the shade.
    Silly immigrant. Didn’t he know he needed an argument with full citation to understand
    that it was cooler in the shade and that unless he had such an argument he was not
    entitled to recognize what was obvious.

  20. Anthony Watts
    Posted Aug 1, 2007 at 9:02 PM | Permalink

    RE27 well the point is, We’re not out to measure mixing rates. We’re measuring site compliance. This site is clearly out of compliance by published and known NOAA standards. Thats what this project is about, to get a census of compliance. Those that keep trying to tell me that I need to do a bunch of other things at the same time should just either mind their own beeswax, or go do those studies themselves. Its a volunteer effort, done on my own time and systems and the generous time of others so funds for such suggested research diversions aren’t readily available.

    We have one goal, completing a USHCN site compliance census.

    You’ll have a hard time convincing anyone that A/C units near temperature sensors is irrelevant just because the localized mixing rate is unknown. Just one A/C nearby, much less 22, brings the issue forward. The magnitude of such +/- micro-site biases is for the future.

    Other studies will follow I’m sure. Perhaps you can do one from the perspective of an HVAC engineer.

  21. Jim Edwards
    Posted Aug 1, 2007 at 9:13 PM | Permalink


    I generally agree w/ you that the effects of HVAC on these network sites is, at times, perhaps overstated [I’ve posted to that effect…] That’s different from saying a site can’t be criticized for diverging from published suggested site guidelines without first proving the site is actually heat contaminated by HVAC. You’re pushing it.

    These sites are meant to be representative of the continental US, not the developed continental US. All of these sites are fine with me, as long as a proportionately larger number of sites are installed in the vast reaches of under-developed land in this country.

    I’d agree that we can’t predict whether a site is biased up or down w/o a pretty complex understanding of the site. I’d also agree that probably fewer than a half dozen of those A/C units have any significant chance of affecting measured temp, so mentioning 22 is overkill. Sample frequency and A/C duty cycle will be key combination affecting precision of measured temp. Look at the noise in the data.

    Anthony states the wind direction was reported as S to SW. I assume this means from the N to the S, which would tend to direct some of the discharge from the condenser of unit #13, above, directly to the Stevenson screen. If it means from the SW toward the NE, it would tend to do the same thing for A/C unit # 8. It looks like about a half-ton unit, and maybe it only runs 20% of the time, but that would still generate about ~1500 btu / hr off the condenser, on average. Even if the air ‘is well mixed’, as you say in your #13, above, that potential heat contribution can’t be good if this site is being sold as representative of undeveloped portions of the US.

    Personally, I might be more concerned with the proximity of the screen to single-glazed, aluminum-framed windows – where the building interior is likely heated during the winter and during cold nights. This would tend to elevate measured temps during the winter and at night.

    Proximity to asphalt looks like a near non-issue to me – even if it violates guidelines – b/c the asphalt appears to be well-shaded by evergreens year-round.

    Watering can’t be an improving factor on data quality if watering frequency doesn’t match rainfall patterns of the local cell this site is suppossed to be representing.

    I share your legalistic reading that window A/C units are not ‘large, industrial heat sources.’ That doesn’t mean the recommendations shouldn’t warn against measuring temp near split system or packaged condensing units. I submit to you that these EPA recommendations may have been written at a time when A/C was less prevalent.

    A rule from the law of negligence is that compliance with customary practice is not proof that one’s conduct was reasonable, it is merely evidence that one’s actions weren’t negligent. Just because scientists are placing network stations near A/C units that aren’t specifically warned against doesn’t mean they’re not negligent.

  22. JS Polloi
    Posted Aug 1, 2007 at 9:38 PM | Permalink

    I looked around and saw many “lights = zero” sites that had microsite problems.

  23. krghou
    Posted Aug 1, 2007 at 9:41 PM | Permalink

    At Home Depot window units ranged from 6000 to 12000 BTU/hr. I picked 8000 BTU/hr as it seemed to be most common. Converting to watts gives a little over 2000 each, and a little over 50KW for 22 of them. Taking a radius of 100ft around the screen, which includes all the A/C units, yields an area of about 3000 m^2. 50 KW / 3000m^2 is 16.7 W/m^2.

    I guess 4W/m^2 leads to global catastrophe, and 16W/m^2 is insignificant and not measurable.

    Obviously I could just pick the close one as well, say one unit within a circle of 15 ft radius. 2000W / 4.4 m^2 = 454 W / m^2. Obviously insignificant as well.

    The numbers for the distances are just guesses from reading the thread. Please feel free to correct if they aren’t correct.

  24. TCO
    Posted Aug 1, 2007 at 9:43 PM | Permalink

    Is there a correlation of “problems” with heatup rate over the last several years? What is the value of the slope and the R2?

  25. Jim Edwards
    Posted Aug 1, 2007 at 10:11 PM | Permalink

    #34, TCO:

    I agree with you there. I would not want to guess whether / how much temp is affected. I made this same general point on July 27 [7-22-07 Thread: “Turbo bias”]. I can easily imagine the effect to be non-insignificant, however, even with plenty of mixing w/ ambient air. Air discharged from a condenser will have a +20-30 degree F dT, as you know.

    If I quote from my own comment # 41, [7-15-07 Thread: “Pain in Maine, but they can measure rain”]

    If a mere 1% of the air entering the air sensor has been previously heated 20-30 F above ambient, that’s a net ~0.25 F increase to measured ambient air temp.

    That seems to me to be a reasonably likely description of the addition of significant positive noise. I won’t say that it will happen, but it seems likely enough to expend the effort required to avoid the possibility of contamination when the risk is foreseeable. Think of it as a precautionary principle.

    I think it’s pretty clear to anybody who’s worked in HVAC that condenser fans are not going to be blowing unmixed discharge air directly onto temp probes. If they did, the RC crowd would be reporting ~10-15 degrees F of recent warming, rather than ~0.5 F. I agree with you [if you’re saying that…] that some of the discussion seems to imply that hot discharge air is blowing into Stevenson screens, which is: 1) wrong, and 2) implies gross recklessness on the part of climate scientists, rather than simple negligence on the part of desk-bound eggheads.

  26. Bill F
    Posted Aug 1, 2007 at 10:21 PM | Permalink

    The “gross recklessness” is not the unpaid volunteer weather observer who puts a station in a less than ideal location. The “gross recklessness” is when a climate scientist claims that the station is rural and represents a homogenous station location with little or no contamination from development when he or she has done next to nothing to actually determine if that statment is correct or not. Guys like Karl and Hansen write papers claiming such conditions for their “rural” stations and editors and reviewers let it pass without asking for the first stitch of proof that they have done anything to actually verify that it is true. Guys like Anthony don’t have the burden of proving that the station data is not affected by the various developments around it. The guy who made the unquestioned claim in a peer reviewed article has the burden to prove that the statement in his article was not false when confronted with overwhelming evidence that his rural and uncontaminated sites may be anything but rural and uncontaminated.

  27. Dave Dardinger
    Posted Aug 1, 2007 at 10:23 PM | Permalink

    re: #33

    I suppose it was just a trap to see if TCO was paying attention [wink], but you have a math error in your calculations.

    a circle of 15 ft radius. 2000W / 4.4 m^2 = 454 W /m^2

    You forgot to square the 4.4m (15ft)and then you forgot to multiply by pi. the result should be 33 W /m^2. Still large but not as gigantic as 454. The first calculation is also missing pi.

  28. Jim Edwards
    Posted Aug 1, 2007 at 10:31 PM | Permalink

    Dave Dardinger:

    And don’t forget the heat is contaminating a volume, not an area [and hot air has an inconvenient tendency of rising, which eliminates most, if not all, of the effect of 2nd floor units], so I expect the positive heat bias will be lower than implied in these calculations.

  29. VirgilM
    Posted Aug 1, 2007 at 10:31 PM | Permalink

    This doesn’t appear to reflect very well on the 150 NWS employees, who struggle to find volunteer observers to take observations of temperature, rainfall, snowfall, and snow depth once a day, 365 days a year. Considering that it is the local NWS offices that find new observers and maintain the equipment and also considering that the local NWS office is focused more on precipitation than on temperature (drought and flooding), I am not surprised that a number of sites are not located very well for temperature. Was anyone who was managing the observation network 100 years ago (hell even 15 years ago) thinking that this data would be used to determine decadal temperature trends important for climate change research? Can we site a station today in a rural area and know for sure that that area will remain rural in the year 2100? The city I live in today only had about 1000 people 100 years ago. Now it is at 100,000. Were the Weather Bureau employees thinking 70 years ago: “If I put the station here, will they build a mall parking lot across this gravel road?” I think we are asking the NWS to do an impossible job with limited resources. Ideally, they should install automated equipment in the middle of nowhere on top of vegitation that is representative of a large area around it. That costs money and it is something that Congress is unwilling to fund. I wouldn’t make any conclusion based on the USHCN network in terms of temperature trends. I wasn’t designed from the outset to detect temperature changes of less than 2 degrees.

  30. VirgilM
    Posted Aug 1, 2007 at 10:47 PM | Permalink

    I have to agree with TCO that the air conditioning units likely don’t have much impact on the temperature. In my condensor unit, the fan blows the hot air upward and since it is warmer than the surrounding air, it will continue upward and become a very, very, very small part of the UHI. I have to get within 5 feet of my unit to feel noticible warmth. The closeness to a building is much more problematic. The snow melts very fast south of my house within 20 feet because of light reflected from the side of my house to the ground. My house is also emitting IR radiation.

  31. Dave Dardinger
    Posted Aug 1, 2007 at 10:53 PM | Permalink

    re: #28 Jim,

    Sure, but the increased back radiation from the atmosphere comes precisely from absorption of IR from the surface and is converted into column temperature increase. It doesn’t stay on the surface or the earth would soon be covered with a layer of lava. So the fact that the 33 watts of heat energy is directly mixed into the air column merely eliminates the middle man, so to speak.

  32. TCO
    Posted Aug 1, 2007 at 10:57 PM | Permalink

    Jim is talking about the air conditioners, Dardie.

  33. Jim Edwards
    Posted Aug 1, 2007 at 11:16 PM | Permalink

    #32, 33:

    Yes, solar energy is like a planar wavefront with ~1400 W flux – if you move up and down in altitude, you’re still going to get hit w/ around the same flux. An A/C condenser is more of a point source, and effects would be similar to inversed square relationship if the heat weren’t escaping up, and since the discharge air mixing w/ ambient isn’t likely to be isotropic. Result isn’t so easy to predict, I predict.

    Two important factors missing from calculations, above, are also duty cycle [as TCO notes] and efficiency. A half-ton unit removes 6000 btu / hr from the conditioned space, but due to inefficiency will emit maybe 30% more heat than that into the outside environment [~ 7800 btu / hr]. This assumes the unit runs 100% of the time, however, which very few do – except on the hottest days. A more reasonable duty cycle for a place like this might be more like 20-30 %, which is why I came up with a value of ~ 1500 btu / hr of heat emitted in the outside environment for a 6000 btuh unit in my first post, above.

  34. paul graham
    Posted Aug 2, 2007 at 12:10 AM | Permalink

    Give the lack of quality control I feel any studies that utilizes this date should contain the post fix.
    * may contain nuts!
    ! Non Uniform Temperature Statistics

  35. JS
    Posted Aug 2, 2007 at 4:53 AM | Permalink


    Don’t forget about when the wind comes across the top of the building the a/c unit is attached to and creates an eddy which takes the heat from that a/c unit (not to mention the heat from the roof of the building) right into the temperature station.

  36. MarkW
    Posted Aug 2, 2007 at 5:02 AM | Permalink


    No, it is up to you to show that these violations of the standard have no affect. You are the one who wants to use the data, it is up to you to prove that the data is valid.

  37. MarkW
    Posted Aug 2, 2007 at 5:07 AM | Permalink

    As we know from the Barrow, Alaska study, even small towns can have significant UHI. And Barrow doesn’t have anywhere near as much asphault as this place does.

  38. MarkW
    Posted Aug 2, 2007 at 5:15 AM | Permalink

    It’s true that hot air should rise. In still air. But air is rarely still. Wind, of as little as a mile or two per hour is enough to cause turbulence that can keep hot air closer to the ground. Especially if the wind is coming from behind the building.

  39. TCO
    Posted Aug 2, 2007 at 6:04 AM | Permalink

    Wind also gives you mixing. In any case, stop taking it to the bank that 100 foot away ACs are a problem. My initial guess (and I’d bet on it) are that they aren’t and that people here have an overblown view of their effect. Antony disagrees. Well fine. I admit to not knowing. But my point is that it’s NOT a no-brainer that Antony is right. I’m not making a sophist point here. This is honestly my betting Bayesian view. Before taking this stuff to the bank give us an estimate of impact. We’ve had the spectacle of people quoting BTUs to impress us with the aura of some sort of thermodynamic problem-solving. But it’s never drilled down to a delta T. This is because the mixing problem is NOT trivial and becayse Antony and the like, lack any feel (even a plug and chug use ASHRAE tables for large plenums) for this problem.

  40. Andrey Levin
    Posted Aug 2, 2007 at 6:08 AM | Permalink

    Weather stations use day/night max/min temperature metrics. Couple of minutes of wind carrying AC hot exhaust directly to temperature measuring unit could corrupt day max temperature data. No need to heat it all day long to get biased reading.

  41. TCO
    Posted Aug 2, 2007 at 6:22 AM | Permalink

    Mark: If you want to publish a paper that says that the authors have not done micro-site calibrations on the GISS sources, fine (pretty weak contribution btw). What I am criticizing is the huffing and puffing about the air conditioners as if it is proven that they are a problem.

  42. Michael Jankowski
    Posted Aug 2, 2007 at 6:29 AM | Permalink

    Re#30, but what change in temp does it take for you to notice the “warmth?” And I have to be within about 5 feet of a vet in my house to notice the warmth of blowing air when the heat is on, or the coolness of the air coming from the A/C, but surely those warming and/or cooling effects extend far beyond 5 feet (or else my house would never warm or cool due to my HVAC system)?

  43. TCO
    Posted Aug 2, 2007 at 6:29 AM | Permalink

    41. That’s an interesting observation. Still need to show the impact in some mathematical form. But your observation is shrewd and additive.

  44. MrPete
    Posted Aug 2, 2007 at 6:53 AM | Permalink

    TCO, you are truly missing the point. No need for mathematics to prove a simple yes/no compliance assertion. Either a site is in compliance or it is not. Scientists should only be using data from compliant sites.

    It is a painful and ultimately wasteful exercise to analyse the exact degree of impact for various out-of-compliance factors. Sure, analyse the general impact of various compliance requirements…that’s how we get good compliance rules. But in general, the rules are set up to provide good confidence that nothing is affecting the data.

    Out of compliance sites do not give good confidence about the data.

    Let’s find the known-good sites, and only use their data. It’s that simple.

  45. Steve McIntyre
    Posted Aug 2, 2007 at 7:13 AM | Permalink

    #45.I agree with Pete.

    It’s quite possible that the impact of non-compliant things is negligible, that good sites will give similar results. While air conditioners themselves are a potential defect, their presence tends to be associated with other non-compliance issues e.g. paved areas within 100 feet, nearby obstructions. Surely the appropriate way to proceed is, as PEte says, to try to identify compliant sites and see what their results look like, rather than trying to guess at parameters for air conditioner adjustment, pavement adjustment etc.

  46. Boris
    Posted Aug 2, 2007 at 7:22 AM | Permalink

    RE27 well the point is, We’re not out to measure mixing rates. We’re measuring site compliance.

    Then all anyone can say from your research is that certain sites are out of compliance. You might want to remind folks of that.

    Looking forward to those A/C experiments. Should be interesting…

  47. MarkW
    Posted Aug 2, 2007 at 7:26 AM | Permalink

    Getting quite good with the strawmen there aren’t you TCO. Not all of the AC units are 100 feet away.
    You are quite aware of this fact. So why do you pretend that 22 air conditioners, ranging from 15 to 100 feet away, are just one unit, 100 feet away?

  48. MarkW
    Posted Aug 2, 2007 at 7:29 AM | Permalink


    Why do you assume that 22 AC units are not a problem? Why do you assume that the other issues raised are not a problem.
    The AGW crowd has declared that the temperature network is a high quality network. These pictures (and many others) show that this claim is in doubt.

    It is up to those who want to restructure the world’s economy, to show that the data they are basing their claims on is good data.

    It is the job of a reviewer to show that there is reason to doubt. It is then up to the producer of the paper to prove that the doubts are either inconsequential, or to redo his paper.

    That’s the way science works. It is not up to the reviewer to prove anything.

  49. MarkW
    Posted Aug 2, 2007 at 7:31 AM | Permalink

    So Boris, you are finally in agreement with the rest of us that out of compliance weather stations should not be used to make climate projections?

  50. Paul Penrose
    Posted Aug 2, 2007 at 8:00 AM | Permalink

    Please keep in mind that at this point Anthony and his volunteers are only collecting data; the analysis has not been done yet, so we don’t know yet what kind of affect these microsite issues have had on the data. Some people are leaning in the direction of assuming that the data is hopelessly corrupted and that the claims of 0.6 degrees of warming for the 20th century are unfounded. Others, like TCO, seem to be in the opposite camp; they are assuming that the microsite issues are negligible and that the current claims about 20th century warming are valid.

    Both camps are making unfounded assumptions at this point. The reality is that right now we just don’t know. I understand people wanting to talk about the data as it comes in, but I think it would be wise to refrain from making any conclusions just yet. Obviously any anthropogenic heat sources in the general vicinity of the temperature sensor will have some kind of effect, and so it’s worth noting when making a census of the sites. Later analysis (hopefully) will determine what the actual impact is, whether .01 degrees or 1 degree. In the meantime I have no problems with curious people asking questions and pondering what the effects are, but intelligent people will realise that this kind of contemperaneous discussion is not a substitute for proper analysis and will hold off making any final conclusions until that analysis is performed. In my opinion, TCO’s shrill diabribe against these discussions is very annoying and unhelpful.

  51. Boris
    Posted Aug 2, 2007 at 8:22 AM | Permalink


    So Boris, you are finally in agreement with the rest of us that out of compliance weather stations should not be used to make climate projections?

    No, you have to prove that the siting issues have an effect, which no one here has proved and for some strange reason no one wants to even attempt to prove. Until you prove something, don’t be surprised if no one pays attention. It should be easy to do and publish if it’s so obvious.

    You see, you guys are still assuming that the errors are both large and small–large enough to be an impact on global temps, but small enough not to be noticed when GISS looks for anomalous data.


    Please keep in mind that at this point Anthony and his volunteers are only collecting data; the analysis has not been done yet, so we don’t know yet what kind of affect these microsite issues have had on the data.

    But Watts and etc. aren’t gathering enough data. Take AC for instance. You need to know installation dates, usage patterns if possible–or at least the usage history of the buildings, you need to see if the thermometer can pick up the AC heat, and how much, and what time of day/season. A picture of an AC unti tells you there’s an AC unit there and nothing else. And since an AC unit–if it affects temp at all–could also introduce a spurious coolng trend, you don’t even know the sign of any anomalies you have at a single site.

  52. JS
    Posted Aug 2, 2007 at 8:35 AM | Permalink

    When NCDC compared the ref station in Rhode Island to the MMTS in Durham New Hampshire they noted that Durham was over 1 degree warmer in the summer. They assumed it was the equipment, probably because they never went to the site. Durham has an a/c unit 20 feet from the sensor.

  53. TCO
    Posted Aug 2, 2007 at 8:49 AM | Permalink

    46, actually I disagree with “coordinating” air conditioners as an issue with other site problems. At least if it is done in the way that you criticized MBH. In that event, you did not quantify the impact of the different posited or certain flaws (and in several cases, such a robustness calculation was trivial). Also, you used flaws with LARGE rhetorical impact and low mathematical impact, but did not quantify the impact. And you conflated the impact of different flaws in such a manner that you used a (rather agreed upon) flaw of decentering to carry the impact of a (not agreed on, not even asserted to be a flaw) of correlation versus covariance matrix.

    Also, air conditioners are not part of the standard. They might have an impact. They might have a huge impact. But if the only thing that is relevant is compliance to the standard, ACs are not an issue.

  54. Jim Edwards
    Posted Aug 2, 2007 at 9:14 AM | Permalink

    #52, Boris:

    Your argument re: A/C is a off here. You say:

    Take AC for instance. You need to know installation dates, usage patterns if possible’€”or at least the usage history of the buildings…

    That would be helpful to correlate hypothetical thermal contamination w/ the historical temp record – if somebody wanted to make historical corrections or demonstrate the effects of A/C over time. If Dr. A is stating his temp measurements are uncontaminated, however, all B has to do is show contamination exists – not how much and when. Let Dr. A go through the work to justify using historical data ‘corrected’ for historical A/C usage.

    You also said:

    And since an AC unit’€”if it affects temp at all’€”could also introduce a spurious cooling trend, you don’t even know the sign of any anomalies you have at a single site.

    This is not right. A heat pump may introduce a cooling influence on the outside environment, but normal A/C units will not. Their condensers emit hot discharge air ~20-30 F above ambient. Heat pumps are usually found regionally – almost none in some parts of the country, many more in various pockets. They may be common where you live, but that would not be representative of the US as a whole.

    An operating A/C unit will either have zero effect [possible], a huge effect [+10-25 F; unlikely, b/c such an effect would be noticed right away], or a very small effect due to mixing [+0.05-0.3 F, as I argued in #25, above]

    Where the data is inherently noisy, and observers are looking for very small annual / decadal average dT to attribute to AGW, it’s quite possible that “Large and small” contamination effects due to A/C would not be noticed as noise.

  55. Mark T.
    Posted Aug 2, 2007 at 9:21 AM | Permalink

    No, you have to prove that the siting issues have an effect, which no one here has proved and for some strange reason no one wants to even attempt to prove.

    The assumption by the climate scientists is that these sites are “good,” and they clearly state this as has been documented in numerous sources. The onus is on them to prove that assumption, not the other way around. An auditor’s only “requirement” is to identify problems, not prove the problems are meaningful. Yet again, science seems to get modified for whatever works in favor of the alarmist.


  56. Michael Jankowski
    Posted Aug 2, 2007 at 9:28 AM | Permalink


    No, you have to prove that the siting issues have an effect

    If siting issues didn’t have an effect, why would there be siting guidelines created in the first place?

  57. TCO
    Posted Aug 2, 2007 at 9:30 AM | Permalink

    If the “onus is on them”, than why didn’t they already do it? Going in, to try to publish something that essentially says that “the onus is on them” is a likely non-starter. Anyone could publish then and just demand that authors account for whichness of whys.

    Every experiment has potential confounding variables. How much work is done to allow for them is a soft issue. Previous authors have made some efforts to account for potential problems. And reviewers and editors have looked at it and allowed publishing as is, as a sufficiently scientific product to be additive (note this is not a promise, endorsement that no one will ever find a confounding variable that changes the original conclusions.)

    You might be able to go in with a much, MUCH, more moderate, and even-handed study that just shows the percent air conditioners or whatever it is (essentially the value of this publication would be as data reporting, NOT position arguing). You could make a moderate comment that the data might be interesting if people want to try to understand microsite contamination, but that you have not quantified that and don’t make a claim.

    Or really much better yet, do a bit more work and contribute by actually showing an impact. That’s the sort of thing that is worthy of taking up space in a journal. In a library.

  58. TCO
    Posted Aug 2, 2007 at 9:41 AM | Permalink

    57: I don’t know. Maybe it does, maybe it doesn’t. We don’t seem to have much information on the rationale for siting or instrument design. I don’t know how in depth and “nuclear” (detailed) that work is. Are there calibration standards and frequencies? Standard-setting studies? Follow-on published work to look at confounders? Heck, let’s get the damn rooftop study.

    Let’s at least see what is really known already! Not be surprised by it. I was amazed that Antony did not post the standards first thing as part of his project. That he was full steam ahead and someone had to suggest to him to put that on his site as a sorta important document!

  59. Mark T.
    Posted Aug 2, 2007 at 9:46 AM | Permalink

    If the “onus is on them”, than why didn’t they already do it? Going in, to try to publish something that essentially says that “the onus is on them” is a likely non-starter. Anyone could publish then and just demand that authors account for whichness of whys.

    Don’t get me wrong, TCO, I agree that such a statement isn’t going to get you far in the literature particularly. However, the first step is to identify problems, which is what Anthony is doing. He’s not done yet, so analysis, and subsequent conclusions, are not ready for prime-time.

    Boris’ statement is completely contrary to accepted scientific practices, and the only point I was making was that this is how science has been corrupted with the new faith. The climate scientists that fail to prove assumptions _should_ be ignored, but that’s not what’s happening. Right or wrong, it will be up to the skeptics to prove the case, whichever direction it leads. Personally, I just want the science to be _right_ before we go screaming about the sky falling.


  60. Steve Moore
    Posted Aug 2, 2007 at 9:55 AM | Permalink

    It’s apparent to me that not too many folks here have ever gone through an ISO 9000 Certification.

    Most companies have Standard Operating Procedures. Not all of them adhere to their own procedures. In ISO, you document what you DO; not what your published procedures are. A third party can then look at what you document and decide if your actual procedures provide the level of quality they desire.

    I look at as a kind of preliminary ISO certification: they are trying to document reality. When they are finished, we can look at the result and decide if the quality is there — not before.

    By the way, done properly, an ISO can serve as a “wake-up call”. Maybe that will be the case here.

  61. Mark T.
    Posted Aug 2, 2007 at 9:55 AM | Permalink

    I have, several times… man does it suck.


  62. Boris
    Posted Aug 2, 2007 at 10:00 AM | Permalink

    An operating A/C unit will either have zero effect [possible], a huge effect [+10-25 F; unlikely, b/c such an effect would be noticed right away], or a very small effect due to mixing [+0.05-0.3 F, as I argued in #25, above]

    A spurious trend could be introduced by changes in AC/building usage, especially if the station was moved to near an AC site and the jump in warmth is accounted for in the GISS adjustments. Obviously this will not happen very often–or at all–but it shows the limitation of the snapshot audit.

    And we are assuming that the AC can affect temps at many of these sites. Why people are afraid or unwilling to test this assumption I do not know.

  63. VirgilM
    Posted Aug 2, 2007 at 10:02 AM | Permalink

    Let’s say that it is windy. The point heat source (the air conditioner) now has to heat a much larger volume of air. If the wind is blowing at 10 miles an hour, a very large volume of air will be moving past the air conditioner unit. Is the unit adding enough heat to this large volume to make a detectable change in temperature 15 feet away from the unit? Something to think about. I’ll have to find myself a digital thermometer and do some experiments with my unit.

  64. Dave Dardinger
    Posted Aug 2, 2007 at 10:05 AM | Permalink


    You’ve been picking on ACs. We’ve posted some messages already which indicate that a constantly running AC 15′ away from a thermometer (actually this isn’t totally right since the math merely calculated an average over a 15′ radius but we’re just doing BOE calculations so far), will produce about 33 w/m^2. Treating it like the sun, let’s divide by 4 giving 8 w/m^2. This is still twice the forcing of doubled CO2 albeit in a localized area. I think this would be enough of a calculation to justify being concerned about the effect, even though it doesn’t prove anything.

    But in picking on ACs, you’ve cannily ignored more important bias such as building siting. The spec calls for a distance of greater than 4 times the height of a building for placement of a station.

    Now the math for determining how much affect this produces will be rather hairy, but important. Can anyone come up with either some previous papers or do some quick calculations about what the results would be? Some things to be considered would be path of the sun through the sky, material of the wall, distance from and height of the wall, etc. Measurements from IR meters of actual walls and what the air and ground temperatures are in the vicinity are would also be useful.

  65. Mark T.
    Posted Aug 2, 2007 at 10:06 AM | Permalink

    Why people are afraid or unwilling to test this assumption I do not know.

    Why don’t you let Anthony finish his survey before criticizing him for any potential conclusions that might be drawn from the survey?


  66. david cobb
    Posted Aug 2, 2007 at 10:08 AM | Permalink

    the existance of any, let alone many, grossly contaminated
    sites demonstrates that the QC procedures used are,at best,
    inadequate and at worst, biased.

  67. TCO
    Posted Aug 2, 2007 at 10:08 AM | Permalink

    Boris, one could imagine a useful contribution made by studying some small number of ACs in a very detailed way, to get some view as to whether they matter or not, some at least, better than current zero level of knowledge that exists on their impact. If the impact for very close ACs is shown to be negligeable, then you just publish that and disregard all the detailed sampling, picures, etc.

    If you do find an impact, then you could publish something showing that basic understanding and then making a list of what places have an AC. It would not be sufficient to correct all the sites. But would allow for some feel as to at least bounding the issue. Comments would need to be even-handed and eschew conclusions. (this is not the pattern with the screamers here.)

    Of course even better would be a study that shows all of what you are asking for. But I would argue that this may be hard to get and would be unreasonable, if say very strong impacts were noted as at least possible.

    I think we should add knowledge and not let “lack of perfect” stop what is good. (I feel the same way when people here want to throw out every tree ring inference.) The important thing is to be open to imperfect measurements (since all are imperfect in some sense, and since useful inferences are contained in them), but to be fair and cautious about making strong inferences.

  68. Douglas Hoyt
    Posted Aug 2, 2007 at 10:10 AM | Permalink

    I would toss this site out rather than use it in any temperature reconstructions. Problems include nearby pavement, nearby wall, and overhanging trees, let alone the ACs’.

    The only way to let it back in a temperature reconstruction would be detailed studies of the nearby heat flows and some reasonable proof that biases of a few tenths of a degree can be calculated and removed.

  69. TCO
    Posted Aug 2, 2007 at 10:17 AM | Permalink

    69: Decisions to exclude data should not be made sanguinely. (this is true whether a warmer or denier does it.) Note, I’m not saying to include all data either. I’m saying don’t be sanguine about excluding things. What if you scrapped every Census response that had a flaw on a single question? Would that be a valid adjustment? What about the lost information in the other questions? this is a much trickier issue than hoi polloi realize. I know enough to know that I’m stupid here, that a sampling statistician would help. You lot are just sanguine in your ignorance and in decisions.

  70. SteveSadlov
    Posted Aug 2, 2007 at 10:19 AM | Permalink

    RE: #61 – That is a great point. This is indeed like the early days of ISO9000. And the reactions I am seeing from many remind me a lot of the reactions I saw to ISO9000 back then …. LOL!

  71. KevinUK
    Posted Aug 2, 2007 at 10:19 AM | Permalink

    #58 TCO

    I used to enjoy your posts on this blog. You quite rightly used to push Steve on ceratin points which is good but your lastest posts on Antony Watts efforts are changing my opinion of you which is sad. You’re looking more and more like a troll everytime you post.

    It’s patently obvious to anyone who cares about the application of the scientific method (which you used to do) that what Antony, Russ and the other volunteers are doing (unlike Karl et al and it would most climate scientists) is applying the scientific method. They are attempting to falsify the claim made by Karl et al that the USHCN network consists of ‘high quality’ weather stations.

    The adjusted (homogenised) measurements from these weather stations lie at the heart of the claims for ‘unprecendented’ warming due to AGW in the later half of the 20th century. You should be welcoming the fact that this claim is being subjected to srutiny/audit and not nit-picking over Antony, Russ et al’s unfunded efforts.

  72. MarkW
    Posted Aug 2, 2007 at 10:24 AM | Permalink

    I see are AGW alarmists are back to the claim that unless other people can disprove their assumptions, then their assumptions must stand.

    And to think, they accuse others of not doing science.

    Boris, if you want Anthony to gather every conceivable piece of information about each site, can I presume that you will be willing to help pay for this?
    Or are you just going to use this as yet another reason why we should just give up and assume the data must be perfect.

  73. MarkW
    Posted Aug 2, 2007 at 10:29 AM | Permalink

    Next time I submit a paper my prof asks why I didn’t account for a factor Z, I’m going to tell him that TCO and boris have assured me that unless the proffessor can prove to my satisfaction that factor Z in question makes a difference, I don’t have to account for factor Z.

    Who here wants to bet that my prof is going to say OK, and give me an A anyway?

  74. Anthony Watts
    Posted Aug 2, 2007 at 10:30 AM | Permalink

    Just for the record, TCO. my name is spelled A N T H O N Y

    not “Antony”

  75. MarkW
    Posted Aug 2, 2007 at 10:33 AM | Permalink

    A station that fails to live up to minimum standards should be discarded, unless someone can prove that the standards violations have no impact on the data, or can prove that they have a method for accurately removing the contamination from the data.

    That’s not only an easy decision, it’s basic science.

    It is not the job of the reviewer to prove that the contamination is serious.

  76. TCO
    Posted Aug 2, 2007 at 10:34 AM | Permalink

    Sorry, I really do think you are a nice guy, Anthony.

  77. Dave Dardinger
    Posted Aug 2, 2007 at 10:34 AM | Permalink

    re: 370, Taco,

    think about why a point source with mixing is a different problem with a plane with one way impact.

    I submit there is no real difference. The enhanced greenhouse effect comes from the low level air being heated by additional CO2 which absorbs IR from the ground and reemmits some of it toward the surface. Please note that this is primarily because of the additional CO2, not because of the additional temperature of the ground resulting from the initial heating (which is a secondary effect). Therefore, while the heating of a mixed atmosphere in the vicinity of an AC would be localized, it should not be neglected as far as changing low-level air temperatures. And low-level air temperatures are what the thermometers are measuring after all.

  78. KevinUK
    Posted Aug 2, 2007 at 10:49 AM | Permalink

    #73 Anthony

    Profuse apologies for spelling you name wrong THREE times. Won’t happen again honest!

  79. Posted Aug 2, 2007 at 10:51 AM | Permalink


    Taco … LOL

  80. Earle Williams
    Posted Aug 2, 2007 at 10:56 AM | Permalink

    Re #63


    And we are assuming that the AC can affect temps at many of these sites. Why people are afraid or unwilling to test this assumption I do not know.

    You have flat out refused to take this on, and yet you criticize others for not doing so. Things that make you go hmmmmm.

  81. Bill F
    Posted Aug 2, 2007 at 11:01 AM | Permalink

    If the “onus is on them”, than why didn’t they already do it?

    We have a winner!!!! Ding ding ding!!!! Congratulations TCO, you just asked the question that is at the heart of the whole surface stations project! GISS and NOAA claim to have a high quality temperature monitoring network where they can distinguish rural uncontaminated sites from urban sites affected by anthropogenic influences. Yet they have never published any audits or undertaken steps to publically PROVE that claim. Guys like Karl, Hansen, Jones, etc. claim that they are using rural sites unaffected by urban or developmental bias…but they show NO PROOF that they have actually gone to any of the sites or audited them to show that their assertion is actually valid. The million dollar question is why isn’t anybody asking these guys to actually prove that claim during the peer review process? Pielke has published plenty of papers pointing out potential station siting problems, yet the peer reviewers for guys like Hansen and Jones seem to ignore that evidence and accept the claims of high quality stations at face value with an absence of proof.

    As Steve M cited above, Hansen uses “lights” as evidence of possible anthropogenic bias at a station, so why does he get a free pass on using lighting as a proxy for possible bias, but the close proximity of AC units that generate way more heat than a light are being questioned as a possible source of bias? That just doesn’t make sense to me. Nobody is trying to do a study of microsite biases here…they are trying to show that just because there are zero lights, it doesn’t mean that you can assume the station has no possible biasing factors, which is precisely what Hansen, Jones, and others have done in their published papers and nobody has called them on it during review.

  82. David Smith
    Posted Aug 2, 2007 at 11:43 AM | Permalink

    Article on tree coverage, parking lots and temperature ( link ).

    There are also several interesting references in the literature section.

  83. Murray Duffin
    Posted Aug 2, 2007 at 11:58 AM | Permalink

    Boris and TCO, adjustments are being made for assumed UHI or land use effects in the official surface temperature averages. Are you demanding that these adjustments be based on known measured effects, applied with mathematical rigor, or do you believe that to be the case without question, or does your concern apply in omly one direction? Please reply to the question. Murray

  84. Jim Edwards
    Posted Aug 2, 2007 at 12:56 PM | Permalink

    #63, Boris:

    You say:

    A spurious trend could be introduced by changes in AC/building usage, especially if the station was moved to near an AC site and the jump in warmth is accounted for in the GISS adjustments. Obviously this will not happen very often’€”or at all’€”but it shows the limitation of the snapshot audit.

    I don’t think it’s that “obvious” that the spurious trends you admit could occur are accounted for and won’t happen often, or at all. The one thing clear from our whole argument here is that nobody knows how much a particular site is heat contaminated, even after we’ve documented the site environs much better than those who ‘adjust’ or use the data for publication or policy-making have done. How can we say with any certainty the frequency of contamination and the correct amount to ‘adjust’ them ?

    Well, you could say the same for a “snapshot audit” of temperature. You can’t demonstrate a temperature trend with a single temperature reading any better than you can prove a spurious trend with one visit to a measurement site. Should we stop measuring temp, then ? [That’s a rhetorical Q, if you can’t recognize it…]

    You’re really arguing the need to ‘audit’ the measurement site at more frequent / regular intervals. Who’s responsibility should that be ? [Another rhetorical Q…]

    Aren’t you really arguing Anthony’s fundamental underlying point: that NOAA should have been checking these sites themselves all along ?

  85. TCO
    Posted Aug 2, 2007 at 1:30 PM | Permalink

    Dave, please do read Jim’s comment. Take a second and do it, before responding. It is not as simple a question as you think it is to use heat fluxes to estimate impact on the thermometers. Mixing is an important issue here. What you need to know is how much the volume of air immediately adjacent to the boxis heated up. This is not a simple function of a hemispherical spreading problem. The degree to which mixing occurs to transmit that heat to other volumes is important.

    82: They made an effort. They went through gate keepers. Maybe they messed up. But to move the discussion forward, you need some quality of argument here. Just saying that they did not look carefully enough is insufficient.

    84: My concerns do not only apply in one direction. I am truthful and fair. I don’t see the need to discuss independant flaws though. I really see this as the type one fallacy. The Clinton lied so Libby should get away with it mentality. And it really is intellectually independant. I’ve been asked this question in a huffy puffy manner about a butt billion times and I won’t generally respond to it.

  86. TCO
    Posted Aug 2, 2007 at 1:32 PM | Permalink

    84. In fact, I’m primarily a skeptic, not a warmer. It’s exactly for the reason that I think that truth and ethics and searching for insight regardless of what side it is on, that I make the comments I do, when I see poor quality of argument by the skeptic side. But, it really really doesn’t matter who I fault for what. Either the criticism is valid or not.

  87. Kenneth Fritsch
    Posted Aug 2, 2007 at 2:07 PM | Permalink

    Anthony Watts has and is providing a service that is rather uniquely suited to the internet and particularly a blog site. He has received much public attention from his efforts and drawn attention to them. He is not preparing to write or submit a peer reveiwed paper or even do a prescribed audit. He is calling attention to a potential problem and doing it in the truest and best traditions of the internet.

    What is more important to observe right now (as opposed to whether the process being used is up to publication level) is the reaction that he will receive from the climate science community in general and those responsible for these sites and using data from them. They are the people that would ultimately publish papers to quantify the problems and to recommend corrective actions for potential problems.

    It is understandable that those with the most at stake in the validity of these data would not necessarily react favorably to what is being exposed and to almost extinctively indicate that all has already been considered. There are also those amongst us who have gut feelings that scientists or at least groups of scientists are better at qualifying their methods than someone either as a scientists from outside the area of interest or a non-scientist. I think most of us here would tend to give “expert” scientists some benefit of the doubt in their field of expertise, but that does not mean ignoring potentionally important issues such as Anthony has exposed here and asking the necessary questions of the involved scientists. One does not need to be an expert to judge the worth of the scientists’ responses.

    Quite frankly I would not at this point want to touch a hair on Anthony’s head, or at least not until I have heard some satisfactory responses from those involved. If he did a perfectly scientifically based analysis with a published paper as a goal, I would suspect that several exchanges of papers on the subject would take much longer to illicit a response from the responsible parties and it would occur far from the public eye that can sometimes rightly or wrongly encourage a quicker solution.

  88. Dave Dardinger
    Posted Aug 2, 2007 at 2:28 PM | Permalink

    re: #86 TCO,

    Mixing is an important issue here.

    Of course it is, but you have a rather selective way of pushing for it. How about admitting that 22 air conditioners within a hundred feet of a site will result in mixing which will impinge on that site? That’s the whole thing with UHIs as well as microclimates. And I’ll add the obligatory point that we’re really interested in trends not absolute temperatures. We all know that’s the important thing, but passersby may think we don’t because not every post states it. Ultimately we’d like to know when the ACs were added, how their use has increased (or not) and that sort of thing. But for the moment we want to list what potential biases may be present.

    You get mad because we ooh and ahh over the wrapping paper, thinking, perhaps, that we forget that doesn’t determine if the present’s any good or not. But if you say you’ve never shaken a wrapped present trying to figure out what it is, I won’t believe you. Let us shake our ACs and parking lots. The world needs fun and frolic as well as rigor and realism!

  89. TCO
    Posted Aug 2, 2007 at 2:28 PM | Permalink

    84: I responded to you, but it was deleted.

    John/Steve: Either let me answer or erase 84. Prefer erasing 84 as “what TCO complains about in Mann-land” is off topic.

  90. Posted Aug 2, 2007 at 2:28 PM | Permalink

    TCO says: August 1st, 2007 at 8:42 pm

    Antony, so what is the delta T impact from one AC 100 feet away? (I’m not trying to say that it is small with this rhetorical question, I’m traying to say that you DON’T know it.) You don’t know the mixing rate. The throw. You have some BTU numbers there to impress the polloi. But the key issue is mixing (and in the out doors!)


    Impress this engineer. Tell me what the number is. Tell me what it is in summer. Tell me what it is in winter. Tell me what it is in the morning. Tell me what it is at night.

    Then we can have a little correction party to make the numbers come out right.

    Other wise I will have to assume as you do that the data from this station should not be used in a climate survey.

    BTW siting rules are designed to eliminate the very uncertainties you decry. The rules should be followed.

  91. Bill F
    Posted Aug 2, 2007 at 2:29 PM | Permalink

    86, How do you know they made an effort? They won’t reveal how they make their adjustments or in some cases even what sites they used for their “rural network”. They don’t publish any methodology for how they make the decision to adjust station data upwards or downwards, and people using their “adjusted” data are not asked to justify the adjustments as a part of the peer review process.

    How can you say they have passed through a “gatekeeper” when the gate keepers aren’t even willing to require them to make data available, much less ask that they justify the adjustments that are made to the data. Everybody just accepts that GISS and other agencies’ adjustments to raw data are valid, when they never produced anything to justify on a station by station basis exactly what they did and why they did what they did to adjust the data. Arguments about measuring the heat output of AC units pale in comparison to the flippancy with which people accept GISS adjusting data up or down by several degrees C without so much as a notation of why the adjustment occurred or what methodology was used.

  92. TCO
    Posted Aug 2, 2007 at 2:30 PM | Permalink

    My point is that you can’t solve the problem by referring to heat flux, Dardie. Please give me an estimate, even a bottom of the envelope estimate of delta T from a 100 foot away AC.

  93. TCO
    Posted Aug 2, 2007 at 2:32 PM | Permalink

    Simon: I don’t know the answer. I’m reacting to the people who said that they knew it would have an impact. Or made other huffy puffy comments to that effect. (numbered fallacy whatever: excluded middle).

  94. Dave Dardinger
    Posted Aug 2, 2007 at 2:42 PM | Permalink

    re #91 TCO,

    My point is that you can’t solve the problem by referring to heat flux, Dardie. Please give me an estimate, even a bottom of the envelope estimate of delta T from a 100 foot away AC.

    I’m not sure why some of your messages are being deleted. Possibly you’ve been sufficiently active to catch the eye of the Spam Catcher.

    BTW, since you went back to “Dardie”, let me ask a question which didn’t get officially answered when you were here before. Near the end of that time there was a message or two claiming you weren’t an individual but a consortium of grad students or the like all using the same “style”. For me this seemed possible since there were so many messages so quickly from you and often it seemed like you’d forgotten things you’d just been told. I tried to find out from Steve if this were the case or not, but neither he or you denied it in so many words. I don’t really care what the answer is, but it makes a bit of a difference in figuring out how to respond to your messages.

  95. Posted Aug 2, 2007 at 2:45 PM | Permalink

    #71 Steve S.,

    I was involved in the ISO wars. All ISO guarantees is that there will be a procedure. It will be documented. It will be followed.

    The correctness of the procedure is beyond ISO 9000.

  96. Posted Aug 2, 2007 at 3:04 PM | Permalink

    TCO asks:

    Please give me an estimate, even a bottom of the envelope estimate of delta T from a 100 foot away AC.

    Bottom of the envelope? Less than 50 deg C. Depending of course on whether the flow is laminar or turbulent. Any guesses as to the Reynolds number at various air speeds? Do the trees help or hurt?

    Will there be a winter effect from heating the building?

    If the AC is on will that affect the radiation temperature of the building?

    If you put pig swill in fine wine bottles does that make it a fine wine?

    BTW ever wonder what the effect of an air conditioner 15 ft away is? I have. From time to time. What do you think it is? Less than 50 deg C? Less than 10 deg C? Why?

    I think I have the perfect solution to global warming. Turn off all air conditioners within 100 ft of a monitoring station. Think of all the CO2 that could be saved. And weather data.

    Note the 4X rule for obstructions is barely adequate. A good site would use a 10X rule as is used in siting wind turbines or measuring fluid flow in pipes.

  97. Jim Edwards
    Posted Aug 2, 2007 at 3:11 PM | Permalink

    #91, TCO

    OK, here’s my weak attempt.

    3 mph wind,

    A bounded box** similar to a wind tunnel 200′ x 200′ x 50′ high exists, through which wind blows,

    all condenser discharge air isotropically mixed within sample box,

    No condenser discharge air leaves the box through the sides or roof – all mixes w/ ambient wind and passes by temp probe before exiting box at far end,

    total A/C output = 22 units x 0.5 tons x 1.3 [inefficiency] x 25% duty cycle = ~3.5 tons of heating polluting sample box,

    Condensers emit ~ 350 cfm / ton cooling at dT of +25 F while operating
    = 11 tons x 350 cfm / ton x 60 min / hr x 25% duty cycle = 58k ft3 of discharged air w/ dT = +25 F

    Flow of ambient air through box will be 3 mph x 5280 ft / mile x 200 ft [width] x 50 ft [height]= 158.4M ft3 per hour fresh ambient air

    net dT would = (58k ft3 x 25 F) / 158.4 M ft3 = +0.0091 F for all of the 22 A/C units w/in 100 ft. One would expect higher dT with same (unrealistic…) assumptions if focus was on nearby units, only.

    ** Note that I cheated and used a box, rather than a sphere.

    Alternative calculation method would add 3.5 tons / hour = 40k btu to each 158.4 M ft3 of fresh ambient air. That would be 0.00025 btu / ft3 increased energy.

  98. Posted Aug 2, 2007 at 3:15 PM | Permalink

    The “air conditioner adder” will vary with temperature and wind direction.

  99. Posted Aug 2, 2007 at 3:19 PM | Permalink


    Is turbulent flow realistic with a 3 mph breeze?

    If you are in a laminar flow region don’t you have the possibility of “hot spots”?

  100. Posted Aug 2, 2007 at 3:24 PM | Permalink

    I have a song title:

    “If It Wasn’t For Bad Data, I Wouldn’t Have No Data At All”.

    to be followed by

    “Why I Sing The Blues”.

  101. SteveSadlov
    Posted Aug 2, 2007 at 3:29 PM | Permalink

    RE: #94 – And in the current case, those who are in a snit about site inspections are apparently upset about the mere act of establishing that the siting procedures are being followed. Indeed, we have not even begun to assess the efficacy of the siting procedure.

  102. SteveSadlov
    Posted Aug 2, 2007 at 3:36 PM | Permalink

    RE: #98 -In fact, I would argue that the equivalent wind to an enclosure cooling fan system would be knocking down structures and bowling over trees. The winds which Parker claimed to blow away UHI, scaled down to an electronic enclosure, would be nearly nothing. Parker found that the difference between Tmin on “windy” and “non windy” nights was miniscule, because the wind on most of the nights in question was barely putting a dent in UHI!

  103. Posted Aug 2, 2007 at 3:36 PM | Permalink

    Frankly, I find TCO’s comments to be irreleant.

    Anthony is undertaking a phenomenal effort (with the help of volunteers). After all this effort, the most important bit of data we will have will be a single dummy variable: NONCOMPLIANCE.

    NONCOMPLIANCE = 1 if station complies with none or almost none of the standards

    NONCOMPLIANCE = 0 if station complies with all or almost all of the standards

    Yes, there will be a lot of shortcomings such as not knowing the history of compliance/non-compliance. Such as a bunch of judgment calls in the classification of some stations.

    None of that matters much.

    With that variable and the already existing balanced panel data, we will be able to simply test if there is a temperature trend difference between compliant and non-compliant stations and put a magnitude on that. We do not know what that result will be. I do not even want to speculate on what the direction of the difference would be if one is found to exist.

    All that matter for those of us who believe that historical temperature measurements give us very little information about climate trends and that the current station network is not geared towards proper analysis of such trends are two things:

    1) # of compliant stations and non-compliant stations

    2) whether there is a difference in trends between them

    If most of the stations turn out to be compliant or if there does not seem to be much difference between the two groups, that would constitute evidence against my position.


  104. steven mosher
    Posted Aug 2, 2007 at 3:55 PM | Permalink

    RE 95.

    TCO thinks reynolds number is an aluminum foil spec. ( sorry TCO)

    He has a point. An agnostic point. A good point. The point is: “If you claim that AC units corrupt the record
    show the math. Otherwise, STFU.”

    I don’t think I have seen him claim or prove that they have no effect. So TCO, would it be safe
    to say you are agnostic about the impact of AC units? shrugs.

    The most we can say is the site photos put the quality into question. Some confirm the quality.
    Others don’t confirm the quality. And photos don’t in and of themselves, display the delta T, but they raise questions.

    There are 4 main Points to Anthony’s study.

    1. Challenge claims by previous studies that the “network” is high quality.

    This is ankle biting. Biting ankles is respectable. some may like to aim higher
    but that is a matter of taste.

    2. Establsh a OPEN documentation baseline. Nobody has even mentioned this. perhaps
    NOAA can be convinced to complete it rooftop study, to photo doc all the sites,
    to provide researchers with complete documentation of the sites, to develop a
    more robust metadata system.

    3. Refine the estimates of global warming. Restricting analysis to only those sites
    that meet guidelines may lead no change or some change. It will be fun
    and instructive to look at this. To some extent Hansen and Gavin have already
    telegraphed their next punch. “The good sites will show the same trend as the bad sites.
    and the US is only 2% of the land mass. who cares if we missed the century warming trend in the US
    by .2C” NOTE for example, how little is made of the differences between CRU and GISS.
    Finally, the observational record does not matter much to them.

    4. Identify historical Sites that can be coordinated with the NEW NETWORK, the CRN.
    currently the plan is to “homogenize” the CRN with the nearest neighbor. I think
    this requires further scrunity. Best neighbor might be a better criteria. Open question
    of course.

  105. steven mosher
    Posted Aug 2, 2007 at 3:59 PM | Permalink

    Click to access Doc4.3(2).pdf

    The precise exposure requirements for specific instruments used at climatological stations, aimed at
    optimizing the accuracy of the instrumental measurements, are discussed in the Manual on the Global
    Observing System, Part III (WMO No. 544), and in the Cuide to Meteorological Instruments and Methods
    of Observation (WMO No. 8). However, the representativeness and/or homogeneity of the climatological
    record are closely related to the choice of location of the observing site. A station sited on, or near a steep
    slope, ridge, cliff, hollow, building, wall or other obstruction, is likely to provide data which are
    representative of only the site. A station which is or will be affected by the growth of vegetation, including
    even limited tree growth near the sensor, growth of tall crops or woodland nearby, by the erection of
    buildings on adjacent land, or by increases (or decreases) in road or air traffic including that due to
    change in use of runways or taxiways will provide neither representative nor homogeneous data.
    A climatological observing station should be sited at a location which permits the correct exposure of the
    instrumentation (this may occasionally give rise to a split site), and allows the widest possible view of the
    sky and surrounding country if visual (i.e., non-instrumental) data are required. Ordinary and Principal
    Climatological Stations should be sited on a level piece of ground covered with short grass; the site
    should be well away from trees, buildings, walls, and steep slopes and should not be in a hollow. A plot
    size of about 9 m x 6 m is sufficient for outdoor temperature and humidity sensing instruments, and within
    this an area of 2 m x 2 m of bare ground is ideal for observations of the state of the ground and soil
    temperature measurements. A slightly larger plot (say 10 m x 7 m) may be necessary to enclose a rain
    gauge in addition to the other sensors. A rule used in many Services is that the distance of any obstacle,
    including fencing from the gauge must be more than twice, and preferably four times the height of the
    object above the rain gauge. Prevention of unauthorized entry is a very important consideration, and may
    demand enclosure by anything from low, open palings to open high wire fences. Sunshine recorders, rain
    gauges and anemometers must be on sites with suitable exposures and need not be on the same site as
    other instruments. In general terms, anemometers require exposure at a distance from any obstruction of
    at least 10, and preferably 20 times the height of the obstruction, while gauges need some shelter from
    the wind. Automatic stations will normally need a high level of security to protect against animal and
    unauthorized human entry, and require the availability of suitable power supplies, and possibly additional
    protection against floods, leaf debris and blowing sand. A local caretaker is normally needed to cut any
    grass and for other “housekeeping”.
    In addition, Ordinary and Principal Climatological Stations should be located at such a place and under
    such administrative conditions that will allow the continued operation of the station, with the exposure
    remaining unchanged, for a decade or more. For stations used or established to determine long-period
    climate change, such as reference climatological stations and other baseline stations in the GCOS
    network, constancy of exposure and operation is required over many decades. Additional constraints on
    siting apply to Global Atmosphere Watch (GAW) stations es tablished to provide data on atmospheric
    chemical composition (see chapter B2 of WMO No. 49, Technical Regulations). These include the need
    for no significant changes in land use practices within 30-50 km of the site of global stations, and freedom
    from the effec ts of local and regional pollution; for example, from major population centres, industrial and
    extensive farming activities, highways, volcanic activity, forest fires and dust storms. Both global and
    regional GAW stations should be within 50-70 km of an up per air synoptic station.
    When a major rdocation of a climatological station is necessary, or when one station is to be replaced by
    another nearby, it is highly desirable to operate both observing stations for an overlapping period of at
    least one year to determine the effects of changed instruments or sites on the climatological data. There
    are thus distinct advantages in supporting long -term operation of two climatological stations in
    climatologically homogeneous areas.

  106. Jim Edwards
    Posted Aug 2, 2007 at 4:31 PM | Permalink

    #98, M. Simon:

    Of course, you’re right, I said the assumptions are unrealistic. There is no way this system will be accurately modelled w/ pencil and paper; I’ve said that. But it’s also very unrealistic to presume a small A/C unit 100 ft away will have a significant effect on the temp measured [closer units, something else…]. It’s not going to be perfect laminar flow near where the fierce condenser fan is forcibly mixing hot air w/ ambient. There will be a strong vortex w/ high and low pressure zones near the unit. Some of the discharge air even loops through for a second trip through the condenser coil.

    But so far we have:
    zero mixing, hot spot = +25 F dT
    zero mixing, no hot spot = 0 F dT
    100% mixing, 3mph wind, all A/C condenser heat w/in 100 ft magically directed toward temp sensor, whether upwind or downwind = ~+0.01 F dT

    Therefore, if there is thermal inhomogeneity, and a relatively warm plume passes through the Stevenson screen along with a 3 mph wind, it probably affects temp measurement > +0.01 F dT

  107. steven mosher
    Posted Aug 2, 2007 at 4:57 PM | Permalink

    Interested parties with accounts can have a look at this:

    nice stuff here too:

  108. steven mosher
    Posted Aug 2, 2007 at 5:00 PM | Permalink

    Interested Parties.

    perhaps we should be comunicating with OKE

    Found this. havent read it, if someone has access, fire away

    A technique to detect microclimatic inhomogeneities in historical records of screen-level air temperature

    Runnalls K. E. and T. R. Oke

    JOURNAL OF CLIMATE 19 (6): 959-978 MAR 15 2006

    Abstract: A new method to detect errors or biases in screen-level air temperature records at standard climate stations is developed and applied. It differs from other methods by being able to detect microclimatic inhomogeneities in time series. Such effects, often quite subtle, are due to alterations in the immediate environment of the station such as change,, of vegetation, development (buildings, paving), irrigation, cropping, and even in the maintenance of the site and its instruments. In essence, the technique recognizes two facts: differences of thermal microclimate are enhanced at night, and taking the ratio of the nocturnal cooling at a pair of neighboring stations nullifies thermal changes that occur at larger-than-microclimatic scales. Such ratios are shown to be relatively insensitive to weather conditions. After transforming the time series using Hurst resealing, which identifies long-term persistence in geophysical phenomena, cooling ratio records show distinct discontinuities, which, when compared against detailed station metadata records, are found to correspond to even minor changes in the station environment. Effects detected by this method are shown to escape detection by Current generally accepted techniques. The existence of these microclimatic effects ire a source of uncertainty in long-term temperature records, which is in addition to those presently recognized such as local and mesoscale urban development, deforestation, and irrigation.

  109. steven mosher
    Posted Aug 2, 2007 at 5:33 PM | Permalink


    Explain how you are figuring duty cycle. Just curious.

  110. Anthony Watts
    Posted Aug 2, 2007 at 5:56 PM | Permalink

    RE109 Pielke has it here in review:

    And AMS Allenpress has the full pdf here:

    but it looks like you have to subscribe.

  111. Jim Edwards
    Posted Aug 2, 2007 at 6:19 PM | Permalink

    #110, Steven Mosher:

    Experience in the field. Midafternoon on the hottest day in August one might expect the unit to run 100% of the time. During the rest of the year it will be less than that, as the unit is sized for the maximum load – so the unit can cool to set point and turn off quickly during the Spring, early morning, or evening. 30% duty cycle, on average, is a reasonable estimation of duty cycle for an operating A/C unit. Of course there are outliers, but I have no data on peculiar heat loads or low set points here, so I’m using a default value. I chose 25 % for the problem, above, b/c I assumed sensible and latent loads may be a little lower than in the city. I’m also assuming the units are appropriately sized. I could be wrong, maybe it’s 40 % duty cycle. (?)

  112. steven mosher
    Posted Aug 2, 2007 at 6:44 PM | Permalink

    RE 113.

    Funny, when I lived on a slab in LA, we never ran it during the day, but ran it all night long
    as the concrete gave up its heat when the sun went down.

    so at what outside temp would expect the units to go on at? ballpark

  113. steven mosher
    Posted Aug 2, 2007 at 7:02 PM | Permalink


    checking the daily data from USHCN for 1981

    June 17 – sept 25

    About 80 days with a Tmax over 90F
    about 30 days with a Tmax over 100F.

    Not sure this refines your guess at all. FWIW

  114. Jim Edwards
    Posted Aug 2, 2007 at 7:57 PM | Permalink


    That’s going to depend on too many factors, primarily the chosen setpoint for the controlled space and the amount of sensible and latent heat that is being produced within the building, and also how thermally isolated the inside of the building is from the outside. Depending upon internal load, it could turn on when outside temp = 62 F; it may not cycle in earnest until outside temp = 80 or 85 F. I can’t give you a ballpark number w/o knowing the application. We don’t know how many offices are even occupied, here.

    If you’re talking about a residential setting w/ a slab floor and good insulation, it may be that the interior doesn’t get hot during the day while residents are at work, but evening A/C is necessary to counteract the sudden load when people come home and start turning on lights, TV, opening fridge, and cooking. Just sitting on the couch releases heat that must be removed by the A/C. This is why Inuits above the Arctic circle need A/C, now that they’re relatively affluent. [Your concrete is giving off heat during the day, too, by the way, it may not seem as obvious as when you have cool, evening air temp.]

    But that leaves ~285 days w/ T

  115. steven mosher
    Posted Aug 2, 2007 at 9:38 PM | Permalink

    Thanks Jim

  116. steven mosher
    Posted Aug 2, 2007 at 9:44 PM | Permalink

    RE 111.

    I suppose I will subscribe. I get tired of going there and not being able to
    read stuff.

    It might be interesting to strike up a convo with OKE. Everything of his I’ve had the
    opportunity to read has been quite accessible and I actually feel like I’ve learned
    stuff from his papers.

  117. VirgilM
    Posted Aug 2, 2007 at 9:47 PM | Permalink

    I borrowed a hand held electronic thermometer and annemomenter from the office and I took some crude measurements at varying distances from my condenser. I measured the temperature coming out of the unit and compared it to the temperature about 50 feet away from my house. I did this three times and the dT was about 13 degrees F. I will note that it was overcast this evening with nearly calm winds.

    I then took measurements at 30 ft, 20 ft, 10 ft, 5 ft, 2 ft, and 1 ft away from the unit at a height of about 5 feet above the ground.

    I would say that there was no detectable change in temperature at 30, 20, and 10 feet. At 5 feet, I got .8 F colder, .7 warmer, and .8 warmer. During the .8 F colder measurement, there was a slight 2 mph breeze from the north…right to left across my measuring line. There is a definiate warm up from 2 to 1 feet. About 1 deg F. But between 2 and 10 feet it was difficult to quantify the delta T because of the occasional occurance of slight winds and to the fact that temperatures were falling in general.

    I also measured the speed of the air coming out of the top of the unit. It was 20 mph. I could only feel the turbulance from the fan 18 inches from the edge of the fan at 7 feet above the ground. My unit does a good job of sending the warm air up and the fact that the air is 13 degrees warmer than the surrounding air would also help the air rise above a thermometer nearby.

    So under these conditions, within 5 feet, the temperatures will certainly be contaminated. Beyond 10 feet, should be okay. Between 5 and 10 would be questionable. Very crude experiment with lots of flaws, so question my results greatly.

  118. Anthony Watts
    Posted Aug 2, 2007 at 10:23 PM | Permalink

    VirgilM -118

    Thats good to hear about your experiemnt.

    This is a ground mounted unit with a top venting fan that draws air in from the sides, correct?

  119. steven mosher
    Posted Aug 2, 2007 at 10:24 PM | Permalink

    RE 118.

    Virgil you realize that measuring the wind will lead to insanity.

    Heisenberg said: When I meet God, I am going to ask him two questions: Why relativity? And why turbulence? I really believe he will have an answer for the first

    The other day I was on a flight with my kid. He looked out the window and stared at the wing.

    It had Vgs ( vortex generators, those little tiny tringular shaped things)

    He asked what they were. I explained that he didnt really want to know.

  120. MarkW
    Posted Aug 3, 2007 at 5:14 AM | Permalink


    That’s just one A/C unit. Of course there are 21 other units, each of which is progressively closer to the sensor.

  121. MarkW
    Posted Aug 3, 2007 at 5:18 AM | Permalink


    I don’t think it’s a good assumption to assume that these units are sized for the maximum load. More than likely they are sized to the maximum available power.
    In years past, out of the way places like this rarely had the kind of electric service that you would find in bigger cities. You didn’t want to blow any breakers if all 22 came on at the same time. I’ve visited places like this in the summer and I can remember the temperature in the building climbing even though the AC was running full blast. It didn’t catch up again until the sun started to go down.

  122. MarkW
    Posted Aug 3, 2007 at 5:22 AM | Permalink

    With all of the buildings and the trees at this site, any air flow is going to be much more turbulent. That will tend to keep the heat from these units closer to the ground.

  123. Geoff Sherrington
    Posted Aug 3, 2007 at 6:39 AM | Permalink

    Has this question been answered before? If the response time of the temperature measurement device is rapid, say a second, a single gust of atypical wind will affect the temp pattern. This could happen several times a day. If was atypically hot and it exceeded the normally hottest part of the day, a bias would be created.
    I can so no point in this whole discussion, except that if scientists are going to use data, they sould ensure that it is collected to agreed, documented standards, is repeatable if capable of replication and that error estimates of both accuracy and precision accompany the write-up.
    Is that asking too much?

  124. reid simpson
    Posted Aug 3, 2007 at 6:55 AM | Permalink

    Jim Edwards: the point of Andrey Levin in #41, imo, is that it only takes contamination at Tmax to corrupt the data and it is at Tmax that the A/C’s would likely be running. Would the duty cycle make much difference around the time of Tmax on those days hot enough to run the A/C’s?

  125. VirgilM
    Posted Aug 3, 2007 at 10:22 AM | Permalink

    Anthony RE119,

    It was a ground unit about 2.5 feet high. The diameter of the fan is about 18 inches. The intake is on the sides except for half of the side of the unit next to my house.

    If I was really scientific, I would have to install a unit 100 feet away from my house to ensure that the warmth from the house was not biasing the results.


  126. TCO
    Posted Aug 3, 2007 at 10:53 AM | Permalink

    It’s also unlikely even on a very hot day, that all the units will be running. Some rooms will not be occupoied.

  127. SteveSadlov
    Posted Aug 3, 2007 at 10:58 AM | Permalink

    To uplevel the discussion, what is the impact of the human development known as Happy Camp on surface temperatures measured within said development?

  128. Bill F
    Posted Aug 3, 2007 at 11:03 AM | Permalink

    It only takes one TCO…most likely the one 15 feet from the sensor. If it is running at Tmax then there will likely be some impact. I know I shouldn’t make that statement without providing a complete literature search, thermodynamic calculations, and peer-review. But what the heck, I am feeling a little daring on a friday. The fact that the curator has noted an impact on the temperatures due to the sprinkler system already tells us that the data at the station is flawed…but lets just assume that it all balances out somewhere right? Thats what the “professionals” seem to want to do.

  129. Sam Urbinto
    Posted Aug 3, 2007 at 2:41 PM | Permalink

    #18 #40, others later Proves my point, thanks for agreeing with me TCO. “…what is the delta T impact from one AC 100 feet away… …You don’t know the mixing rate. The throw…”

    We don’t know what the effect is = we can’t compensate for the effect. We don’t KNOW if the temperature being measured is accurate. We don’t even KNOW what we’re measuring. I appreciate it.

    This is not about measuring anything, it is about “Does the station meet the standards?” NO. Period. NO. A side issue is “Does lights=0 mean a station meets the standards?” or “Does lights=0 mean the station is rurally situated, free from site contamination?” NO. Period. NO.

    Perhaps now that we agree we can’t measure the temperature, I’m sure you’ll stop asking questions about it that have nothing to do with “if the sites meet standards” and are not pertinent and take us off the discussion of if the sites meet standards and the other issues (lights=0=rural?) that come up while seeing “if the sites meet standards.” As I’m sure you now understand many do not!

    #52 Ah, I see Boris agrees with me too that we don’t know what we’re measuring! “…an AC unit’€”if it affects temp at all’€”could also introduce a spurious coolng trendyou don’t even know the sign of any anomalies you have at a single site…” Glad to see you too know that all that’s being done is finding out if the site meets the standards or not, so we can later study what effect it has and remove/fix it, also!

  130. Sam Urbinto
    Posted Aug 3, 2007 at 3:31 PM | Permalink

    #55 Jim, an A/C could produce a cooling influcence if the hot air being exhausted up drew in cooler air than normal towards itself, I would think. And in #85, bingo.

    #57 Yes, my point exactly; there are standards for a reason.

    #59 Ooops, spoke too soon. Actually Mosher quoted why there are standards in #19 and linked to more in #106, and it’s obvious if you’re trying to measure something, you try and minimize the things that might influence what you’re measuring.

    #63 Dang, wrong again. Nobody’s testing that assumption because it’s not what the project is of course, if there even is that assumption. Although I’d say it’s safe to assume that at least in some cases, the temperature would be affected in some way at some times.

    #87 (maybe 86 too) TCO, if that is true, then I fail to see what issue you have with finding out if the sites meet standards or not. And even more confusing why you keep treating this as if it’s an AC easter egg hunt. Some of the other things you’ve posted in other threads confuses me on this issue, also.

    #91 I’m worried that it even needs to be pointed out that siting rules are designed to eliminate uncertainties. #97 lol AC delta T is “less than 50 C”!

    #105 I think you summarized the survey’s implicit or explicit points pretty well. That’s just me though.

    1. Find out if the network is high quality rather than take it for granted.
    2. Establish an open documentation baseline.
    3. Refine estimates of global warming.
    4. Identify sites to coordinate with CRN.

    #118 123 124 129 Same thing I said about 55; the A/C is going to at least change the air flow of the area in some way at some times. (I suppose on a calm day, one could blow smoke into the general area of the AC unit and see what kind of flow there is (as well as measure it/photograph it with an infrared camera or sensor or whatever))

    But again, as I also said “We don’t know what the effect is = we can’t compensate for the effect” and we don’t really care (at least at this time) anything other than “if the sites meet standards.” This one doesn’t. Next.

  131. Sam Urbinto
    Posted Aug 3, 2007 at 4:26 PM | Permalink

    I agree. Why are we talking about AC units? There’s 22 around this site, it doesn’t meet standardds, let’s move on.

    If the discussion has gone this way, perhaps it’s because of what you say and how you say it? If so, stop and we won’t be “[snip]” any longer.

    We’ve pretty much already ascertained that we don’t know the specific effects of whatever contamination variables are there by themselves, much less as a whole. And AC is just one.

    I know enough about AC units to know they create air, therefore at least in some way impact air flow. How exactly? I don’t know. I don’t care. Not the purpose.

    It’s not what the 22 are doing, specifically, since we don’t know it. It’s just that the site is out of standards so badly — that is the key here, and the topic.

  132. steven mosher
    Posted Aug 3, 2007 at 4:29 PM | Permalink

    RE 131.

    I have to admit that I have this visceral adverse reaction to adjustments. Call me a radical

  133. steven mosher
    Posted Aug 3, 2007 at 4:56 PM | Permalink

    Killing one stone with two birds.. or something like that.

    I loved the work that Jim did. [snip].

    So it made me think. If I put a thermometer in a CRS ( a wooden Box) and blew Hot air

    directly on the structure what would the response of the thermometer inside the structure look like?

    What if I blew cold air on it?

    The two birds: AC units and Parkers UHI study.

    other question. are aspirated units more or less susceptible to hot air contamination?

  134. Sam Urbinto
    Posted Aug 3, 2007 at 5:20 PM | Permalink

    #134 Right! You radical empiricist you. 🙂 As I mentioned before if the goal is to have good data, why not make it as pure as possible, regardless of how “important” it is?

    #135 I will answer your question for you TCO. I answered your post having a feeling yours would be deleted as well as my reply to it. However, only the word of yours I used was removed, and I had written my post in the context that it would still be understandable even if yours was not there. This one here might get snipped too; not up to me to decide what stays and what doesn’t. It’s obvious from what I wrote in #133 that you said something about the amount of discussion on AC units. And I talked about that on topic — pointing out it wasn’t the effect you’re fixated and obsessing on that makes the number surprising, but the fact the number points to a more out of whack site than usual. I don’t think it’s that difficult to understand. I would guess that Steve is thinking that you must obviously understand it; you write too well at times to be that hard-headed. So what purpose does one think somebody else has that doesn’t seem to get it?

    Advice: I got one post snipped, and stopped writing anything like what got snipped. It’s not too difficult; read your post, edit your post, keep snarks and woggles and jibes and blouncy out of it, and stay on topic. This one’s about a bad site. Answer civilly and without that tone to your “voice” and maybe the posts will stay. That’s all I can tell you.

    Steve’s free to edit or delete anything he likes, and it’s fine with me. It’s in the rules even if it wasn’t fine with me.


  135. Posted Aug 3, 2007 at 6:19 PM | Permalink

    I’m just about ready to delete this entire thread. So please people do the following:

    1- Remain civil
    2- Stop using crudity to describe
    3- Stick to topic

    TCO I ask that you just take a breather and consider your words more carefully, Mosher, you too. Everybody else, don’t encourage these ad hom’s by piling on.

  136. Posted Aug 3, 2007 at 7:02 PM | Permalink

    Comments are now closed for this thread

  137. Posted Aug 4, 2007 at 12:05 AM | Permalink

    Comments are open again

  138. Posted Aug 4, 2007 at 12:07 AM | Permalink

    RE 109, I have the Oke Paper on micosite variances posted on

    here is the link to the PDF

    “A technique to detect microclimatic inhomogeneities in historical records of screen-level air temperature”

    Runnalls K. E. and T. R. Oke

    JOURNAL OF CLIMATE 19 (6): 959-978 MAR 15 2006

  139. Kenneth Fritsch
    Posted Aug 4, 2007 at 12:17 PM | Permalink

    In today’s edition of the Chicago Tribune, Al Olson asks our local meteorologist, Tom Skilling, about a thermometer located on the North side of his house about 8 feet above the ground reading 8 to 10 degrees (F not C, I assume) too high.

    Tom replies by noting that The National Weather Service guidelines for thermometer exposure are stringent and that failure to adhere to them will result in erroneous readings. He says “Unfortunately, the criteria are practically impossible to satisfy in an urban setting”. He relates that “A thermometer (or its sensor) should be located over grass in a white, ventilated shelter 4-6 feet off the ground, at least 100 feet from all paved surfaces and at least 500 feet from any building”. He suggests to Olson that the high day time readings could result from the thermometer’s exposure to heat radiated from a nearby sunlit surface or from heat released from an air conditioner.

  140. Jim Edwards
    Posted Aug 4, 2007 at 12:42 PM | Permalink

    #122, Mark W:
    Well, I have to assume something, and ANY competent HVAC installer would size for load, not power available. When we had to install units beyond what the electrical service would supply, we had an electrician come out to upgrade the service.

    I agree with you that what goes on in the country is not necessarily what goes on in the city, but this is a federal gov’t installation – I expect they could get the power upgraded if they wanted.

    On the other hand, A/C units are often undersized – not b/c any competent HVAC installer would ever think about putting in undersized units, but b/c internal thermal loads change over the years. People add coffee pots and copy machines, they hire a new employee and move from a 19″ TV to a 60″ plasma screen. Some of these units could easily be undersized, but we have no data that any are or are even operational.

    #125, Reid Simpson
    Will the duty cycle matter for a single Tmax measurement ? Probably not, If the algorithm allows a single 10 sec reading to represent Tmax for the whole day. Even on cool Spring days the A/C will be on for 45 sec or so before cycling off. And the danger here, of course, is that w/o more detailed records of the time series, it will be impossible to resolve the cycling of the A/C to look for an effect in the temp signal. [Unless one shuts all of the A/C units down for a week or so, periodically, and records the fact in notes for the site, so a step function could be sought. Call it macro-cycling.]

    #131: Sam Urbanito,
    An A/C unit draws in ambient air, often mixed with a small amount of condenser discharge air that circles around in a vortex. It can’t bring in air cooler than ambient into the area unless there’s a nearby source of sub-ambient air. It’s possible that a very large unit that pulls in a lot of ambient air could ‘flush out’ the warming effects of concrete, say, but I doubt it. If possible, I wouldn’t call that a “spurious cooling trend.”

    I agree with your general take that A/C is likely to have some effect, but we don’t have to quantify the contamination

    I also agree with TCO’s prior comment that small A/C units don’t appear to be covered by the standards listed by Steve M. in #5, above. Even though A/C isn’t a large industrial heat source, it shouldn’t be near the measurement station. Perhaps the standards need to be updated, now that everybody has air conditioning.

  141. TCO
    Posted Aug 4, 2007 at 1:11 PM | Permalink

    Jim, good point in your last sentence. Ken, should we go by the standard or by those numbers from an interview? And are there some calibration studies that back up the 500 foot distance and the like?

  142. BarryW
    Posted Aug 4, 2007 at 4:00 PM | Permalink

    If the sites met the other siting requirements the whole issue of AC units would probably be moot.

  143. Dodgy Geezer
    Posted Aug 6, 2007 at 8:24 AM | Permalink

    There has been some discussion about the effects air-conditioning outputs will have on nearby temperature stations.

    I do not know if any of you have used the Computational Fluid Dynamics modelling tool FLUENT? It is a common student package in UK universities. It will simulate many aspects of airflow in complex conditions, including temperatures.

    I suspect it would be easy for an engineering undergraduate to model the Happy Ranch buildings and run the package to find out where the air goes. It sounds like a typical project. Pehaps you could find a tame student and try it?

  144. SteveSadlov
    Posted Aug 6, 2007 at 2:53 PM | Permalink

    RE: #144 – I would think that Flo Therm (commercial off the shelf thermal management design software) could also be adapted to this.

  145. SteveSadlov
    Posted Aug 6, 2007 at 2:57 PM | Permalink

    However, Flo Vent may be even better:

    (I had not looked at that site in a long time …. they now have S/W that seems perfectly tailored for evaluating microsite issues)

  146. Sam Urbinto
    Posted Aug 6, 2007 at 3:09 PM | Permalink

    Jim, 141, That’s the point, two-fold: We don’t have to quantify it and even if we did, we probably couldn’t without extensive testing of air flow and temperature over time. On the other explanation, that’s also pretty much the same point; we don’t know. Since the AC is affecting airflow in some way that’s the idea (bad siting), we have no idea of the specifics of how the airflow and temperature effects are in play, nor do we need to.

    But I think we’re losing the object of this. The issue here is just because lights=0, that doesn’t mean the site is really “rural” or that the site doesn’t suffer from the same contamination issues that the urban sites do.

    In addition, the purpose of this all is to gather information on sites meeting standards or not, so in the future we can audit the trends between the sites that meet standards and those that don’t, and how the adjustments in and between sites affect the whole. Of course, we need all or most of the stations categorized in order to do that. So what one station is doing in any particular way is not really important; we don’t want exact detailed information, just overall trends.

    I think trying to figure out the exact effect at any one place is a distraction and a waste of time. Just knowing it doesn’t meet the standards and is probably contaminated “in some way” is enough.

%d bloggers like this: