Berkeley Earth, Very Rural and Not

Jet exhaust as climate forcing

Guest Post by Willis Eschenbach

The good folks over at the Berkeley Earth Surface Temperature project have published their paper about urban heat islands. It’s called “Influence of Urban Heating on the Global Temperature Land Average using Rural Sites Identified from MODIS Classifications”, by Wickham et al, hereinafter W2013.

They find no urban heat island effect, saying at the close of the Abstract:

Time series of the Earth’s average land temperature are estimated using the Berkeley Earth methodology applied to the full dataset and the rural subset; the difference of these is consistent with no urban heating effect over the period 1950 to 2010, with a slope of -0.10 ± 0.24/100yr (95% confidence).

Having read the paper, I can’t say I’m surprised that they find no difference.

The W2013 authors have used the MODIS 500 metre dataset to identify urban areas. It’s a dataset that uses the MODIS satellite data to classify the planet into land-cover classes, including urban. The W2013 study uses the MODIS data to divide all the world’s temperature records into “very rural” and “not very rural”. Then they show there is little difference in the trends between the two groups.

GEhenderson field 2003 grid 100%Figure 1. Actual MODIS grid overlaid on Henderson Field, Guadalcanal, Solomon Islands. Picture is from 2003, prior to recent development. All gridcells would be classified by MODIS as not built-up, since less than 50% of the area of any given gridcell is built-up area. The ocean is at the upper right. Click to embiggen.

Let me talk a bit about the W2013 study and the MODIS map, and demonstrate by example the reason why I didn’t expect their “very rural” split to have much discriminatory power regarding temperature trends. As an example, let me use one of my favorite airports, Henderson Field on Guadalcanal.

Here is the description from the W2013 study of how they identified the “very rural” stations.

The MOD500 map is available as a raster image, providing a binary classification (urban or not urban) for a global grid with pixels of size 15 arc-seconds.

Rather than compare urban sites to non-urban, thereby explicitly estimating UHI effects, we split sites into very-rural and not very-rural. We defined a site as “very-rural” if the MOD500 map showed no urban regions within one tenth of a degree in latitude or longitude of the site. This choice should minimize errors that occur from MODIS classifications in fringe areas. We expect these very-rural sites to be reasonably free from urban heating effects.

This led me to investigate the MODIS “MOD500″ map. The MODIS dataset is described in A new map of global urban extent from MODIS satellite data. Basically, it uses multi-spectrum radiation to discriminate between types of land cover, and to identify the signatures of various built-up areas..

Of immediate interest to this discussion is their definition of “urban areas”:

a. Urban areas

In both datasets, urban areas (coded class 13) are defined based on physical attributes: urban areas are places that are dominated by the built environment. The ‘built environment’ includes all non-vegetative, human-constructed elements, such as buildings, roads, runways, etc. (i.e. a mix of human-made surfaces and materials), and ‘dominated’ implies coverage greater than or equal to 50 percent of a given landscape unit (here, the pixel). Pixels that are predominantly vegetated (e.g. a park) are not considered urban, even though in terms of land use, they may function as urban space. Although ‘impervious surface’ is often used to characterize urban areas within the remote sensing literature, we prefer the more direct term ‘built environment’ because of uncertainty and scaling issues surrounding the impervious surface concept. Finally, we also define a minimum mapping unit: urban areas are contiguous patches of built-up land greater than 1 km^2. SOURCE

Now, each “pixel” that they mention above is one of the gridcells shown in Figure 1 above. Those are the actual gridcells used in the MODIS map. If more than half of a gridcell is “built environment” (houses, roads, runways, etc.) then the gridcell is counted as “built-up land”. However, there’s a final hurdle. You need to have five adjacent gridcells of built-up land to have those gridcells classed as part of an urban area in the MOD500 map.

This is kind of an odd definition. It means that any small hamlet is rural, since it won’t cover a square kilometre (about 250 acres). You have to have five contiguous built-up gridcells, which totals just over 1 km^2, for those gridcells to be classified as urban in the MOD500 map.

Figure 2 below shows how much of the land shown in Figure 1 could be built up, covered with houses and roads and parking lots, without anything being called “urban”.

GEhenderson field 2003 no contig 60%Figure 2. As in Figure 1, overlaid with theoretical possible buildup without any areas being classified as “urban” in the MODIS dataset. Gray areas are 100% built-up, white triangles show 40% built-up. A total of about 70% of the land is built-up. None of it is classified urban because nowhere are there are five contiguous built-up gridcells.

In any case, that’s the history. Figure 3 shows the situation today:

GEhenderson field 2013 gridFigure 3. Actual MODIS grid overlaid on current view of Henderson Field. Picture is from January 2013. All gridcells would still be classified by MODIS as not built-up, since less than 50% of the area of any gridcell is built-up area. The gridcells shown in red are the nearest to being classified as built up, as they all have about 40% of their area covered with roads, houses, and runways.

Again, in the MOD500 map, none of these gridcells would be classified as urban. Even if those four gridcells were more than 50% built-up, still none of them would be classified as urban because there aren’t five of them and they aren’t contiguous.

However, the fact that the MOD500 map classifies all of these gridcells as being rural makes no difference—the Stevenson Screen housing the thermometers is still in a horrible location. Here’s why, borrowed from Anthony’s post on the subject. This shows a ground level view of the station taken from the terminal. Note the white box of the Stevenson Screen housing the thermometer and other instruments in the green area just left of center, below the smoke visible in the distance.

Here comes a plane! Weather station visible to the left of the plane’s tail, right of the white mast.

Coming into the terminal…

Hey, park it over here!

Uh, oh, look where the jet exhaust is pointed:

Hmmm, a new high temperature today?

Back to normal.

Now, that’s the problem with the location of the Henderson Field weather station. There are lots of good places to locate the Stevenson Screen in the local area … and it’s in a bad place. This situation is mirrored in many airports and weather stations in general around the planet.

The difficulty I have with the approach of using the MOD500 map to distinguish urban from rural is simply stated:

Many of the siting problems have nothing to do with proximity to an urban area.

Instead, many of them have everything to do with proximity to jet planes, or to air conditioner exhaust, or to the back of a single house in a big field, or to being located over a patch of gravel.

And sadly, even with a map averaged on a 500 metre grid, there’s no way to determine those things.

And that’s why I didn’t expect they would find any difference … because their division into categories has little to do with the actual freedom of the station from human influences on the temperature. Urban vs Rural is not the issue. The real dichotomy is Well Sited vs Poorly Sited.

It is for this reason that I think that the “Urban Heat Island” or UHI is very poorly named. I’ve been agitating for a while to call it the LHI, for the “Local Heat Island”. It’s not essentially urban in nature. It doesn’t matter what’s causing the local heat island, whether it’s shelter from the wind as the trees grow up or proximity to a barbecue pit.

Nor does the local heat island have to be large. A thermometer sitting above a small patch of gravel will show a very different temperature response from one just a short distance away in a grassy field. The local heat island only needs to be big enough to contain the thermometer, one air conditioner exhaust is plenty, as is a jet exhaust …

The only approach that I see that has any hope of success is to painstakingly divide the stations, one by one, based on what is in their viewshed, and exactly how far are they from a variety of ground covers. The work done by Michel Leroy of METEOFrance in 2010 lays out one way to do this, as discussed here.

Because even with the outstanding resolution of the MODIS dataset, it still can’t tell us whether the siting of the weather station is up to snuff, or whether we’re just measuring the temperature of a jet engine at 50 metres … and that’s why I don’t find the results in the W2013 paper at all persuasive.

Best regards,

w.

PS—One curiosity. The published paper is a slightly polished version of their earlier pre-publication paper with the exact same title, available here. The curiosity is the re-ordering of the authors on the two title pages, viz:

W2013 author list

Internal politics?

I also note that my friend Steven Mosher is now listed as an author on the paper, my congratulations to him.

[UPDATE] Steven Mosher has pointed out that in the analysis, this particular station (Henderson Field) is classified as urban. I have no problem with that, as it is not a long ways from the capital and is likely classified urban by BEST (although likely not by MODIS) on that basis for their study.

Nor was the classification of that particular station my point, which I quote here from above:

And that’s why I didn’t expect they would find any difference … because their division into categories has little to do with the actual freedom of the station from human influences on the temperature. Urban vs Rural is not the issue. The real dichotomy is Well Sited vs Poorly Sited.

It is for this reason that I think that the “Urban Heat Island” or UHI is very poorly named. I’ve been agitating for a while to call it the LHI, for the “Local Heat Island”. It’s not essentially urban in nature. It doesn’t matter what’s causing the local heat island, whether it’s shelter from the wind as the trees grow up or proximity to a barbecue pit.

Steven also adds the following comment:

My main concern is that people will think the article is about urban/rural (see the title … very rural, not) when the text is about “siting”.

Also, folks may get the idea that airports can never be CRN1 [a measure of station quality] … something which Fall 2011 doesn’t support, as many of CRN1 and CRN2 are at airports.

A clarification might help prevent future misunderstandings and gotchas.

My thanks to Steven for calling that classification to my attention, and also for his clarification.

About these ads
This entry was posted in Land use land cover change, Surfacestations paper, UHI and tagged , , , , , , , . Bookmark the permalink.

257 Responses to Berkeley Earth, Very Rural and Not

  1. seanbrady says:

    How many total thermometers in the data set? How about a crowd sourced division based on Google Earth into LHI or no-LHI? Then run the split through their same methodology to see if there is a difference?

  2. Anthony Watts says:

    Thanks Willis. As I point out here, http://wattsupwiththat.com/2012/07/29/press-release-2/ the resolution of the study we did is down to 10 meters. Leroy 2010 sees 100 meters as the limit to the effects he observes relating to siting. Like Marcott et al, this is a case of a low resolution sample missing the “spikes”.

    Might be fun to see if we can link high temperature spikes at Henderson Field to flight arrival/departure times. The data must be there somewhere.

  3. UHI exists. If the BEST team can’t find UHI, they should keep trying. And quit pretending UHI doesn’t exist.

  4. Mike Bromley the Canucklehead in Switzerland says:

    I can’t believe how chintzy science has gotten. Of everything else they trumpet about, the real world is not one of them.

  5. JunkPsychology says:

    Why would a knowledgeable reviewer recommend accepting a study that used this method?

  6. Local Heat Island! I love it! willis’s eagle flies ever higher!

  7. Resourceguy says:

    If oceans cover 71 percent of the planet and we have more efficient temperature monitoring systems on the ocean and overhead in satellites, I’m more interested in the regional and hemispheric differences that present themselves between these major monitoring systems of land vs. ocean over time. Slicing and dicing land-based data sets does not cut it for this important issue.

  8. HankHenry says:

    WOW this is news! No such thing as UHI? Just about every weatherman in the country from time to time mentions UHI in their weather reports. But hey, it’s science so I guess it’s undeniable.

  9. David Schofield says:

    Just watched the BBC weather man; “overnight it will be 2 degrees in the towns and below freezing in the countryside”. Obviously didn’t get the memo.

  10. John Tillman says:

    But, as usual, it’s even worse. Hansen recognized UHIs, but his long secret algorithm to adjust for them always made them warmer.

  11. Hey, park it over here!
    That’s Hilarious!!!

  12. No mention of Henderson Field is complete without the original history of this Japanese air base, captured by the Marines and scene of fierce fighting in a tropical jungle with 200 inches of rain per year. This 2500 square mile volcanic and raised coral reefs dot in the Pacific is covered with impenetrable jungles, hulks of war machines and ghosts of heroes. The land, sea and air battles included the last sinking of a war vessel by ramming, when a New Zealand destroyer rammed a Japanese submarine. This battle was described in “Guadalcanal” by Richard Frank. There were over 200,000 documented cases of Japanese cannibalism documented in WW II war crimes trials, which were ordered sealed for ‘national security’ reasons, per “Flyboys”.

  13. Nick in Vancouver says:

    Realistic? Not very realistic.

  14. Michael Cohen says:

    I think it would be helpful for casual readers to emphasize that the paper is reporting no effect on the global land surface temperature trend from urban heating, not that there is no urban heating.

  15. Chris @NJSnowFan says:

    Good read..

    Pavement pavement pavement, Pavement does not emit water vapor like trees except after ir rains or during. Proven fact one full grown oak tree emits 7 tons of water vapor every day through it’s leaves. Water vapor cools the air and that is a proven fact. Pavement warm the air to dry and hot.

    Urban heat islands are real.

  16. Henry Bowman says:

    A few years ago, this kid (with help from his father, I believe) showed reasonably well the difference between rural sites and urban sites in the U.S. I’m sure there are some problems with his methodology, but surely someone could do a good job, inasmuch as the data are out there:

  17. JeffC says:

    so a thermometer located in center of a patch of concrete in a 100 acre field would be considered rural … right …

    did they even bother to run their model and then actually compare a “rural” area to what they could see with their eyes … I would imagine that almost every airport is considered “rural” …

  18. dorsai123 says:

    [snip]

  19. Gene Selkov says:

    Anthony says:

    > Might be fun to see if we can link high temperature spikes at Henderson Field to flight arrival/departure times. The data must be there somewhere.

    I will be very interested to find that out, although I suspect a thermometer needs to be very sensitive and fast to pick up a gust of warm jet exhaust. I can more easily imagine runways having a substantial effect on thermometers located near them than the jet traffic on them. You can estimate the magnitude of the effect using this paper outlining the temperature contours of jet exhaust at different throttle positions:

    http://www.boeing.com/commercial/airports/acaps/7471sec6.pdf

    You need to combine that with the estimate of the “duty cycle” — how many times an hour is there a jet on the runway aiming its exhaust right at the weather station, and for how long.

  20. Reg Nelson says:

    Muller would have saved himself a lot of wasted time and effort if would have just read this first:

    http://wattsupwiththat.com/2013/01/20/noaa-establishes-a-fact-about-station-siting-nighttime-temperatures-are-indeed-higher-closer-to-the-laboratory/

  21. MAK says:

    Comparing very rural to very urban should not have any difference in the trend at all – assuming that both stations do not change their classification. If a site is already UHI-corrupted, comparing it’s trend to rural is stupid. You will not find anything.

    The whole thing changes if a rural site gradually changes to urban site. This kind of change you don’t find using MOD500 dataset for site classifications.

    Using MOD500 even to try to sort this out clearly tells us these scientist don’t know a thing what they are trying to accomplish. Or maybe they do – but then they really don’t want to find UHI effect.

  22. Matthew W says:

    OK, so the jet exhaust peaks out the temp.
    Can’t we just fix that with smoothing?
    /sarc

  23. Curious George says:

    The BEST methodology simply does not work. It is a commendable attempt at an objectivity, but unfortunately it fails.

  24. Rud Istvan says:

    Willis, excellent post. You actually point to a much bigger problem. Rather than define an issue (local heat island is excellent, even if there are more of them in urban areas) and then go actually get real data in terms of the definition, it seems more and more that what is done is to see what happens to be lying around, then repurpose it by massaging to get a paper published that is inherently shaky.
    The correct way is what Anthony’s brilliant crowd sourced station siting database did. Actually go check the sites for the parameters needing to be measured. It is stunning that NOAA and gang still haven’t gotten out of their taxpayer funded offices to do the same. It must be easier to sit in the office drinking coffee and dreaming up more undocumented homogenizations to wish the problem away.
    Let’s hope Anthony’s paper gets published soon.
    Regards

  25. Bob says:

    Steve Mosher, I guess we shamed you into becoming an author. Congratulations/

  26. Brad says:

    Traditionally, the first author did most of the work, the last author had the funding or supervised the work, and the rest can be from most to least important, or simply alphabetical.

  27. Ford Prefect says:

    Edmonton Alberta – you know the most evil place on the planet – is a good place to compare “The Influence of Urban Heating on the Global Temperature Land Average using Rural Sites Identified from MODIS Classifications”, Wow you just can’t make a name like that up.

    Edmonton has two airports one in the middle of the city and one out side the city. The overnight temp for both airports is given and one is usually cooler. If you were to bet which airport is cooler the one who picked the country would soon have all the money.

  28. Duster says:

    I have some concerns about “rural” vs. “wild” that Anthony can directly observe around Chico, especially to the west where the Sacramento Valley is heavily agricultural. There is little justification for thinking that merely because land is “rural” with a relatively scant human population, that there is little anthropic effect on temperatures. Agricultural land is just as extensively altered as urban land. While it won’t have the heat discharge from air conditioners, industry and transportation that a city does, most agricultural modifications increase insolation, and they do it on a far larger scale. Few if any sites in the USHCN are really in wild lands, possibly a few linked to fire look outs, otherwise they are historically biased to to places that were important to the Department of Commerce’s purposes. That is, places where people lived and worked.

    The CRN in general seems to be better sited, but even there occasional sites like Whiskeytown in California are close to massive human alterations. The Whiskeytown CRN station is upslope from Whiskeytown reservoir, a large, artificial reservoir on Clear Creek. Being directly upslope from the lake means that warmer water at night should moderate cooling around the lake during the night.

  29. The Village Idiot says:

    This is exactly why we can’t trust the instrumental record. It’s just wrong. Tony said: “I don’t “deny” the instrumental record” . Just can’t see the difference between don’t trust and deny.

    http://wattsupwiththat.com/2013/03/18/monday-mirthiness-watch-the-genesis-and-retraction-of-of-a-smear/

    Although any record showing cooling gets the thumbs up :)

  30. Bloke down the pub says:

    Perhaps once the USCRN has got a couple more years under it’s belt, they’ll no longer be able to ignore the differences to the thermometers that are stuck up an airplanes tailpipe. That’s created a mental image that only Josh could do credit to.

  31. joshv says:

    I think both you and BEST are missing the point somewhat. How are your poorly sited stations introducing a trend? If it’s hotter near a runway, then of course that station will run hot, but it won’t show an artificial trend, on decadal time scales it will show an constant offset from a better site location nearby.

    The UHI affect as I understand it is related to *changes* in urbanization, and I think this point is being lost on most everyone. Find locations were urbanization has been static, and compare those to ones where urbanization has changed. It doesn’t really matter if they are poorly sited or not – if the local conditions have remained static, that site should be a good control for a site that’s nearby where the conditions have not been static.

  32. DayHay says:

    Exactly what question were the authors trying to answer? Whether MODIS can detect UHI? Then they have proved it cannot. That is important
    Perhaps if they started with “Does siting have any affect on homogenized temperature data trends?”
    I guess none of the authors ride motorcycles. Take a ride at dusk and feet the real temperatures as you exit the city. Even as little as a 50 foot elevation change is very noticeable. But at least we know the world land temp to 0.00001 deg C……..ugh.

  33. Steven Mosher says:

    sadly willis, Zeke and I did a sensitivity test where we excluded airports for just this reason

    So, no fruit cup for you

  34. Steven Mosher says:

    MAK says:
    April 4, 2013 at 10:10 am
    Comparing very rural to very urban should not have any difference in the trend at all – assuming that both stations do not change their classification. If a site is already UHI-corrupted, comparing it’s trend to rural is stupid. You will not find anything.

    The whole thing changes if a rural site gradually changes to urban site. This kind of change you don’t find using MOD500 dataset for site classifications.

    ##########################

    Thats why I checked the changes in population from 1900 to present.

    the rural sites were virtually unchanged. the urban sites had large changes in population density.

  35. George V says:

    I saw an example of the JEHI (Jet Exhaust Heat Island) effect a couple of winters ago while on vacation in Traverse City Michigan. One morning I was checking the forecast temp for the day on the NOAA web site and checked the history file to see how cold it was the night before (because it seemed pretty durn cold). Sure ’nuff, it was cold, dropping to single digits around midnight from what recall. But the next hourly reading (some minutes after 1AM – can’t recall exactly) was 25 deg. F. The next reading was back in single digits. I suspect some bizjet or turboprop had taken off since there are no commercial flights at that time.

  36. Curt says:

    I believe that both “Urban Heat Island” and “Local Heat Island” (or “micro-ste”) effects must be taken into account, and separately. A station in New York’s Central Park might not have significant LHI effects, but would have large UHI effects. A rural site in a parking lot would have LHI but not UHI.

    But if we’re analyzing trends, I agree that we must look primarily at changes in the environment around the station, and evaluate them carefully. For a long time now, many scientists have believed that the magnitude of the UHI effect is roughly a logarthmic function of population, so a rural site where local population increases from 1000 to 2000 would have the same increase as an urban site with an increase from 1,000,000 to 2,000,000.

    Airport sites are particularly problematic, as the airports are often placed initially in rural areas where land is cheap and there are few neighbors to annoy. But there is an almost inevitable development of supporting infrastructure around the airport over the years — but since no one lives in this infrastructure, population estimates of UHI don’t catch this (though MODIS-type estimates could).

  37. Steven Mosher says:

    ‘Now, each “pixel” that they mention above is one of the gridcells shown in Figure 1 above. Those are the actual gridcells used in the MODIS map. If more than half of a gridcell is “built environment” (houses, roads, runways, etc.) then the gridcell is counted as “built-up land”. However, there’s a final hurdle. You need to have five adjacent gridcells of built-up land to have those gridcells classed as part of an urban area in the MOD500 map.”

    Actually not. 4 cells will do. If you like you can double check the accuracy of the urban classification by MODIS Urban by using the NLCD 30 meter impervious surface data.
    you wont find anything. This issue was raised early on in web comments, so of course
    I had to go get 30meter data to check.

    it looks like this

    http://stevemosher.files.wordpress.com/2012/10/nlcdisa2967178.png

    http://stevemosher.wordpress.com/2012/10/08/sample-suhi/

    http://stevemosher.wordpress.com/2012/10/02/city-size-and-suhi/

    http://stevemosher.wordpress.com/2012/09/26/rastervis-to-the-rescue/

    http://stevemosher.wordpress.com/2012/11/25/modis-r-package-tutorial/

    or you can double check with the ORIGINAL SOURCE data for the modis urban dataset which allows you to use a 500 meter rule as opposed to the adjacent cell rule. The orignal datasets are done on an 8 day basis. huge amount of data. But you can actually go check the classification using: the 500K urban dataset ( with the 1km rule) or the 500 meter dataset without the 1km rule. bet you didnt check that. and then you can check the sensitivity of the classification at SUB 500 meter resolution using 30 meter data. its a lot of work, but go ahead.

    I used some of that here

    http://stevemosher.wordpress.com/2012/10/11/pilot-study-small-town-land-surface-temperature/

    Oh, you can also check the data using 300 meter ESA data

    http://geonetwork.grid.unep.ch/geonetwork/srv/en/metadata.show?uuid=geodata_%202055

    If you want to know how to eliminate airports, see my metadata package.

    Doesnt make a difference.

  38. Steven Mosher says:

    sunshinehours1 says:
    April 4, 2013 at 9:20 am
    UHI exists. If the BEST team can’t find UHI, they should keep trying. And quit pretending UHI doesn’t exist.

    #######################

    Dude I’m still waiting for somebody to accept my challenge.

    1. Define Urban or rural ex ante in a way that is objectively measureable.
    2. i will divide stations into urban and rural per your definition.
    3. i will compute the difference.

    A cookie for anyone who can find the signal.

    So, there is the challenge.

  39. rogerknights says:

    joshv says:
    April 4, 2013 at 10:42 am

    I think both you and BEST are missing the point somewhat. How are your poorly sited stations introducing a trend? If it’s hotter near a runway, then of course that station will run hot, but it won’t show an artificial trend, on decadal time scales it will show an constant offset from a better site location nearby.

    Things that will increase the trend at an airport:

    If larger planes are used that consume more fuel.
    If jet planes are used, which blow their heat sideways.
    If air traffic increases.
    If the airstrip was originally dirt or grass.
    If the airport was originally rural but the area around it becomes urbanized.

  40. Nik Marshall-Blank says:

    But we cater for this. We take the median for each day and discard all outliers greater than 25% of the median. This means in a single day out of 100 observations we have only 8 valid readings. Mmm… same times as the planes landed. Curious.

  41. Bob Kutz says:

    While I don’t disagree with your analysis, I do hold out that there is a UHI effect. There certainly is an LHI effect as well, but I see the UHI effect every single day.

    My house is in town. I work in an industrial park outside of another town. The distance between is some 20 odd miles. The outdoor thermometer on my house agrees to my truck’s thermometer and the local cable provider’s ‘current temp.’ from their offices about 8 blocks from my house.

    When I jump in the truck and head to work, it drops between 4 and 7 degrees F by the time I am out of town and near the local airport. Hot or cold; the temp drops when I go outside of town.

    If I take the other way to work, it’ll drop another 4 to 6 when the road drops down to the river bottom for several miles. That is just local temperature variation. Being in the bottom of a hole certainly causes that.

    The industrial park where I work is actually a WWII air corps training base. The runway is still in use as a local airport. This is different from the airport I pass on along the way.

    The temperature here usually agrees between my truck, the ‘official’ temperature at the airport, our lab’s outdoor ambient temp. reading, and the temperature at the meteorology lab at the local community college, just across the industrial park from my office. Sometimes not. Sometimes our lab is effected by the rather large burner we run. Sometimes the airport thermometer is several degrees higher than the other two. My truck and the aeronautics/meteorology lab thermometers don’t disagree by more than a degree.

    Now if I run into town during the day, another 6 miles to the south, the temperature is always 2 to 5 degrees warmer than out here at the industrial park. Always. And this town is down on the river. ‘In a hole’ so to speak. So the coolness of the indigenous riverbottom landscape is more than offset by the UHI effect of 50,000 people.

    So I can point to two very definable ‘UHI’ here in southeastern Iowa where neither community is over 50,000 people.

    I don’t know how this plays out in terms of data, but there is little doubt that the average temperature for these two, geographically dis-similar towns that are in relatively close proximity.

    Then there is the large town about 50 miles to the northwest. The UHI effect there can be seen just by watching the local TV weather broadcast. The host occasionally points out that it can be close to 10 degrees below zero in most of the region before the thermometer at the downtown TV broadcast station dips below zero. Going into that town on a hot day proves the point every single time. It can be hot and humid out in the cornfields. But you get into the heart of that metro area and it is scorching. It can be more than 10 degrees difference between just 15 miles to the southeast vs. right down town. And it is worse at night. All that concrete holds the heat about as well as any heat sink you might intentionally devise.

    As I said; I don’t know how this plays out in the climate data, but there is a UHI effect, distinctive from the LHI effect, which can be as local as the Stephens screen at an airport. The UHI effect covers most of the downtown and outlying areas in two small towns (say five or six square miles) and a fairly broad swath of territory in a large central Iowa city (probably approaching 100 square miles). Looking at google earth it is doubtful it is more than 10 sq. miles by the ’50% in 5 squares’ method.

  42. Don B says:

    I have a car thermometer. Two or three times each week before sunrise I drive from town to the country and return within an hour. It is cooler away from town.

    All of those authors, and none of them have a car thermometer.

  43. Steven Mosher says:

    Bob says:
    April 4, 2013 at 10:16 am
    Steve Mosher, I guess we shamed you into becoming an author. Congratulations/

    #################

    Not. I have never asked and will never ask to be a co author. i do the work I do and don’t care very much about were the credit is given. acknowledgments or byline? phhht. I do what i do.

  44. Alec Rawls says:

    Wickham et al:

    We expect these very-rural sites to be reasonably free from urban heating effects.

    It is not the absolute level of UHI effects that matter but the CHANGE in UHI effects, and as any economist will tell you, the law of diminishing returns really is something close to a law of nature: it is a very widespread phenomenon that the marginal effect of the first increments to any input will typically be larger than the marginal effect of further increments.

    So the place where the weather station is located adds a bit of paving and some heating and air conditioning over the decades. The marginal effect on the very rural station is much greater than on the less rural and urban stations. The expectation going in could even be that very rural stations tend to experience greater UHI than less rural stations, yet Wickham treats the very rural stations as if they experience no UHI at all.

    As an earlier commenter noted, it would be necessary to look at the transition of stations from more rural to less rural and compare to stations that changed less. What is the correlation between rate of transition to urban and the speed of warming over the 20th century?

  45. Mike Ballantine says:

    Data mining in corrupted data yields corrupted results. Willis’ example highlights a major problem with the instrumental record. Temperature measuring in the middle of an airport has only ONE valid and very important use. It gives the pilots information they need to set safe, efficient take off and landing speeds. Any other use is like using pliers to hammer nails.
    IMHO, ALL airport sited temperature stations should be excluded from the climate data sets.

  46. Nik Marshall-Blank says:

    @Bob Kutz – Well there’s a whole new area to investigate now. Do these areas where the readings are taken correctly represent the temperatures are they are expected to cover?

  47. davidmhoffer says:

    Steven Mosher;
    Dude I’m still waiting for somebody to accept my challenge.

    1. Define Urban or rural ex ante in a way that is objectively measureable.
    2. i will divide stations into urban and rural per your definition.
    3. i will compute the difference.

    A cookie for anyone who can find the signal.

    So, there is the challenge.
    >>>>>>>>>>>>>>>>>>

    Sure. Define a methodology that doesn’t address the problem and in fact reinforces it instead and then bravely bet a cookie on it.

  48. ralfellis says:

    Same problem with the Central England Temperature series. This consists of three locations, averaged. But one of those locations is on Manchester airport, with is a major international airport with arivals every three minutes. And the Lat-Long I have puts the met station just across the grass from the engine run-up bays, in the middle of all the taxiways.

  49. Robuk says:

    Henry Bowman says:
    April 4, 2013 at 9:56 am

    A few years ago, this kid (with help from his father, I believe) showed reasonably well the difference between rural sites and urban sites in the U.S.

    But that would mean venturing outside which could be dangerous, cars, cycles etc.

  50. Steven Mosher says:

    rogerknights says:
    April 4, 2013 at 11:18 am
    joshv says:
    April 4, 2013 at 10:42 am

    I think both you and BEST are missing the point somewhat. How are your poorly sited stations introducing a trend? If it’s hotter near a runway, then of course that station will run hot, but it won’t show an artificial trend, on decadal time scales it will show an constant offset from a better site location nearby.

    Things that will increase the trend at an airport:

    If larger planes are used that consume more fuel.
    If jet planes are used, which blow their heat sideways.
    If air traffic increases.
    If the airstrip was originally dirt or grass.
    If the airport was originally rural but the area around it becomes urbanized

    ####################################

    The one thing people miss is that airports in some areas actually are cooler and have cooling trends. why?

    UHI is a function of these variables:

    Incoming solar
    outgoing LW
    Storage capacity
    Anthro heat.
    Turbulance

    For incoming solar airports are great because the building tend to be low and you dont get a radiative canyon ( see Oke for a description )
    For outgoing LW they are great because of a sky view factor of 1. (see OKE)

    Storage capacity: the is basically the heat that gets stored in concrete/asphalt etc.

    This can impact Tmin. BUT you have to look at the wind speed. Airports are built with long fetches. See the fetch on willis’s airport. At 7m/s UHI disappears. if the surface roughness is low this happens at lower wind speeds

    Anthro heat. yup the heat from the jets.
    But this delta T has to occur at just the right time. It has to happen around Tmax.

  51. Justthinkin says:

    “sunshinehours1 says:

    April 4, 2013 at 9:20 am

    UHI exists. If the BEST team can’t find UHI, they should keep trying. And quit pretending UHI doesn’t exist.”
    sunshinehours1 and others here. NOBODY is claiming that UHI do NOT exist. We just want you and others to show your homework as to how this is affecting the GLOBAL temperature and/or you new name,Climate Change. BTW…anybody with two active brain cells and not in it for fame/fortune do not claim there is no climate change,just what is causing it. Its been changing for 4 billion years,and will carry on doing so.

  52. Steven Mosher says:

    Mike Ballantine says:
    April 4, 2013 at 11:24 am
    Data mining in corrupted data yields corrupted results. Willis’ example highlights a major problem with the instrumental record. Temperature measuring in the middle of an airport has only ONE valid and very important use. It gives the pilots information they need to set safe, efficient take off and landing speeds. Any other use is like using pliers to hammer nails.
    IMHO, ALL airport sited temperature stations should be excluded from the climate data sets.

    ########################

    thats a good point. i did that. guess what? no difference. some airports cool other warm.
    remove them all and the global trend doesnt change.

    Also, you can actually compare CRN stations 9 the one Anthony likes ) to nearby airports on an hour by hour basis.

    guess what? yup. no difference.

  53. Steven Mosher says:

    davidmhoffer says:
    April 4, 2013 at 11:29 am
    Steven Mosher;
    Dude I’m still waiting for somebody to accept my challenge.

    1. Define Urban or rural ex ante in a way that is objectively measureable.
    2. i will divide stations into urban and rural per your definition.
    3. i will compute the difference.

    A cookie for anyone who can find the signal.

    So, there is the challenge.
    >>>>>>>>>>>>>>>>>>

    Sure. Define a methodology that doesn’t address the problem and in fact reinforces it instead and then bravely bet a cookie on it.

    ####################

    Ah, so your opinion about UHI is not based on objective criteria. Cool.

  54. Steven Mosher says:

    Bob Kutz says:
    April 4, 2013 at 11:19 am
    While I don’t disagree with your analysis, I do hold out that there is a UHI effect. There certainly is an LHI effect as well, but I see the UHI effect every single day.

    #############################

    yes there is UHI effect. That has never been the question.

    the question is does that effect BIAS the record.

    The UHI effect is not homogenous in space or time. in fact, some rural sites are warmer than nearby urban sites ( differences in surface emmissitivity and albeo drive that )

    Example: here are the land temperatures for different land classes. is urban the warmest?

    NOPE! As Oke remarked long ago the urban – rural DIFFERENCE depends more on the rural than it does on the urban. So, sometimes the difference is big, sometimes its small, sometimes its NEGATIVE.

    See temperatures by land class.

    http://stevemosher.files.wordpress.com/2012/11/daylst-fig3.png

  55. Stephen Skinner says:

    “the Stevenson Screen housing the thermometers is still in a horrible location.”
    I was beaten to it by Mike Ballantine above, but the thermometers are in exactly the right place for their purpose and must be to provide the most accurate weather info for the pilots. In addition I don’t think the jet exhaust will have any affect as the wind sock in the background shows the wind blowing to the right while the jet coming onto the ramp is likely to be idling.
    As Henderson Field doesn’t appear to be very busy it should be easy to see any jet exhaust affect as a spike in the temps at the times of the flights.
    I like the idea of LHI as opposed to UHI as the latter implies that urban is the only place a temp can be above ‘normal’, but so called rural such as drained farmland can be warm as well. Ask any glider pilot.

  56. kim says:

    There are some people whose daughter’s papers I will no longer read.
    =================

  57. Don B says:

    Willis, Roy Spencer did some work with UHI and population density. Whatever happened with that research?

    http://www.drroyspencer.com/wp-content/uploads/ISH-UHI-warming-global-and-US-non-US.jpg

  58. Steven Mosher says:

    for the chart above if you thought land class 7 was urban you are wrong.

    read the rest here and see how land class and temperature varies.
    to understand WHY you have to understand albedo and emissitivity. Bottom line: urban aint the hottest. its warmer than most rural surroundings but you have to control for the type of rural aurrounding to understand why in some cases its high and in some cases its low and in some cases its negative. ON AVERAGE with the actual stations we have it turns out to be below the noise floor.

    http://stevemosher.wordpress.com/2012/11/10/terrain-effects-on-suhi-estimates/

  59. Willis Eschenbach says:

    joshv says:
    April 4, 2013 at 10:42 am

    I think both you and BEST are missing the point somewhat. How are your poorly sited stations introducing a trend? If it’s hotter near a runway, then of course that station will run hot, but it won’t show an artificial trend, on decadal time scales it will show an constant offset from a better site location nearby.

    Thanks, Josh. Poorly sited stations do not necessarily introduce a trend. But it it quite common. For example, at Henderson Field they lengthened the runway in the 1980s to bring in the larger jets … which increased both the pavement within the viewshed and the number and size of the planes landing and blowing hot exhaust onto the thermometer. Since development is much more common than de-devlopment, most changes increase the surrounding temperature.

    w.

  60. davidmhoffer says:

    Steven Mosher;
    Ah, so your opinion about UHI is not based on objective criteria. Cool.
    >>>>>>>>>>>>>>>>

    No, my opinion is that your criteria is not objective. Not cool.

  61. Steven Mosher says:

    Don B says:
    April 4, 2013 at 11:50 am
    Willis, Roy Spencer did some work with UHI and population density. Whatever happened with that research?

    http://www.drroyspencer.com/wp-content/uploads/ISH-UHI-warming-global-and-US-non-US.jpg

    ####################

    eh, his work was not replicatable. part of his issue is using an innaccurate population dataset. The dataset he uses actually clusters people together. A much better source is TIGER census data. That allows you to work with actual census tract counts rather than data that is gridded by an algorithm which is what Roy used.

    Finally, population is a poor metric and Oke who first proposed it, later abandoned it because it was not dimensionally correct, so now we look at the impervious surface area.

  62. Mosher, if you are going to write papers about something that exists and then whine about your failure to find it and offer cookies to people so they try and do your job, you really are a big failure.

    Why not find places with UHI and write a paper about the difference in temperatures?

    Concluding “consistent with no urban heating effect” is just not believable.

    Concluding “we know UHI exists but we couldn’t find it” is believable.

  63. Steven Mosher says:

    ‘Airport sites are particularly problematic, as the airports are often placed initially in rural areas where land is cheap and there are few neighbors to annoy. But there is an almost inevitable development of supporting infrastructure around the airport over the years — but since no one lives in this infrastructure, population estimates of UHI don’t catch this (though MODIS-type estimates could).”

    That is one of the reasons why I tell people that they cannot use population as a metric.. or rather why certain datasets ( like those used by roy spencer0 are not good solutions.

    there are two types of population datasets: ambient and residential.

  64. Willis Eschenbach says:

    Steven Mosher says:
    April 4, 2013 at 10:48 am

    sadly willis, Zeke and I did a sensitivity test where we excluded airports for just this reason

    So, no fruit cup for you

    Even more sadly, Steven, there’s not one word in the study about your magic airport test, or about your claimed study of the question, so once again it’s just another jolly Mosher anecdote … have you given up entirely on providing citations for your claims, or is this just a passing phase and your unscientific lack of citations will “self-correct” about when the rest of climate science self-corrects?

    In any case, you’ve taken a map (MOD500) where up to three quarters of the land surface can be parking lots and it’s still identified as rural and you are using it to try to tell good sites from bad … how about you deal with the underlying questions about scale and whether the MOD500 map can show us bad sites, and not with just the one example I picked?

    My point had nothing to do with airports. It had to do with the foolishness of trying to divide the sites into classes based on the MOD500 map. You say you and Zeke have divided the sites based on airports vs. not airports, but so what? That’s just as bad as dividing them based on the MOD500.

    The only thing worth dividing them on, as I said above, are the actual characteristics of the actual site. Not whether it’s rural or urban. Not whether it’s on the MOD500 map or not. And not whether it’s an airport or not. All that you and Zeke proved is what I’m saying—you can’t divide the sites on unrelated or weakly related lines and see a difference. You need to divide them on the basis of the actual sites themselves.

    As to the fruit cup, well, that’s still in question. I may not win it, but I doubt greatly whether it will go to your study …

    Best regards,

    w.

  65. How about a new classification for thermometers, thusly: “Out in the woods nowhere near any developed land/structures!” and, wait for it, “NOT out in the woods nowhere near any developed land/structures!” Would that clear it up? Or, if there ARE no thermometers “NOT out in the woods nowhere near any developed land/structures!” let’s put some out there and see which correlate to which, project should take a few months to produce relevant data at a 4- or 5-Sigma level. Anyone who wanted the correct answer to this question would have done this a long time ago. The rest of you are trying to prove something else, and what that might be I cannot ascertain.

    May I have my grant now?

  66. yes there is UHI effect. That has never been the question.
    the question is does that effect BIAS the record.

    Is there any doubt that today different stations experience different amounts of UHI contamination? For clarity, I think we need to specify whether Local Heat Island (LHI) is or is not a subset of UHI. For the benefit of this study, LHI is not included in UHI. Agreed?

    So the question then becomes, “Does UHI or LHI BIAS the record”.
    LHI is real.
    Different stations experience different LHI today.
    To assume that LHI does not bias the readings means you must believe that a poorly sited LHI station today was poorly sited when it was set up. This strains credulity. Did people set up a station next to a parking lot or did the parking lot expand toward the station?

  67. Robuk says:

    Steven Mosher says:
    April 4, 2013 at 11:39 am

    Mike Ballantine says:
    April 4, 2013 at 11:24 am
    Data mining in corrupted data yields corrupted results. Willis’ example highlights a major problem with the instrumental record. Temperature measuring in the middle of an airport has only ONE valid and very important use. It gives the pilots information they need to set safe, efficient take off and landing speeds. Any other use is like using pliers to hammer nails.
    IMHO, ALL airport sited temperature stations should be excluded from the climate data sets.

    ########################

    thats a good point. i did that. guess what? no difference. some airports cool other warm.
    remove them all and the global trend doesnt change.

    Also, you can actually compare CRN stations 9 the one Anthony likes ) to nearby airports on an hour by hour basis.

    guess what? yup. no difference.
    ——————————————

    Total garbage,

    Try comparing them since say 1945.

  68. Steven Mosher says:

    “Thanks, Josh. Poorly sited stations do not necessarily introduce a trend. But it it quite common. For example, at Henderson Field they lengthened the runway in the 1980s to bring in the larger jets … which increased both the pavement within the viewshed and the number and size of the planes landing and blowing hot exhaust onto the thermometer. Since development is much more common than de-devlopment, most changes increase the surrounding temperature.

    w.

    ########################

    the temperature falloff for a jet plume is rather dramatic. I’ll refer you the ground handling safety guidelines published by airline manufacturers. Just request them or you can find a few of them on the web. Also, if the blowing of exhaust onto thermometers was a regular occurance it would be visible in the one minute data from airports. Its not. if it was persistent then you’d see it by comparing CRN data to nearby airports. Again, nothing.

    This isnt to say that it cannot happen, only that the conditions are rare and not easily found in the actually data.

    So

    1. the plume falls off rapidily.
    2. if it were a common occurance it would show up in 1 minute data. it doesnt.
    3. it doesnt show up comparing the best (CRN) to nearby airports.

    All that said, airports, statistically speaking, make no difference whether they are included or excluded from the dataset. You can find worldwide airports on the internet. the FAA has taken down their resource because of terrorism. The open datset includes all airports and heliports.
    volunteer built just like surface stations.

  69. Of course the media, and unfortunately the BESTers, still can’t lower themselves to the plebeian truths. All those highly trained individuals, quite embarrassing.

    Thermometer Placement Oversight.

    Where are the official plans to move and correct the stations that get built into a hot spot? Anybody know?

    Why should politically aspiring scientists want stations moved, when they serve their agenda left just as they are?

  70. Willis Eschenbach says:

    Steven Mosher says:
    April 4, 2013 at 11:14 am

    ‘Now, each “pixel” that they mention above is one of the gridcells shown in Figure 1 above. Those are the actual gridcells used in the MODIS map. If more than half of a gridcell is “built environment” (houses, roads, runways, etc.) then the gridcell is counted as “built-up land”. However, there’s a final hurdle. You need to have five adjacent gridcells of built-up land to have those gridcells classed as part of an urban area in the MOD500 map.”

    Actually not. 4 cells will do.

    Actually not. The cells (at the equator) are 463 m x 463 meters. Four of them make about 0.85 square km, and the MOD500 data (quoted above) says there must be 1 square km of contiguous built-up gridcells to be called an urban area. So your claim is 100% wrong.

    This is why I went through the MOD500 information in detail, because I was pretty sure you guys didn’t understand it … and as it turns out, you didn’t.

    Actually, it gets worse from there. At 60°N, it takes ten gridcells to make up one square kilometre. So your claim is not just wrong, it’s risible.

    In addition, you have defined “very rural” as being more than a tenth of a degree of latitude or longitude away from any “urban” spot on the MOD500 map … which means that you are allowing sites to be “rural” near the poles which would not be rural at the equator. That’s a weird way to do business, using an elastic measure like degrees of longitude. Might not make a difference, but it might, so why do it that way?

    Finally, as I showed above, the land surface of the planet could be 70% parking lot, and the MOD500 map might not show one single urban area … and I’m pretty sure you guys didn’t realize that either .

    Regards,

    w.

  71. Go Canucks Go!! says:

    Thanks Willis.
    I ran 2 trend lines at TreesforWood. Both BEST and RSS were land only from 1980 until 2010. Best was running about .27 per decade and the satellite was running about .2 degrees. I don’t think the satellite data has UHI or LHI problems

  72. Reg Nelson says:

    Wouldn’t it make more sense to take the poorly sited stations and set up nearby stations in better suited locations, then compare the results over a period of time? How difficult would that be?

    The NOAA has already done this (to some extent) and found that stations located closer to their building have higher nighttime minimum temperatures than those located further away. That’s actual empirical evidence (and common sense). Anyone who claims that there is no UHI effect is therefore clearly a “denier”. LOL

  73. AJ says:

    What’s the status of Watts et. al (2012?/2013?) ?

  74. Steven Mosher says:

    ‘Even more sadly, Steven, there’s not one word in the study about your magic airport test, or about your claimed study of the question, so once again it’s just another jolly Mosher anecdote … have you given up entirely on providing citations for your claims, or is this just a passing phase and your unscientific lack of citations will “self-correct” about when the rest of climate science self-corrects?”

    the sensitivity tests are mentioned. Not in much detail, the answers didn’t change. In short,
    When Zeke and I came on board we had just finished a massive sensitivity test. That basically allowed me to reclassify the stations using a screen that used, modis, isa, nightlights, airports, 10km screens, 25km screens. None of them gave different answers than the .1 degree screen.

    so, the simple .1 degree screen gave an answer, sensitivity tests about that screen gave the same answer. Not really interesting.

    At some point I have to figure out how to write up a paper that says i tried all this stuff and could not find a UHI signal.

  75. According to Richard Muller, himself,

    “Urban areas are heavily overrepresented in the siting of temperature stations: less than 1% of the globe is urban but 27% of the Global Historical Climatology Network Monthly stations are located in cities with a population greater than 50000.”

    http://notalotofpeopleknowthat.wordpress.com/2011/10/23/mullers-problem-with-uhi/

    How many more would there be if they took the population down to, say, 5000? And excluded airports?

    If BEST really want to exclude UHI, let them ignore all of these sites completely and give us figures based on only on sites that have been guaranteed as reliable by the local Met Offices. If such guarantees cannot be given, that country’s sites should be ignored completely.

  76. Geoff Sherrington says:

    Australia has climate records for over 1,000 sites. A few years ago I selected about 50 of those least subject to the hand of man, which I called pristine.
    The object was to set a baseline. Ideally, all pristine sites should have a similar trend over time, being natural climate change. This would make a baseline above which UHI would show.
    It did not. The complete demolition of methods used for quantification of UHI using regional remote methods follows in these figures.
    Sorry BEST, you have no story until you can explain what is going on here.
    http://www.geoffstuff.com/Pristine_Summary_1972_to_2006.xls
    (Data provided, so please be patient with download time).

  77. I think one of the things Willis’s analysis shows is just how dangerous it is to rely too much on statistical wizardry and algorithms and not enough on proper local and practical knowlege.

    GHCN’s Icelandic adjustments are a classical example of this. Who knows better the temperature history of Iceland? The experienced, knowledgable Iceland Met Office, or some computer programmer in GHCN?

  78. Steven Mosher says:

    ‘Actually not. The cells (at the equator) are 463 m x 463 meters. Four of them make about 0.85 square km, and the MOD500 data (quoted above) says there must be 1 square km of contiguous built-up gridcells to be called an urban area. So your claim is 100% wrong.

    This is why I went through the MOD500 information in detail, because I was pretty sure you guys didn’t understand it … and as it turns out, you didn’t.

    Actually, it gets worse from there. At 60°N, it takes ten gridcells to make up one square kilometre. So your claim is not just wrong, it’s risible.”

    The easy way to do it is this.

    1. you take all the stations and classify them using the urban dataset ( 1km rule)
    a when you do this make sure you fix the projection )

    2. you then take a 500 meter dataset ( the source of 1 ) and do the classification using that.

    3. you then take the 300 meter dataset and do the classifiaction using that

    4. you then take a 30 meter dataset ( US and alaska) and triple check.

    You’ll be surprised, but no nobel prize

  79. Willis Eschenbach says:

    Steven Mosher says:
    April 4, 2013 at 11:18 am

    … Dude I’m still waiting for somebody to accept my challenge.

    1. Define Urban or rural ex ante in a way that is objectively measureable.
    2. i will divide stations into urban and rural per your definition.
    3. i will compute the difference.

    A cookie for anyone who can find the signal.

    So, there is the challenge.

    For me the challenge is to convince people get off of the false “rural/urban” dichotomy entirely. There are terrible stations everywhere, being out in the country is no guarantee of good siting.

    As I said above, the problem is in the name. If you called the phenomenon by a more accurate name, the “Local Heat Island”, you would see that (as you point out implicitly in your challenge) likely there is no way to divide stations into good and bad local sites using ex-ante rural/urban criteria. I think it can only be done by examining each site one by one, and using the actual siting criteria. There’s no magical shortcut, rural/urban, MOD500, or otherwise.

    The result of your study, Steven, is only that you’ve shown that using MOD500 can’t separate good stations and bad stations from a satellite … you have to do it from the ground.

    Finally, I was curious why you didn’t use your technique to investigate things like the Oke paper or the McKitrick paper which found evidence of UHI. I’d think you could use your method and the Oke or McKitrick datasets to determine whether they were wrong or not and if so why … seems like if you’re going to implicitly claim the previous studies are wrong, you should show it directly using their data.

    w.

  80. John Bell says:

    Great article, Willis! The AGW crowd are engaged in what Irving Langmuir called “Pathological science” and they know how to tweak the little things to twist facts to suit their theory.

  81. Willis Eschenbach says:

    Steven Mosher says:
    April 4, 2013 at 11:23 am

    Bob says:
    April 4, 2013 at 10:16 am

    Steve Mosher, I guess we shamed you into becoming an author. Congratulations

    #################

    Not. I have never asked and will never ask to be a co author. i do the work I do and don’t care very much about were the credit is given. acknowledgments or byline? phhht. I do what i do.

    Well said, Steven.

    w.

  82. GingerZilla says:

    Syncronicity as I pondered this only a day ago. Stood on a train platform freezing my unmentionables in a bitter Easterly wind. Train leaves platform giving gust of diesel and sudden onset of mediterranean exhaust warmth. Appreciated despite vile smell but gone as sudden as onset. Pondered weather stations in back fire range going from sub 30s F to high 50s (heatwave by current Brit standards). However despite Willis pictures BEST have said ‘nothing to see here move along’. Poor cold GingerZilla gazing at snow in South England in April now confused as told heat all in mind. Many Brits also in state of Mass Hysteria seeing imaginary snow in town that melts on contact but by a miracle settles down the country road-all deniers do not believe country folk. Rural people total liars about cold – all in their denier heads. No heat in town.*

    /sarc

    *this is factually correct as even in the pub someone keeps opening the door denying me heat.

  83. Willis Eschenbach says:

    Steven Mosher says:
    April 4, 2013 at 12:28 pm

    ‘Even more sadly, Steven, there’s not one word in the study about your magic airport test, or about your claimed study of the question, so once again it’s just another jolly Mosher anecdote … have you given up entirely on providing citations for your claims, or is this just a passing phase and your unscientific lack of citations will “self-correct” about when the rest of climate science self-corrects?”

    the sensitivity tests are mentioned. Not in much detail, the answers didn’t change. In short,
    When Zeke and I came on board we had just finished a massive sensitivity test. That basically allowed me to reclassify the stations using a screen that used, modis, isa, nightlights, airports, 10km screens, 25km screens. None of them gave different answers than the .1 degree screen.

    so, the simple .1 degree screen gave an answer, sensitivity tests about that screen gave the same answer. Not really interesting.

    At some point I have to figure out how to write up a paper that says i tried all this stuff and could not find a UHI signal

    Thanks, Steven. So I take it that you did NOT do a sensitivity test on the MOD500 results reported in this paper, showing the “very rural” stations with and without airports? I’m not claiming it would show anything, just asking if it was done or not.

    In any case, the issue is not airports. It is bad siting of the Stevenson Screen, which has nothing to do with airports.

    w.

  84. James at 48 says:

    To understand superposition of anthropogenic energy loss / flux upon background temperature in a given zone, one must perform the integral of all anthropogenic energy sources and anthropogenically created energy imbalances due to albedo modifications within that zone. The whole concept of an “urban” heat island is flawed and antiquated. The above approach is what is needed. This approach can apply to any zone be it an urban area or at the opposite extreme, supposed “wilderness.”

  85. ralfellis says:

    Steven Mosher says: April 4, 2013 at 11:39 Temperature measuring in the middle of an airport has only ONE valid and very important use. It gives the pilots information they need to set safe, efficient take off and landing speeds.
    _____________________________________

    Not quite, Steven.

    Temperature is irrelevant to landing speeds, which vary according to weight only (plus an allowance for high winds and gusts).

    For take off, temperature is mainly used to determine thrust, and it is thrust and weight that determine the take off speed, not temperature. On our jet, temperature alone only adds one knot to V1 and V2 airspeeds, between 0 and 40 degrees c.

    .

  86. Willis Eschenbach says:

    Steven Mosher says:
    April 4, 2013 at 12:38 pm

    ‘Actually not. The cells (at the equator) are 463 m x 463 meters. Four of them make about 0.85 square km, and the MOD500 data (quoted above) says there must be 1 square km of contiguous built-up gridcells to be called an urban area. So your claim is 100% wrong.

    This is why I went through the MOD500 information in detail, because I was pretty sure you guys didn’t understand it … and as it turns out, you didn’t.

    Actually, it gets worse from there. At 60°N, it takes ten gridcells to make up one square kilometre. So your claim is not just wrong, it’s risible.”

    The easy way to do it is this.

    1. you take all the stations and classify them using the urban dataset ( 1km rule)
    a when you do this make sure you fix the projection )

    2. you then take a 500 meter dataset ( the source of 1 ) and do the classification using that.

    3. you then take the 300 meter dataset and do the classifiaction using that

    4. you then take a 30 meter dataset ( US and alaska) and triple check.

    You’ll be surprised, but no nobel prize

    Steven, you claimed four gridcells were enough to be counted as an urban area by MODIS in the MOD500 map. That was not true.

    A simple “Gosh, I guess I was wrong about the four cells being sufficient in MOD500″ would have sufficed.

    Giving me an unwanted lecture about some other method of investigation you advise us to take is just handwaving to distract people from your error. Man up, admit your error, and move on. Then we can discuss your next proposed analysis.

    w.

  87. Mike Jonas says:

    They have learned nothing since their original paper. In fact, it looks like a regurgitation of the exact same method. I’m particularly disappointed to see that (a) they do not cite Watts(2012) and (b) Judith Curry is still a co-author. There was a lot of discussion on Judith Curry’s blog of the first paper, and it is very disappointing that she in particular appears to have learned nothing from it. They surely must have been aware of Watts(2012). Since it shows that their method is totally flawed, it is breathtaking that they go ahead without a mention. Surely responsible scientists would at the very least cite the paper and say why it didn’t apply to them. [I'll answer my own question: The obvious problem is that it does apply to them and it tells them that what they are doing is invalid. ie, they are not reponsible scientists. It pains me to say this about Judith Curry, as she has been prepared to take a lot of flack in the interest of good science.]

  88. GingerZilla says:

    Michael R. Moon on April 4, 2013 at 12:14 pm

    Appart from a lack of a hockey stick you did not mention ‘Climate Change’, ‘Weird Weather’ or ‘Unprecidented’. Sorry your grant has been declined.

    We, the people of the gate, suggest you apply the FEAR Index for a statistically significant study/press release that the media will pay attention to.

    /sarc

  89. ralfellis says:

    Oh, and as an aside. Pilots like temperature sensors that are in the ‘wrong’ locations. We don’t want to know the temperature out in the countryside, we want to know the temperature of the air that our engines are sucking in during take-off – which is the temperature out on the burning-hot black ashphalt runway.

    Side-of-runway locations give a great temperature for aircraft, but a lousy temperature for weather/climate studies. It warms up more than rural locations in the day, especially if the runway is black, and it cools down more at night. Luckily, grooved concrete is becoming more common for runways, which one assumes does not have such temperature extremes.

    .

  90. knr says:

    The basic trouble with using airport based measurements is that airports are not typical of their surrounding environments. That did use to to matter becasue these measurements where ‘only intended ‘ to be used for aircraft movements and and out of the airfield so knowing what the conditions on the airfield was fine .
    Once they became used for 101 other things , largely because they were there already and so it meant you did not have to create new sites not becasue there were ‘good sites’ , that is when the trouble started.
    And even that it did no really matter because people accept weather is chaotic and hard to predict so ‘approx ‘ was good enough . But with AGW and its ‘settled science’ we seen claims of super accuracy with the data taken political importance , whilst the old problems over airport based measurements still remain .

  91. ed mister jones says:

    This Page is a Photo Riot of Thermometer siting.
    http://www.bobbyshred.com/fools/falsetemps.html

  92. knr says:

    ralfellis before jets , when most airport where set up , knowing the weather conditions was a lot more important . There are still issues , such has the need to de-ice, that need temperature data .

  93. catweazle666 says:

    My wife has told me I mustn’t read blogs like this because it puts my blood pressure up!

  94. ed mister jones says:

    Steven Mosher says: April 4, 2013 at 11:39 Temperature measuring in the middle of an airport has only ONE valid and very important use. It gives the pilots information they need to set safe, efficient take off and landing speeds.

    I’ve been flying for . . . too many, 36(?) years – That is news to me! What the other guy said.

    My primary concern with temp at the airport is airfoil contamination by Ice/Snow/Frost/Freezing precipitation, or whether or not I’ll have to have a few hundred gallons of Glycol mix sprayed at very high cost. Temperature WILL affect air density and therefore Ground Speeds, and will affect landing and take-off distances, but it has >NO< significant impact on what AIRSPEEDS the Pilot or Automation systems reference to operate the Machine. To be thorough: As temperature increases, density decreases (pressure constant) – a given AIRSPEED (mass flow) will require a higher GROUNDSPEED. Hot days or High Elevation airports require longer runways or lower aircraft weights, all else being equal.

  95. Matt G says:

    This is not addressing the main issue with regarding rural and urban.

    An airport location if it is warmer will generally stay warmer all the time without changing trend. A badly located station will always remain on the warm side with the trend not changing. The trend is no good to use in these studies as it is totally irrelevant. When a rural location suddenly has a airport built, the station will only change for one years data suddenly and after that the trend’s influence on the months will be same over the long term. It only takes a change like this to offset the same data set over a recent period from one much longer in the past. The trend says nothing about how sudden changes to local stations have a long term effect.

  96. Jonathan Abbott says:

    I supply meteorological equipment to airports for a living. The exact siting usually depends on the availability of power and comms rather than the position that will give the most accurate representation of the temperature on the runway (which as mentioned above is what the pilots want, though not the meteorologists). There are also ICAO standards that define how far away instruments should be from the runway centreline, for safety reasons. This is typically around 100m minimum.

  97. Matthew Benefiel says:

    This makes me think of an applicable engineering example where one has to qualify his design by running it through a thermal test. Let’s say for this example the customer wants you to perform an eight hour test at -50 and since you usually use Fahrenheit at work you assume that is correct (it happens, information is lost in translation). Let’s also say you are a small company with no thermal chamber of your own so you have to hire a company to test it for you, while supplying someone to travel out there to monitor the design. You do all this, get the results back and after a week of aligning your results with the test house results you write your test report. You send it to the customer and they immediately come back telling you that it was -50 Celsius. Well you don’t want to spend the money to travel out again and you are under the gun to finish this all up so you make an addendum stating that the test run was only 4.44 degrees (Celsius) higher which isn’t much. Furthermore you look up all the components and see that they are all rated to -55 C (for this example) so they will work either way.

    Should be acceptable right? Yet your unit hasn’t run at -50 C yet and from experience a few degrees can make a difference, especially if the design has flaws. A simple operating temperature range for each part doesn’t mean the whole design will work through that range and something as simple as a floating pin can cause problems at low temperatures. In the end the test has to be performed according to the customer’s demands or it hasn’t passed.

    Seems to me the siting issue is very similar. It is expensive to go back and do so it’s easier to try and “fix” it with analysis, but in the end you have to bite the bullet or you will never know. The real work here is the gathering of data, not the analysis of that data (as important as that is). You also end up spending more money and time than if you had gone back in the first place.

  98. Gene Selkov says:

    ralfellis: maybe the values of V1 and V2 for your aircraft are based on the worst-case scenario at the worst airport where the machine is certified to land? The decision speeds seem to be set somewhat arbitrarily, and in different manner for different aircraft.

    Surely your Vy depends on altitude, quite a bit. According to this chart, at 40C, you’re already at 7500 feet above the ground, compared to standard atmosphere:

    http://upload.wikimedia.org/wikipedia/commons/thumb/6/6f/Density_Altitude.png/567px-Density_Altitude.png

    I hear from small aircraft owners that they get stuck at high-altitude airports for days, waiting for a cold enough weather to take off? In what way are their machines different from yours, other than having a lower ceiling?

  99. ed mister jones says:

    This is reminding me of the North Korea at night from space Image . . . What I mean is, that it is a good relative representation but distorts the reality of the contrast area . . . IOW, there’s a Whole Lotta Nothin’ between the Connecticut coast and Plattsburgh NY, between Portland, and Presque Isle Maine, Between Gander Newfoundland and Boston, MA . . . . Between Vancouver, BC and Portsmouth, New Hampshire.

  100. u.k.(us) says:

    ralfellis says:

    April 4, 2013 at 12:49 pm

    Steven Mosher says: April 4, 2013 at 11:39 Temperature measuring in the middle of an airport has only ONE valid and very important use. It gives the pilots information they need to set safe, efficient take off and landing speeds.
    _____________________________________

    Not quite, Steven.

    Temperature is irrelevant to landing speeds, which vary according to weight only (plus an allowance for high winds and gusts).

    For take off, temperature is mainly used to determine thrust, and it is thrust and weight that determine the take off speed, not temperature. On our jet, temperature alone only adds one knot to V1 and V2 airspeeds, between 0 and 40 degrees c.
    ===============================
    I fear I’m in way over my head, but pilots need to know the temperature, because it affects the length of the take-off run, which limits the length of runway left to stop in case of an engine failure.
    (do you take-off and limp the aircraft around for repairs, or slam on the brakes/ reverse thrust and run off the end of the runway).
    Temperature matters for pilots, higher temperatures decrease engine power.

    I’ll leave the nuances of where to place climatic weather stations, to others, for now :)

  101. UK John says:

    They looked for any bias due to UHI and didn’t find it in the Global Temperature Land Average.

    Other things usefully you could look for in the Global Temperature Land Average is evidence of confirmation bias and perhaps what effect a massive human population increase has had.

  102. ed mister jones says:

    How does one measure and classify widespread temperature inversions fairly common to Mega – Urbanized areas surrounded by mountains, like LA for example? Is it a metric of Global Climate, or an outlier? Is it Climate or Anthropogenic aberration? I think all temps should be taken in locations and by uniform methods that are Anthropogenically (word?) Sterile . . . then you would be observing the “Net” Temperature, and not a mish-mash of variable-divergent data.

  103. wayne says:

    Willis, I notice you are not speaking of how grid cells have changed over time. A grid every year over many years. However, maybe I missed it, haven’t read every word. Is that not the real UHI differences we are speaking of that do affect temperature anomalies, not static view at moments, but over time. Seems Mosher lost track of this ‘time’ as BEST got the best of him.

    Other have admitted that urban areas have an offset in the anomalies though those already developed have the same, or close, anomalies. But — the number of developed cells over time have drastically increased as years went by. How can someone not see what these facts lead to… a warmer, not globe, but the readings from a satellite standpoint or the average of temperature stations near development.

    If we are speaking of all-city’s-global-averaged-temperature, I have no problem, cities have grown over time and bring with them warmer local temperatures, that is what we do read, but globally, that cannot actually be more than a few percent taking every swatch of this globe into account.

    So, I see the problem not as just UHI or LHI but UHΔOT, urban heat differences over time. — that does affect the anomalies over time like in all of the time-series we look at. I see no one addressing this.

  104. Chad Wozniak says:

    Just another phony exercise in modeling. More evidence that climate models shoudl be outlawed as the basis for anything, let alone policy decisions. Willis, good job at exposing this latest fraud.

  105. Its worth pointing out our AGU poster on this last year, which used a more restrictive filter for rural (e.g. modis = 0, airports = false, nightlights < 30, impermeable surface area < 10):
    http://wattsupwiththat.com/2011/12/05/the-impact-of-urbanization-on-land-temperature-trends/

    I've still got some qualms about all the global UHI work I've seen to-date (including our AGU poster). There just isn't enough station density (or enough rural stations with long records) in some areas to do a good comparison.

    The U.S. is relatively easy by contrast, as we have 7000-odd co-op stations, many of which have long continuous records. The U.S. is also blessed with accurate lat/lon coordinates of instruments (at least down to 30 meters or so), something not true for much of the rest of the world. Here is our recent JGR paper on U.S. UHI where we found a rather large signal in the raw data (~14-20% of the century scale trend in minimum temperatures), though it was significantly reduced by homogenization: ftp://ftp.ncdc.noaa.gov/pub/data/ushcn/papers/hausfather-etal2013.pdf

  106. Steven Mosher says:

    Fig. 7 The frequency of urban patch sizes (x-axis, log scale,
    spline curves) for each map (excluding IMPSA and HYDE3).
    Note the two peaks for MODIS (orange) and GLC00 (black)
    for patches below 2 km2
    . These peaks correspond to single and
    two-pixel patches, which may indicate speckle in these maps.
    These size distributions are unique to remote sensing-based
    products and much rarer in the GRUMP (blue) and VMAP0
    (red) maps. The modal peak in GRUMP is for patches between
    29 and 33 km2
    in total area. The majority of these mid-sized
    patches are found in East and South Central Asia. The
    maximum frequency of the MODIS patches was 5,213 at an
    area of 0.67 km2

  107. Steven Mosher says:

    ed mister jones says:
    April 4, 2013 at 1:27 pm
    Steven Mosher says: April 4, 2013 at 11:39 Temperature measuring in the middle of an airport has only ONE valid and very important use. It gives the pilots information they need to set safe, efficient take off and landing speeds.

    I’ve been flying for . . . too many, 36(?) years – That is new

    ##################

    err you are confusing me with the reader I was responding to

  108. BarryW says:

    Hmmm, anybody do some transects so there’s some ground truth?

  109. Louis Hooffstetter says:

    Nice job once again, Willis. Thanks. Love the word ‘embiggen’. It’s perfectly cromulent!

  110. Louis Hooffstetter says:

    Moshpit says:
    “i do the work I do and don’t care very much about were the credit is given. acknowledgments or byline? phhht. I do what i do.”

    That’s why we love you Steve! You are truly “an honest broker”. Just because we don’t usually agree, doesn’t mean we have any less respect for you. Keep up the good work! (but try to come around to our way of thinking, OK?). Sarc!

  111. richard verney says:

    It is almost inconceivable that a UHI signal does not exist in the land based thermometer record, and if this signal cannot be found in the data set, then it suggests to me that the data set contains significant problems with resolution.

    Is not the best indicator of UHI a comparison between the satelitte data sets and the land based data sets for the period say 1979 to 1996? Of course, the land based thermometer data sets would need trimming to the same latitude coverage as the satelitte. Further, ocean temps would have to be included since these dampen response.

    I would have thought that that would represent as good as any starting point for investigation. Then a more detailed study could be conducted using only stations that have the highest siting and are truly rural (this would have to be ascertained by physical inspection) and then compare those selected stations with stations that are situated in urban areas where urbanisation has been growing over a period of say 30 to 50 years.

    Once something is fully urbanised, the trend of developing UHI is lost. It is the impact of the growth of urbanisation encroaching on rural or semi rural sited stations that is causing the pollution.

  112. Steven Mosher says:

    Thanks, Steven. So I take it that you did NOT do a sensitivity test on the MOD500 results reported in this paper, showing the “very rural” stations with and without airports? I’m not claiming it would show anything, just asking if it was done or not.

    In any case, the issue is not airports. It is bad siting of the Stevenson Screen, which has nothing to do with airports.

    ###########################################

    Not exactly

    The reported tests were done at .1deg

    Then, I took the list of stations, and reclassifed all the stations using a finer screen
    11km free of modis, plus no ISA 1km ( to address some issues at northern latitudes)
    zero nighlights, and no airports.

    That classification along with some other sensitivity tests were run. Zeke and I came to the meeting expecting the answer to change…… nope, found nothing.

    non results are not that interesting. Engineering wise, of course you blather on about
    all the crap you tried. I’v tried anthromes, biomes, nightlights, airports, land cover,
    modis urban, modis landcover, 300 meter data, 30 meter data, albedo, emissivity,
    population, population density, population growth, etc etc etc. and all combinations.
    folks are welcomed to accept the warning that you wont find anything dramatic there, or they can go break their back looking through data for 4 years, or they can say I dont believe you. Or they can realize that the real issue is siting.

    there was also a study done of airports versus non airports. nothing to show, except some cooling airports in japan. I believe Muller would have shown those charts to Anthony on his visit, since they were shown to me on my first visit and confirmed the independant work I had done.

    In the end, for me, finding something would have much more interesting than finding nothing.
    Thats why I liked Zeke paper so much ( and our poster) because there was at least a small signal in the US data.. and a small signal in worldwide raw data, but in the final analysis.. nothing

    There are two issues that I think deserve more attention.

    A) the small town effect. You will note on my site that this is an ongoing project, but to do it world wide I need to reprocess landsat 28m data. Just time consuming. Maybe its hiding there. Put another way, I can’t rule out that there may be a small effect sub 1km. The hitch is this. Spending a year of data processing to exclude that and not having a publishable result ( hey ma, found nothing again ) isnt on my bucket list anymore. lets put it this way.
    I do an airport/non airport study and find nothing. AGW folks dont care. And Skeptics wont believe it. So whats the point? In the begining the point was figuring it out for myself. And it was kinda nice to have some folks say ‘thanks for finding nothing”. As for those who were disappointed when nothing was found.. sorry. if you want help looking for yourself, i was nice enough to build tools for you. have at it.

    B) microsite. What is needed is an objective measure of good siting. While leroy is a good qualitative start, it is really lacking in terms of scientific backing.. That is, there are no field tests or measurements to back up the CRN1-5 rating. That’s changing…. But the
    biiggest issue here is calculating the ACTUAL viewshed of the sensor. Strangely enough the area directly under the sensor is less important than stuff 100m away. hmm, have a read through design docs for eddy covariance methods for flux towers.

  113. Lil Fella from OZ says:

    Thanks Willis. I have always had a problem where temperatures are measured (and how) in Aus. This was long before the AGW commotion. There are many things which influence temperature readings. There are difficulties in comparing recorded temperatures 20 years apart let alone 100 or more!

  114. Steven Mosher says:

    Geoff Sherrington says:
    April 4, 2013 at 12:30 pm
    Australia has climate records for over 1,000 sites. A few years ago I selected about 50 of those least subject to the hand of man, which I called pristine.
    The object was to set a baseline. Ideally, all pristine sites should have a similar trend over time, being natural climate change. This would make a baseline above which UHI would show.
    It did not. The complete demolition of methods used for quantification of UHI using regional remote methods follows in these figures.
    Sorry BEST, you have no story until you can explain what is going on here.
    http://www.geoffstuff.com/Pristine_Summary_1972_to_2006.xls
    (Data provided, so please be patient with download time).

    #############################

    Now THATS more like it. a man who has done some work.

    A few questions and I will have a look at it.

    1. What methodology did you use for classifying pristine?
    a) what quantitative metrics.
    b) are these metrics tied to known causes of UHI ( see oke, they havent changed for
    decades of research.)

    2. What is your data source.

    3. How did you process the data.

    I will note this. almost every freaky site in the world comes from down under. outside the US it is the only place with long term cooling stations. and urban proxies like population are not very good there ( industry towns report zero population)

  115. little polyp says:

    Willis bula

    From the 2 degrees of separation of the RSYC, thank you.

    It perplexes me that in this age, the likes of BEST or other, less diligent, authors feel the need to arrive at a conclusion. What is it and why have these “scientists” forgotten that either no answer or the wrong answer is just as important. Is it really the case that we are in a 24 hour news cycle and that ego, prominence or money mean that say Marcott and co cant bring themselves to say “as a consequence of this study, we are unable to discern an outcome that would allow us to offer tangible insights into this issue ?”

  116. Doug Proctor says:

    The problem here is the same I see in industry these days: if you can’t dump and troll through datasets, then it ain’t either science or worth it. Manually looking at sites and making a decision is too “subjective” for science these days. Maybe because if you did this, you would have an “opinion”, but if it is just “data”, you are off the hook: hey! that is just what the program says!

    It is like looking at the individual proxies of Marcott and saying, each of these has too much variation to tell us solid things about any case, so why would an averaged group do better? GIGO used to be a real concern. Now if it is digital it is gold.

    If Darwin were working today, he’d be shouted off the stage for not working with computer models and digital data.

    Thanks, Mr. Jobs.

  117. Steven Mosher says:

    ‘Steven, you claimed four gridcells were enough to be counted as an urban area by MODIS in the MOD500 map. That was not true.

    A simple “Gosh, I guess I was wrong about the four cells being sufficient in MOD500″ would have sufficed.

    Giving me an unwanted lecture about some other method of investigation you advise us to take is just handwaving to distract people from your error. Man up, admit your error, and move on. Then we can discuss your next proposed analysis.”

    ##########################################

    Actually Willis, I havent said whether you are right or wrong because there are two ways
    that folks describe what you are talking about. There are over 5000 modis urban pixels that have an area of .67 sqkm. How does that happen?

    Fig. 7 The frequency of urban patch sizes (x-axis, log scale,
    spline curves) for each map (excluding IMPSA and HYDE3).
    Note the two peaks for MODIS (orange) and GLC00 (black)
    for patches below 2 km2
    . These peaks correspond to single and
    two-pixel patches, which may indicate speckle in these maps.
    These size distributions are unique to remote sensing-based
    products and much rarer in the GRUMP (blue) and VMAP0
    (red) maps. The modal peak in GRUMP is for patches between
    29 and 33 km2
    in total area. The majority of these mid-sized
    patches are found in East and South Central Asia. The
    maximum frequency of the MODIS patches was 5,213 at an
    area of 0.67 km2

    And I believe from looking at the actual modis plots that there are areas specified as urban that have areas less than 1km. Rather than call you wrong, I’ll just go look at the data. Solomon islands should be a good choice. and some northern latitude spots. Not a big deal, since one way i cross check modis is by using nightlights, and ISA.. you will find minor differences with all the methods.. meh.

    Ps.

    Lose the conspiracy ideas about the author order. When I reformated the paper, I put Charollte in first cause she was principle and then just put names down as i remembered them. My bad.

  118. Gary Pearse says:

    Certainly, I trust Steven Mosher’s diligence in searching for the truth and believe he was surprised to find no significant affect. I do however, think the question of UHI (or Willis’s LHI) is deserving of detailed experimentation. This would involve setting up temporary thermometers say 100, 500 and 1000m going away from the stationary thermometer(and wind direction and strength instruments) in 3 directions at 120 degrees for, say, a week or two. If the results show no meaningful difference at 10 randomly chosen stationary thermometer locations, then the argument would be largely over. We probably do more energetic work on less important issues than this. Come to think of it, experiment seems singularly absent from climate science if we rule out running iterations of climate models…

  119. Nick Stokes says:

    Willis,
    I see you spotted a plane at Honiara International Airport. But it doesn’t seem very crowded.

    I know you can fly from there to Nadi and Port Moresby. And there is a weekly Virgin flight to Brisbane. But is that traffic going to seriously affect the temperature record?

  120. Willis Eschenbach says:

    Zeke Hausfather says:
    April 4, 2013 at 2:46 pm

    Its worth pointing out our AGU poster on this last year, which used a more restrictive filter for rural (e.g. modis = 0, airports = false, nightlights < 30, impermeable surface area < 10):
    http://wattsupwiththat.com/2011/12/05/the-impact-of-urbanization-on-land-temperature-trends/

    I've still got some qualms about all the global UHI work I've seen to-date (including our AGU poster). There just isn't enough station density (or enough rural stations with long records) in some areas to do a good comparison.

    The U.S. is relatively easy by contrast, as we have 7000-odd co-op stations, many of which have long continuous records. The U.S. is also blessed with accurate lat/lon coordinates of instruments (at least down to 30 meters or so), something not true for much of the rest of the world. Here is our recent JGR paper on U.S. UHI where we found a rather large signal in the raw data (~14-20% of the century scale trend in minimum temperatures), though it was significantly reduced by homogenization: ftp://ftp.ncdc.noaa.gov/pub/data/ushcn/papers/hausfather-etal2013.pdf

    Zeke, always good to hear from you. Thanks for reminder of the earlier WUWT post on the question. Also, your other paper was fascinating, I hadn’t seen that. I’d recommend it to anyone interested in the question of UHI/LHI. It poses and answers many interesting questions.

    Have you looked at the same question(s) using the BEST dataset post-scalpel? It seems like that analysis would be worth doing … so many datasets, so little time …

    w.

  121. Darren Porter says:

    How much energy did all the Nuclear tests release into the atmosphere back in the 50′s – 70′s?

  122. Willis,

    Indeed, too many datasets and too little time. One analysis I wanted to see in that Berkeley paper (but wasn’t included) was the UHI signal (if any) pre-scalpal, as the scalpal may be removing a good chunk of the urbanization.

    My project at the moment is to figure out cases in which automated homogenization algorithms fail, either though lots of small breakpoints, lots of trend biases, or sparse station density using synthetic data (similar approach to that of Williams et al 2011). Should be interesting to see when they work and when they don’t once we get the results.

  123. knr says:

    Nick Stokes airports have lots of concrete and lots of buildings even if their stuck in the middle of no where , if you ever act worked on a airport you notice how the ‘pans’ heat up compared to the rest of the place .

  124. The local heat islands, plus the Highest-Lowest sampling are Catastrophic, Anthropogenic, Global Warming.

  125. Tom in Texas says:

    “…allowed me to reclassify the stations using a screen that used, modis, isa, nightlights, airports, 10km screens, 25km screens. None of them gave different answers…”

    S.M., did you try random phone numbers to see if that gave you a different answer?

  126. Leonard Lane says:

    Has anyone ever correlated quality of science in a paper with number of authors? Is it positive, null, or negative?

  127. Power Engineer says:

    UHI exists even at non urban towns such as ours which has only 4500 people….hardly urban. In the last 50 years traffic has grown 10 fold. Trees are largely gone. Lawns have been paved to provide parking for all the houses that have been converted to doctors’ and lawyers’ offices. If a temp station had been located here, it would show warming due to urbanization even though the surrounding air hadn ‘t warmed at all.

    I agree that there is an “urban” definiton problem.

  128. Steven Mosher says:

    ” The W2013 study uses the MODIS data to divide all the world’s temperature records into “very rural” and “not very rural”. Then they show there is little difference in the trends between the two groups.”

    actually, we look within 0.1degrees of the site ( roughly 11km at the equator) to see if there are ANY urban pixels within that radius. if there are any urban pixels within that radius the site gets classified as not rural

  129. Pamela Gray says:

    Calibration checks are a must. Did you do that? Did you take a good random sample of your rural sites and check them from the ground, close up and personal? Did you take a history of those random sites to check for changes in their settings? This is standard practice. Calibrate. EVERY TIME YOU TAKE DATA!

  130. Steven Mosher says:

    Gary Pearse says:
    April 4, 2013 at 4:08 pm
    Certainly, I trust Steven Mosher’s diligence in searching for the truth and believe he was surprised to find no significant affect.

    ###########################

    yes you can imagine the scene when Zeke and I walk in with our improved screen developed for our poster

    1. I dont use the modis 1km rule I use modis source data.
    2. I use ISA as an ADDITIONAL screen because the modis dataset was a feb dataset
    and there was some evidence of snow covering urban areas showing as rural.
    3. I throw in Nightlights to handle those locations, like in australia where you have a mining town
    that is small but has a big nightlights foot print.
    4. I look 11km around each site, not 0.1 degrees which is a different distance by lat

    So, ya, we walk in with our tighter screen and say, this will find it!!!!

    DONK. surprised is a good word. Especially since you will find me being critical of of BEST over on CA, when I knew it all… haha. wrong.

  131. u.k.(us) says:

    They made me yell “clear” before starting my trainer. I’ve seen doves, make moves to avoid my prop, that would make a fighter pilot cry.

  132. Matthew W says:

    Nick Stokes says:
    April 4, 2013 at 4:17 pm
    Willis,
    I see you spotted a plane at Honiara International Airport. But it doesn’t seem very crowded.
    I know you can fly from there to Nadi and Port Moresby. And there is a weekly Virgin flight to Brisbane. But is that traffic going to seriously affect the temperature record?
    ================================================================
    I think that is less about “how much impact” as the simple fact that there is some impact.
    Correct data collecting should be the first priority, otherwise, too much guess work and assumptions.

  133. TMLutas says:

    If you wish to get a lot of information on how UHI shifts over time, look to the new urbanists. They have lots of maps on how cities have sprawled out. Take a look at Buffalo, NY for a great example. They already have multi-decadal maps of sprawl. You just have to overlay the thermometers and the expected UHI effect over time as the sprawl took place.

  134. Reg Nelson says:

    Steven Mosher says:
    April 4, 2013 at 3:43 pm

    @Geoff Sherrington

    “Now THATS more like it. a man who has done some work.

    A few questions and I will have a look at it.

    1. What methodology did you use for classifying pristine?”
    ——————

    Hang on a tick, Steven. You just said in an earlier post that after exhaustive work you determined that station selection\exclusion criteria didn’t change the trend.

    Now, you turn around a few minutes later and ask Geoff what his selection criteria was. Seriously?

    You already answered your own question, The answer is: “It doesn’t matter”.

    ————
    “I will note this. almost every freaky site in the world comes from down under. outside the US it is the only place with long term cooling stations.”

    So, even if Geoff’s finding are correct, they can still be dismissed because Australia is the “Bermuda Triangle” of climate? Is that what you are trying to imply? Seems like you’ve already come up with an excuse without even looking at Geoff’s work. Your confirmation bias is painfully obvious.

  135. TimTheToolMan says:

    If a handful of proxy locations is good enough to determine the earth’s temperature for proxy studies, you’d think a hand picked (no peeking at calculations first!) set of actual temperature recording weather stations with long records that were properly rural (not according to some sort of algorithm based on pixels) ought to be able to get a reasonable value and settle this once and for all.

  136. Puppet_Master_Blaster_Master says:

    Obfuscation.

    The people at BEST are really trying hard to ‘smoke screne’ the world. Guess they will soon start offering online casino betting plans, bitcoin investment options and ‘life’ insurance just like Al Capone’s ‘companies’ did.

  137. Bill H says:

    The Urban Heat Island is real. Anyone driving through a city can watch their car temp thermometer increase and decrease as the population density changes. It is not a problem however until those temperatures (which are inflated due to changes in land use) are then averaged into the global mean temperature. The land changes cause divergence from reality and show a warming trend, when in reality there is none.

    BEST missed the band wagon on this one. They did all kinds of work and missed the forest through the trees. NOAA and GISS are silent on the CRN sitings and their temp records. That silence is affirmation to me that things are not as dire as the CAGW crowd wants everyone to believe.

    In my humble opinion the determining of what constitutes “built up” and Rural is an exercise in futility and proves nothing..

    Anthony’s approach of individual stations dealt with individually is the only proper course. Once that individual stations bias is determined only then can a correction be placed to bring it back into reality.

    Localized Heat Island (anomaly) is much like weather. Weather, long term, is Climate and LHI long term is UHI.. It is simply a bias due to changes locally.

  138. Reg Nelson says:

    Steven Mosher says:
    April 4, 2013 at 3:43 pm

    @Geoff Sherrington

    “Now THATS more like it. a man who has done some work.

    A few questions and I will have a look at it.
    __________

    And what if Geoff’s reply was, “Why should I show you my work? You’re only going to try and find something wrong with it.”

    Would you be happy with that response, Steven?

    What if Geoff told you he he did a survey and 97% (of some of those) surveyed believed his conclusions to be correct? Would that convince you?

    What if you found out that your taxpayer money payed for Geoff’s work and that he was trying to hide that fact from you? How would you feel about that? Would you trust Geoff?

  139. Steven Mosher says:

    Well Willis, lets have a look at the data. You are concerned that this site will be Rural classified by our proceedure. Correct?

    And to determine that you read a modis paper, but never actually wrote to the PI to get the actual data? Correct? or not?

    Willis There are these sites I find in our records

    Are they the ones you are concerned about?

    128572 GUADALCANAL AAF -9.4333 160.0500 10.100 -9.9999 -9.9999 -9.999 Solomon Islands ZZ NA NA NA 80601 0 0 4
    43 128573 HONIARA/NTF/AWS -9.4330 159.9670 0.000 -9.9999 -9.9999 -9.999 Solomon Islands NA NA NA NA 0 0 1
    44 128575 HONIARA, BRITISH SOLOMONS -9.4170 159.9690 56.200 -9.9999 -9.9999 -9.999 Solomon Islands NA 91517 NA NA 0 1
    45 128576 HONIARA/HENDERSON -9.4170 160.0667 9.000

    Let me tell you how I found them

    I will suggest you look at

    http://www.climateaudit.info/data/station/berkeley/

    Start here

    http://www.climateaudit.info/data/station/berkeley/details.tab

    download it. Its an R object

    load(details)

    Then pick up this

    http://www.climateaudit.info/data/station/berkeley/other_unique_ids.txt

    Thats a file of all the NON RURAL stations

    Got that.

    Then you match the ids in ‘details” with the station numbers in
    “other_unique_ids.txt”

    that will give you our “not rural stations”

    When after you do that just search for stations between -9S and -10 S

    Well, that result is shown above.

    In short, the place you were concerned about was classified as not very rural.

    Bottomline: I promise that people WILL find sites that are misclassified. One of the things that Zeke and I did was a sensitivity on the classification schemes. basically they are not very sensitive to errors and you WILL have classification errors. However, this particular site was classified as not very rural. Now please nobody claim that its rural.

  140. Steven Mosher says:

    Reg Nelson says:
    April 4, 2013 at 5:38 pm
    Steven Mosher says:
    April 4, 2013 at 3:43 pm

    @Geoff Sherrington

    “Now THATS more like it. a man who has done some work.

    A few questions and I will have a look at it.

    1. What methodology did you use for classifying pristine?”
    ——————

    Hang on a tick, Steven. You just said in an earlier post that after exhaustive work you determined that station selection\exclusion criteria didn’t change the trend.

    Now, you turn around a few minutes later and ask Geoff what his selection criteria was. Seriously?

    You already answered your own question, The answer is: “It doesn’t matter”.

    #####################################################

    Perhaps, you misunderstood. I am interesting in what criteria he used for two reasons

    1. so I can check his work
    2. because I could be wrong and if he has a criteria I havent tested, a feature I havent controlled for, then I want to know so I can do a better job.

    I’ve look at a TON of stuff. Not everything.. so I am asking him to tell me something to move the science forward. simples

  141. Bill H,

    The CRN record really isn’t long enough to tell us anything at this point about the accuracy of the existing HCN network. It is worth pointing out that the trend of both is pretty much identical over the period of overlapping coverage, however: http://rankexploits.com/musings/wp-content/uploads/2013/01/Screen-Shot-2013-01-16-at-10.37.51-AM.png

    While we still need more time to draw firm conclusions, it certainly doesn’t provide evidence of bias in HCN stations asofyet.

  142. Bill H says:

    As a interesting exercise to see what biases for a group of stations exists, I would take and set up 6 stations around the current one at equidistant points and well sited. Then note the differences over say a year or two in temps. Stations within 750-1,000 feet of the original and use strict quality control without changing the original stations routines. Use Anthony’s surface-station criteria for a 1st class station.

    It would be interesting to see how they all match up and how much bias the station has given wind directions and travel over roads runways etc. You could pick several airports for this experiment or any college campus where the original station is poorly sited yet there are good areas for well sited data collection points to be set up..

    Just a though on how to identify bias and use a existing criteria, which is in use, to evaluate station siting and quantify the bias.

  143. Steven Mosher says:

    Pamela Gray says:
    April 4, 2013 at 5:20 pm
    Calibration checks are a must. Did you do that? Did you take a good random sample of your rural sites and check them from the ground, close up and personal? Did you take a history of those random sites to check for changes in their settings? This is standard practice. Calibrate. EVERY TIME YOU TAKE DATA!

    ###########################################

    The modis data was in fact the most highly calibrated data we could find.

    One issue with that calibration is that it was done on larger cites. So, I did my own calibration focused on tiny towns. I then doubled checked with other data sources ( several) other sensors
    of course there will be “classification” errors some rural classified as urban, some urban as rural,

    If you make the screen for rural too tight you actually push rural sites into the urban class.. which can cause you to see no difference. Most folks dont realize that.

    What folks dont get that that UHI varies within a city. It caries by season, by latitude, by the rural surroundings. It doesnt happen on cloudy days or windy days.. when its all said and done
    its a small average signal.. swamped by other things.. like siting perhaps.

  144. Steven Mosher says:

    Power Engineer says:
    April 4, 2013 at 5:16 pm
    UHI exists even at non urban towns such as ours which has only 4500 people….hardly urban. In the last 50 years traffic has grown 10 fold. Trees are largely gone. Lawns have been paved to provide parking for all the houses that have been converted to doctors’ and lawyers’ offices. If a temp station had been located here, it would show warming due to urbanization even though the surrounding air hadn ‘t warmed at all.

    I agree that there is an “urban” definiton problem.

    #######################################

    Our “Rural sites” had fewer than 3 people per sq km within a 5minute ( 10km) grid cell
    majority had no people.

    we would classify your town as urban.

  145. Steven Mosher says:

    Doug Proctor says:
    April 4, 2013 at 3:54 pm
    The problem here is the same I see in industry these days: if you can’t dump and troll through datasets, then it ain’t either science or worth it. Manually looking at sites and making a decision is too “subjective” for science these days. Maybe because if you did this, you would have an “opinion”, but if it is just “data”, you are off the hook: hey! that is just what the program says!

    ###################################

    well one of the reasons why we move to an objective basis is because for MANY years the UHI community has used airports as a rural comparator. Oke and his students have tried to put it on a more objective measureable basis. repeatability you know.

    For Modis, the classification system does involve human input and refinement and calibration.

  146. John another says:

    I was kinda hoping for an explanation for the difference between Albany and NYC.

  147. Bill H says:

    Zeke Hausfather says:
    April 4, 2013 at 6:44 pm

    Bill H,

    The CRN record really isn’t long enough to tell us anything at this point about the accuracy of the existing HCN network. It is worth pointing out that the trend of both is pretty much identical over the period of overlapping coverage, however: http://rankexploits.com/musings/wp-content/uploads/2013/01/Screen-Shot-2013-01-16-at-10.37.51-AM.png

    While we still need more time to draw firm conclusions, it certainly doesn’t provide evidence of bias in HCN stations asofyet.
    ========================================

    While not being long enough currently the trend is flat, as in less than -0.1 deg C over the time period. This is in line with many of the Class 1 stations Anthony and others have categorized showing a declining temp trend. Only high density areas are showing an increase in temp.

    This leads me to believe that well sited stations are not showing the warming that high population density areas are. There are poor sited stations in both rural and high density areas. The key is to assess these one at a time and look at them given the growth changes vs Temp records. The problem is, there are not very many with close proximity, well sited stations with which to cross reference the temperatures and trends in an effort to see what the actual bias is.

  148. Reg Nelson says:

    Steven Mosher says:
    April 4, 2013 at 6:41 pm

    Perhaps, you misunderstood. I am interesting in what criteria he used for two reasons

    1. so I can check his work
    2. because I could be wrong and if he has a criteria I havent tested, a feature I havent controlled for, then I want to know so I can do a better job.

    _________

    No, I completely understood. Why would you need to check his work? Why not accept it on faith? That’s how climate science works, isn’t it?

    He showed the data in the spreadsheet. It’s all there. Have a look at it and try to replicate his work yourself — that’s how these things are done, right?

    And you ignored the second part of my post that pointed out your confirmation bias.

  149. Pamela Gray says:

    Steve, you know that your answer didn’t even touch what I asked. Did you or did you not choose a random sample of rural sensors and actually visually inspect them and the records kept for these stations? Yes or no. It’s called field research. Try it.

  150. pottereaton says:

    Steven Mosher says:
    April 4, 2013 at 3:43 pm
    ————————————-

    I was hoping you would say that . . .

  151. Pamela Gray says:

    People don’t warm stations, objects warm stations. And objects can be an overgrown tree or a BBQ next to the only house for miles around. As we have seen over and over again with photos, population density may easily have no bearing on microsite issues that create what is commonly called UHI. It is an unfortunate misnomer and leads one to think that UHI only relates to large towns and cities (IE urban). Without a random on the ground calibration of what your in the air choices were, the results could be meaningless.

  152. Steven Mosher says:

    Finally, I was curious why you didn’t use your technique to investigate things like the Oke paper or the McKitrick paper which found evidence of UHI. I’d think you could use your method and the Oke or McKitrick datasets to determine whether they were wrong or not and if so why … seems like if you’re going to implicitly claim the previous studies are wrong, you should show it directly using their data.

    #######################

    Ross’s paper is flawed because of gross data errors. Its his job to fix his errors. he is well aware of them.

    Let me tell you how he calculated population growth.

    Take the united states population in 1979: Now divide that population evenly into every 5 degree grid cell. yes Alaska and New York get the same density. DONK

    Now take the population in 2000. divide it in every cell evenly. DONK.

    basically he assume that population and population growth at a 5 degree resolution, uniformaly distributed is correct. DONK.

    he also made the mistake of putting 56 million people in antartica and that many on St helena.

    How? those are british stations so he just used UK population. DONK. if I did that you people would have me hung. I should check solomon islands for grins.

    So, folks may give me grief for 500 meter data.. ross created 5 degree by 5 degree data.
    DONK.

    WRT to Oke? oke abandoned his idea of tying temperature to population, said it was wrong and limited. plus it only refers to UHI MAX not average UHI. Plus his formulation changes depending on where you go on the globe. Science has moved on. HOWEVER, population does matter for computing the anthro heat portion of the town energy balance ( TEB) see Sailors work.

    I’ve also got some world wide numbers on anthro heat per sq km.

  153. Steven Mosher says:

    Paul Homewood says:
    April 4, 2013 at 12:33 pm
    I think one of the things Willis’s analysis shows is just how dangerous it is to rely too much on statistical wizardry and algorithms and not enough on proper local and practical knowlege.

    GHCN’s Icelandic adjustments are a classical example of this. Who knows better the temperature history of Iceland? The experienced, knowledgable Iceland Met Office, or some computer programmer in GHCN?
    ####################################

    However, in this case we did in fact classify the station as urban. Trust me you will find mistakes, this was not one

  154. Steven Mosher says:

    Reg Nelson says:
    April 4, 2013 at 10:08 am
    Muller would have saved himself a lot of wasted time and effort if would have just read this first:

    http://wattsupwiththat.com/2013/01/20/noaa-establishes-a-fact-about-station-siting-nighttime-temperatures-are-indeed-higher-closer-to-the-laboratory/

    #####################

    written afterwards I in fact found this and pointed Anthony at it.
    I will wait till its done

  155. Jeff Alberts says:

    For me the challenge is to convince people get off of the false “rural/urban” dichotomy entirely. There are terrible stations everywhere, being out in the country is no guarantee of good siting.

    Indeed: http://whatcatastrophe.com/drupal/surveying_olga_2

  156. Geoff Sherrington says:

    Steven Mosher says: April 4, 2013 at 6:41 pm
    Re : Criteria for Australian data posted above.
    Hi Steven, great to see the official appointment. I know that you handle tons of data because you handled this some time ago, unless I tweeted it and forgot. You have to assign priorities to your time and it’s easy you could have missed the significance.
    Criteria.
    1. Local knowledge. I spent a career in mineral exploration and there are few places in Australia where I have not been.
    2, Population. Varies a little, from a few seagulls per square mile to perhaps a dozen people at most in same sq mile.
    3. Google Earth. Grab a site, take a look. Look for evidence of people. If you find any, see how far away the weather station is. Usually more than a mile.
    4. Airports. If any. Not tarmac, just natural land with trees removed, some graded occasionally.Don’t think any take jets. At best, a light twin piston. Traffic probably a couple a week,
    5. AWS or manual – some AWS because no people reside there.
    6. Variety. Some by the seaside, some way inland. Some to the North, some to the South. Swing high, swing low. etc.
    7. Progression. I started with the most unequivocal sites and kept adding until I thought there was a vary faint possibility that a site could feel UHI. I was VERY conservative.

    Overall, do what Willis did. Have a look at Google Earth. Coordinates to 4 places after the decimal degree will usually pick up the view of the screen.
    Caveats. I do not consider it statistically valid to fit a linear least squares regression as I have done. It is eye candy, but it helps. Second, one has to assume that quality control in remote places can be questionable. However, the range of trends over the decades shows patterns in which some reality is probably embedded.

    If you think that Australia produces difficult results, try Antarctica. If you think Australia is hard, you have to widen your confidence bands.
    Happy to answer any questions. Cheers Geoff.

  157. Steven Mosher says:

    Anthony Watts says:
    April 4, 2013 at 9:19 am
    Thanks Willis. As I point out here, http://wattsupwiththat.com/2012/07/29/press-release-2/ the resolution of the study we did is down to 10 meters. Leroy 2010 sees 100 meters as the limit to the effects he observes relating to siting. Like Marcott et al, this is a case of a low resolution sample missing the “spikes”.

    Might be fun to see if we can link high temperature spikes at Henderson Field to flight arrival/departure times. The data must be there somewhere.
    #########################################################

    1 the field is classified as urban

    Also, I need to update you on my latest.

    using your first set of classifications and various satilltite products I am able to classify stations according to the CRN classification. Basically you use a subset of your classifications ( 1-5 ) to train a classifier and then test the classifier on the held out data.

    The reason that is important is that it will allow one to do the whole globe.

    So. Start with your hand classification: Train the classifier ( using say C4.5 or C5 or SVM) and then test the classifier on the held out data.. That gives you an accuracy figure then you apply to the whole shebang

    Interesting result. Impervious surface surrounding the site ( out to 500m) is statistically more important than the area immediately around the site.

    hmmm some helpful references in here about the physics

    http://www.instrumentalia.com.ar/pdf/Invernadero.pdf

    http://sourcedb.cas.cn/sourcedb_igsnrr_cas/yw/lw/200911/P020091104357019526094.pdf

  158. Reg Nelson says:

    Steven Mosher says:
    April 4, 2013 at 7:43 pm
    Reg Nelson says:
    April 4, 2013 at 10:08 am
    Muller would have saved himself a lot of wasted time and effort if would have just read this first:

    http://wattsupwiththat.com/2013/01/20/noaa-establishes-a-fact-about-station-siting-nighttime-temperatures-are-indeed-higher-closer-to-the-laboratory/

    #####################

    written afterwards I in fact found this and pointed Anthony at it.
    I will wait till its done.
    _______

    I totally agree. The proof is in the pudding. We should wait until these climate models have some predictive value before wasting anymore money on this rubbish. This will take some time, however, as the new trend is to say “may\could happen in 2050 or 2100″.

  159. TRBixler says:

    Fascinating that no one offers a check of a site. 5 meter grid with data logging temp sensors over say 100 meter square. Then compare the temperature represented by the test grid to the temperature sensor in the center of the grid (the temperature sensor being quality controlled). Of course I hear claims of accuracies to .001 but no knowledge if just a few meters away the is a .1 degree difference. Funny how the human body can “feel” those differences. A few meters one way or the other and a trend has been made.

  160. Geoff Sherrington says:

    Steven Mosher & others,
    Here is another page giving coordinates of the weather stations from quite a few of the pristine stations I selected from Australia.
    I culled from this list.
    It is usually best to use the lats and longs because the weather station can be some distance away from where you lob on Google Earth if you simply type in the name.
    http://www.geoffstuff.com/Pristine_157.xls
    There is one site that actually shows a screen in a tourist photo. Point Hicks, near the lighthouse keeper’s residence 1857. Select the photo half way down the black fence to the east. This might be the one site most susceptible to showing a man-made influence, but then that has to be traded against the professionalism of the recorder. There is a tradition of excellence among lighthouse keepers (but also the opposite if they went strange through loneliness). Also, the dominant wind direction is from S-W to W, lessening any effects of people.

    The data are from the Bureau of Meteorology, who have to be commended for a massive data compilation. Sometimes the Bureau does things that are hard to follow, but credit is due for these records. Thank you, BoM.

  161. Reg Nelson says:

    Steven Mosher says:
    April 4, 2013 at 7:57 pm

    “Also, I need to update you on my latest.

    using your first set of classifications and various satilltite products I am able to classify stations according to the CRN classification. Basically you use a subset of your classifications ( 1-5 ) to train a classifier and then test the classifier on the held out data.”

    ——

    Why don’t you just go out and observe what is actually happening in the real world? Conduct real, physical experiments and collect data. Climate Science isn’t cost intensive. How much does a decent, well-sited weather station network cost? Certainly less than what was blown on the Solyndra debacle.

  162. u.k.(us) says:

    Steven Mosher says:
    April 4, 2013 at 7:57 pm

    “hmmm some helpful references in here about the physics”
    ===============
    Yep, well understood I’m sure.
    Consensus, even?

  163. KuhnKat says:

    Steven Mosher:

    G I G O.

  164. Anton Eagle says:

    Mosher,

    if you agree that UHIs exist… and we can clearly see land becoming more urban over time, how could anyone objective argue that a developing UHI would not introduce a bias over time. To argue that there is no bias either makes one an idiot or a political hack… take your pick.

    The fact that you were not able to find the bias with lousy methods doesn’t prove anything.

  165. u.k.(us) says:

    In the olden days, it was 3 guys on top of a Saturn 5 rocket, nobody “mailed” it in.
    Now, all we get is mail.

  166. The BEST methodology doesn’t measure the amount of urbanization that has occurred at a location and that is the variable that matters. As Dr Spencer says,

    The fact that the greatest warming RATE is observed at the lowest population densities is not a new finding. My comment that the greatest amount of spurious warming might therefore occur at the rural (rather than urban) sites, as a couple of people pointed out, presumes that rural sites tend to increase in population over the years.

    In conclusion, the issue is whether increasing urbanization is influencing the surface temperature record, and this study tells us nothing useful in relation to that issue.

  167. Geoff Sherrington says:

    Mosh,
    Has your selection method, applied to Australia, overlapped with any of the ‘pristine’ sites I listed and if so, can you share the results? I selected them from a data sheet of several hundred Aust sites you sent me some time ago near the start of BEST. Ta Geoff.

  168. tobias says:

    Steve M says Powerengineers town of 4500 is urban based on Steve’s own qualification of 3 people/sq km being rural. As far as I can tell PE’s town might be 100.000 sq kms or it might be 1 sq km. Also our local airport used to be called a “rural” airport based on it’s connections, now it is an “International” airport based on its connections, same airport just a longer strip capable of handling larger aircraft in case of emergencies s.a. 911. Semantics can hide anything anyone would like to hide. This whole climate debate is sadly becoming a yelling match. Just tonight D. Suzuki blamed the pine beetle problem on AGW glossing over the natural facts that the pine tree population is 80 to 100 years old and is following a natural cycle documented by scientist and the local native populations as being thousands of years old.

  169. k scott denison says:

    Steven Mosher;
    Dude I’m still waiting for somebody to accept my challenge.

    1. Define Urban or rural ex ante in a way that is objectively measureable.
    2. i will divide stations into urban and rural per your definition.
    3. i will compute the difference.

    A cookie for anyone who can find the signal.

    So, there is the challenge.
    =======================
    “Urban” = any station where any man-made element has been added within a 1,000 yard radius of the station during the history of the tempertaure record.

    “Rural” = any station where there has been no new man-made element added within a 1,000 yard radius of the station during the history of the temperature record.

  170. Steven Mosher;
    Dude I’m still waiting for somebody to accept my challenge.

    1. Define Urban or rural ex ante in a way that is objectively measureable.

    Passing over that this a false dichotomy. An objective measure of urbanization is the surface area of manmade structures.

    Completely rural would have a score of zero, ie no manmade structures. An area where the surface area of manmade structures is several times that of the land area, would be fully urbanized.

    * While horizontal and vertical surfaces have different effects on temperatures, the value of their effects on temperatures are probably similar enough for them to be combined into a single measure.

  171. Willis Eschenbach says:

    Steven Mosher says:
    April 4, 2013 at 6:38 pm

    Well Willis, lets have a look at the data. You are concerned that this site will be Rural classified by our proceedure. Correct?

    Not in the slightest. I thought I was quite clear that the issue was not whether the site was classified as being rural or urban. Let me re-read the post … yeah, here’s what I said:

    And that’s why I didn’t expect they would find any difference … because their division into categories has little to do with the actual freedom of the station from human influences on the temperature. Urban vs Rural is not the issue. The real dichotomy is Well Sited vs Poorly Sited.

    You seem to think that the point is airports, or rural versus urban. It’s none of those.

    My issue is that the MODIS dataset, despite its outstanding resolution, cannot tell well sited surface stations from poorly sited stations.

    w.

  172. Geoff Sherrington says:

    Philip Bradley says: April 5, 2013 at 12:03 am
    I’d define rural as having a temperature trend that is essentially similar to the regional trend of other stations so defined.
    But this does not get us far. I’d like to see a selection of BEST USA sites selected as non-UHI to see if they have a ‘background slope’ if I could call it that.
    I’d look through the data myself except for some severe health problems here that are distracting.

  173. plazaeme says:

    So, let’s try to follow Willis rationale (which I like). Local Heat Island. Of course, there could also be Local Cooling Islands. But we can think, what would be more probable / abundant, cooling or heating islands, compared to areas not affected by human activity? I would say heating ones, and by far. And that makes the point for Willis idea on Local Heating Islands. The more human activity you have, the more heating (wherever). Because (I guess) there are far more human heating activities than human cooling activities. Unless I am wrong.

  174. Peter Plail says:

    As a result of this work by Mr Mosher and his colleagues, perhaps they would communicate their findings to the UK met Office and BBC and ask them to stop the practice of telling us that “of course, the temperatures that we show are for towns, rural temperatures will be several degrees lower”.

  175. richard verney says:

    Willis Eschenbach says:

    April 5, 2013 at 12:26 am
    //////////////////////////////////////////

    What dom the data sets say as to anomaly trends when only Class 1 sites ared used?

    Ditto, when only Class 2 sites are used?

    Ditto, when only Class 1 and 2 sites are used?

    With powerful computers, it should be easy to output these data series (at least as faer as US data is concerned).

    I would have thought that BEST would have looked at the classification of sites, and examined each catagory seperately and jointly to assess to what extent (if any) good siting makes a difference..

  176. Alexej Buergin says:

    “Gene Selkov says:
    April 4, 2013 at 2:02 pm
    maybe the values of V1 and V2 for your aircraft are based on the worst-case scenario at the worst airport where the machine is certified to land?
    Surely your Vy depends on altitude, quite a bit. According to this chart, at 40C, you’re already at 7500 feet above the ground, compared to standard atmosphere:”

    Since you mention small airplanes: The indicated airspeed is measured with a pitot tube, which measures the stagnation pressure (the difference to static pressure, the dynamic pressure, depends on velocity). It is very convenient that all this is proportional to air density, so the pilot can use the same INDICATED airspeed everywhere.

  177. Silver Ralph says:

    Gene Selkov says: April 4, 2013 at 2:02 pm
    ralfellis: maybe the values of V1 and V2 for your aircraft are based on the worst-case scenario at the worst airport where the machine is certified to land?
    Surely your V1 depends on altitude, quite a bit. According to this chart, at 40C, you’re already at 7500 feet above the ground, compared to standard atmosphere:
    =========================================

    V1, V2 and Vr change with every take off, and a new calculation has to be made every time. The main criteria are runway length, runway altitude, runway state (wet?), temperature, atmospheric pressure, weight of aircraft and power setting. We tend to reduce the power to the minimum required, to save the engines.

    However, the thing to note is that the temperature and altitude effect the engine thrust and runway requirements much more than they do the airspeed. The airspeed detector (a pitot) naturally compensates for much of the difference that altitude and temperature makes. Thinner air means lower airspeed detected, so you have to go faster, in terms of groundspeed, to make up the difference anyway. So the take-off airspeed changes very little with temperature alone – just a couple of knots up for the V1 and a couple of knots down for the Vr.

    But the greater groundspeed (not airspeed) caused by high altitude or high temperatures does mean that longer take off runs are required. You will find that high airports often have very long runways. And if you don’t have a long runway, you have to shed some weight.

    .
    .

    u.k.(us) says: April 4, 2013 at 2:02 pm
    I fear I’m in way over my head, but pilots need to know the temperature, because it affects the length of the take-off run, which limits the length of runway left to stop in case of an engine failure.
    ===================================

    True. As above, temperature and altitude effect engine thrust and runway requirements. Most engines are flat rated to 30 degrees c (capable of producing max power up to 30 degrees). Above that, the engine maxes out on internal temperature (gas temperature), not maximum power, and so thrust reduces.

    But Mosher was taking about temperature effecting the take-off airspeed, which is incorrect.

    .

  178. johnmarshall says:

    I live in a rural area in the UK. I am conducting a small experiment on temperature using a pair of identical digital thermometers relaying to a central instrument. One is 1m from the back wall of our house and the second 20m away in ”clear air”. Every day there is a discrepancy between readings of up to 3C. The sensor near the house giving the highest readings. Not unsuspected.

  179. Paul says:

    I think Steven Mosher is holding his own here. Not impressed by Willis renaming the UHI to avoid a challenge!

  180. samsonsviews says:

    I think Steven Mosher is holding his own here. Not impressed by Willis’s renaming the UHI to avoid a challenge.

    On the face of it, suggesting UHI hasn’t impacted the recorded global warming trend seems illogical. Seems to me the challenge is to identify the trend (it has to exist, right?) and quantifying it. However, my starting logic of the ‘UHI has to exist’ and my demanding evidence of its existence while refuting contrary evidence as illogical, seems to me the parallel of the AGW believer’s dogmatic belief in the mystical power of anthropogenic CO2. I mean that power has to exist, right?

    UHI is a storm in a teacup if you ask me. Real debate is on climate sensitivity & the role of feedbacks. The starting challenge would be to prove that the theory of a sensitive climate that responds to forcings is real..

  181. knr says:

    Willis Eschenbach
    ‘annot tell well sited surface stations from poorly sited stations.’

    Nothing can do this but physically checking the stations, the reasons this is not done is becasue of time and money , not becasue it would not improve the validity of the data , which it would, and so lead to improved respect for the data .

    The underlining problem is before AGW theory , people accepted that to some degree weather information would always be bit hit and miss and that forecasts could be wrong so poor sitting was not such a issue . But once this data became all important in a ‘political sense’ to back up the ‘settled ‘ science claims which in turn where used to make great demands for change and spending . Issues such has sitting ,which in the past was consider unfortunate but acceptable, came under review and their failings become much more important.

  182. Alan Millar says:

    Mosher

    Can you please answer the following question to establish for certain what you are trying to say.

    Do you believe that urban areas are hotter on average than rural areas in the same local, yes or no?

    I am going to assume that your answer is yes, as that is the absolute, undeniable truth established by direct observation millions of time a day.

    Now the trend in measured temperature between these areas, should all other things being equal follow the generic cooling/warming of that locale and there should be little if any differenc

    I think that is what you are triyng to say and I agree.

    However, the question I am interested in is….. ‘is there a trend for rural areas to become more urbanised, whilst retaining the definition of rural? Also is there a trend for urban areas to become even more ‘urban’ over time increasing the actual measured temperatures and introducing an artificial trend outwith the generic temperature trend?

    Anthony’s study and anecdotal evidence seems to suggest that such a thing has happened. The only way to check this is to examine the actual site and current placing of the sensors and compare that to the past by whatever means is possible.

    Without this information you and everyone else is guessing. You cannot however deny that if this extra ‘urbanisation’ of the measuring sites has been occuring, then an artificial bias has been introduced into the temperature record.

    Alan

  183. Solomon Green says:

    Steven Mosher
    “Dude I’m still waiting for somebody to accept my challenge.

    1. Define Urban or rural ex ante in a way that is objectively measureable.
    2. i will divide stations into urban and rural per your definition.
    3. i will compute the difference.”

    Try “rural” = no buildings or tarmac roads exceeing 3% of area.
    “urban” = more than 50% covered by buildings and or tarmac.

    Then go and explore, say, Brazilia, Heathrow, Milton Keynes, Mokhatam Hills, Bat Yam, Eilat, Dubai airport and hinterland etc etc etc.

    All of these would fit the definition “rural” seventy years ago and several of these much more recently. They would all qualify as “urban” today.

    But, of course, none of these had thermometers located in their grids all those years ago so there can be no true comparison.

    But satellite data has been around for 30 years or more, perhaps Mr. Mosher could look at those grids that were definitely “rural” when the first data became available but are now “urban” accoridng to the latest data and then confirm that (1) there is no such thing as UHI and (2) that the temperature anomalies for urban areas correlate directly with those of rural areas. I suspect that the latter could be true but the former certainly is not and any interpretation of the data which suggests that it is true must be erroneous.

  184. Scott says:

    Alan Millar says:
    April 5 2013 at 4:55 am
    “Is there a trend for urban areas to become even more ‘urban’ over time …”

    Around here in my semi-urban area, I’ve noticed homeowners are constantly installing groundwater drain tiles to quickly remove surface water after a rain and route it to the storm drains. If they aren’t installing drain tile they are landscaping and sloping the ground to do the same thing. No one wants a part of their yard soggy a day after a rain. Same thing with golf courses and parks, if there is one project that’s a year round constant its surface water removal projects. So over time things dry out much quicker and temperatures rise faster but you can’t really see this urbanization from a satellite.

  185. beng says:

    C’mon, Mosher. If urban areas are demonstrably warmer than surrounding rural areas, then there has to be a greater trend for that site than the rural sites, assuming the urban site was less urban/more similar to the rural sites back in ~1880 or whatever. Yes, there could be a few exceptions depending on the site-specifics, but I’m talking in general.

    If you can’t find it, it means your measuring system/analysis techniques aren’t up to snuff.

  186. Pamela Gray says:

    Uh. No he is not holding his own. He has avoided my question. Did he or did he not calibrate his selection method with a random on-the-ground long-term quality check of stations classified as rural? Rural stations can be impaired over time by vegetative growth, or worse, poorly sited in terms of long-term lake, ocean, and atmospheric oscillation temperature trend effects that can be mistaken for AGW. I so no evidence of calibration here. None.

  187. Stephen Skinner says:

    Below are changes of a few things that can affect surface temperature:

    CO2 30% increase (of itself) (last 150 years)
    The atmosphere has changed by 0.01% (in terms of CO2 content) (last 150 years)
    Cities over 1 million increased by 40,000% (last 150 years)
    Global ground water depletion ∼4,500 km3 (last 100 years) and rate of depletion increasing
    Global grain production increase ∼40% (last 30 years)
    Global vehicle ownership increase ~650% (last 50 years) (corresponding road building/surface change?)

  188. Robuk says:

    I like this graph from Climate Audit, it`s raw I believe, could someone explain how if all the adjustments are done separately to each data set the temperature increase will be the same for both, ie no perceptible UHI signal between Rural and urban.

    http://s446.photobucket.com/user/bobclive/media/peters27.gif.html?sort=6&o=71

  189. Chuck Nolan says:

    knr says:
    April 5, 2013 at 3:25 am
    ……………..
    The underlining problem is before AGW theory , people accepted that to some degree weather information would always be bit hit and miss and that forecasts could be wrong so poor sitting was not such a issue .
    ———————————-
    You’re right because 200 years ago people didn’t try to connect all temperatures worldwide.
    They went through regional droughts, floods, hurricanes, tornadoes, blizzards etc so they knew every location throughout the world was somewhat different at any given moment.
    Now our extremely straight forward scientists and leaders claim to know:
    The entire world’s average temperature + or – 0.01 degrees.
    The average temperature of all the earth’s seas at all depths top to bottom + or – 0.01 degrees
    The average level of all the oceans in all the world to within + or – 0.1mm.
    The average rainfall throughout the whole world within + or – 0.01mm.
    Every puff of wind throughout everywhere on earth, instantaneously.
    Plus they know all this information exactly for billions of years.
    And they all swear life on earth will end and it’s all because of CO2.
    They lie, cheat, steal, adjust and lose data and then refuse access.
    After all this they call us deniers.

    Me? I’m more like Thomas, I doubt most of it.
    cn

  190. Mike Ozanne says:

    Well my prejudices date from doing tier 1 automotive compliance work. Pretty much a no-no to try and validate a measurement point using external criteria. All ours were individually classified by empirically examining both the instrument and its method of measurement, oh the joys of Gauge R&R.. Change the instrument (even for a replacement with an identical model) re-run the validation, Change the method, re run the validation.

    Want to establish the repeatability and reproducibility of these measurement points, go to each one and do the work. Is that a lot of work, yes, do your results have any value until its done , no.

  191. bobl says:

    Stephen has delivered some food for thought – and that is always a good thing. For example Arrhenius tells us that the radiative properties of CO2 in a test tube, a controlled closed environment make it a greenhouse gas. But Arrhenius tells us nothing about how CO2 behaves in a massive, chaotic open system like our atmosphere where it is mixed with massive amounts of constantly moving Nitrogen and oxygen, and even the trace gasses outnumber the poor CO2 molecules. Do we even know for example how much of that added energy is lost in thermal winds, or in the kinetic motion of raining out all that additional water?

    Urban heat islands, could be like that – logic tells us that increased urbanisation around a site SHOULD result in a trend. Stephen says he has failed to find it – I don’t think he has said it doesn’t exist. Willis suggests that siting is the issue and not urbanisation, and logically I think that’s right. The issue really is about sites that have changed in any way – a thermometer placed in a location that has had a tree grow over is will have a trend just as real as one which has has an airconditioner exhaust installed next to it. Indeed at a coastal site, a simple structure , grove of trees impeding the afternoon sea breeze is likely to introduce a trend. In this case land use changes and desertification may be very significant positive trend producers which clearly this study doesn’t address

    One further thing, Max-Min / 2 is a lousy average, the average temperature should include all measurements and be examined for temporal behaviour, if it turns out that all the 0.8 deg warming is occurring because 9AM temperature has increased 10 deg but every other time has cooled, that is quite a different diagnosis from a general increase in temp at all times of day. If only low daily temps are increasing and the hottr hours aren’t, which should in fact happen if humidity increases, what implication has that for the supposed Catastrophe – after all it can hardly be catastrophic if the mins are higher, but the maximums are less. Nor will it be a catastrophe if all the temperature rise occurs in winter, and none in summer. It also is not a catastrophe if all the warming occurs away from the population, other than needing a two brick high sea wall.

    This is my point when I say climate science needs to answer the question of whether Arrhenius actually applies in a massive, chaotic, open system and not only that, they need to be able to state exactly WHEN and WHERE it will be hotter before any protestations of the need for “Action” can be made.

    Stephen, other commenters have rightly pointed out that some of your statements do suggest a confirmation bias. Beware, I suggest you try to prove yourself wrong before you try to prove yourself right, and to do that you DO need to look at siting in detail.

  192. Theo Goodwin says:

    Absolutely brilliant article, Willis. Why is it that everyone who supports CAGW shows no interest in the empirical? Why do none of them want to actually investigate the physical surroundings of weather stations? Why will none of them use Anthony’s surfacestations.org website? Apparently, they suffer a phobia for the empirical.

  193. DCA says:

    I have been a land surveyor in Kansas for over 37 years. “Back in the days” in the mid to late 70’s, we used steel tapes, called chains, to make our horizontal measurements. We learned in college text that these chains were calibrated at 70 deg F and that for every 15 degs over or below 70 deg the chains would shrink or stretch 0.01 feet and the long measurements were corrected. In Kansas the temps range from 0 to 110 degrees depending on the season. The difference didn’t amount to much for a small parcel of land but for rural land, like a section (1 mile sq), the correction can be substantial.

    While working on highway projects we attached thermometers on the chains to aide us in making the measurement corrections due to the long measurement.. We would find as much as a 8 degree difference between asphalt or concrete vs a grass or cultivated field. Another thing I remember is that when measuring over cultivated fields (usually wheat), depending on the season a cultivated field would show differences likely due to ground cover. In most of the growing season the wheat fields were cooler while they were green and golden as well as more humid. However after harvest they got warmer by as much as 5 deg F perhaps because the darker colors of exposed dirt that absorbed heat as opposed to reflecting it.

    If a significant amount of land is cultivated, is this anthropologic influence significant or not?

  194. miker613 says:

    Steve Mosher, if you’re listening, could you explain what you mean by “there’s no signal”? Do you mean (a) Urban areas are not actually any warmer on average than rural areas [that would seem very surprising], or (b) they are, but there aren’t enough of them to make any noticeable overall difference.

    I’d also be interested in hearing your comment on something Steve McIntyre said long ago: It makes sense to use the satellite temperature data as the trusted baseline, and look for problems in the [far more complex] land-based temperature data if there is a discrepancy.

  195. Excellent information, people accepted that to some degree weather information would always be bit hit and miss and that forecasts could be wrong so poor sitting was not such a issue .

  196. Mark Bofill says:

    samsonsviews says:
    April 5, 2013 at 3:03 am

    I think Steven Mosher is holding his own here. Not impressed by Willis’s renaming the UHI to avoid a challenge.

    On the face of it, suggesting UHI hasn’t impacted the recorded global warming trend seems illogical. Seems to me the challenge is to identify the trend (it has to exist, right?) and quantifying it. However, my starting logic of the ‘UHI has to exist’ and my demanding evidence of its existence while refuting contrary evidence as illogical, seems to me the parallel of the AGW believer’s dogmatic belief in the mystical power of anthropogenic CO2. I mean that power has to exist, right?
    ——————
    I don’t think Willis is just renaming UHI. I think he’s made it clear that he thinks the urban / rural distinction isn’t the significant one. He says pretty clearly that poor siting (Is that a word? Poor station site-positioning I guess) is the problem. Going on to say he’s refusing contrary evidence to support a dogmatic belief doesn’t appear to be justified in my view. Or am I misunderstanding something?

  197. highflight56433 says:

    Whatever you refer UHI to, I travel to and from a city daily. In the travel I observe that the temperature in the vicinity of the local airport (w/weather station) and city is typically 5 to 9 F higher than the surrounding area that is pretty much unpopulated and well treed. And there is no change in altitude. The airport has no airline service to blame for extra heating. But the airport is of course devoid and altered from its natural surrounding. The question is: Did the same temperature differential occur before urbanization?

  198. An Inquirer says:

    In reply to Michael Cohen @ April 4, 2013 at 9:55 “I think it would be helpful for casual readers to emphasize that the paper is reporting no effect on the global land surface temperature trend from urban heating, not that there is no urban heating.”

    I had that same impression BEST’s initial positon. However, this paper has these words: “using the Berkeley Earth methodology applied to the full dataset and the rural subset; the difference of these is consistent with no urban heating effect.” They are saying that rural areas are warming at the same rate as the total data set — which would imply either no urban heating or that the rural areas are so pervasive in the total data set that they swamp the urban heating effect from urban areas.

  199. davidmhoffer says:

    1. Since we can easily measure the existence of UHI/LHI with equipment as rudimentary as a car’s dashboard temperature gauge, the only logical conclusion is that it exists. The question is why it isn’t showing up in the temperature data trend.

    2. Part of the problem is exactly what Willis suggested. The methodology cannot capture based on any definition of rural vs urban which weather stations have an artificial influence and which not. It has to be done station by station. But I think there are additional factors.

    3. How would a UHI/LHI effect appear? It would NOT, imho, appear as a trend in the first place. It would appear as a one time step function at the time the artificial influence came into being. Build a building. One time step function. Build a parking lot 10 years later. One time step function. Relocate the building’s air conditioning exhaust 10 years after that. Three step functions spread out over 3 or 4 decades. Then mix them in with thousands upon thousands of other weather stations. Some don’t have step functions, others do, but we don’t know which ones are which and they certainly aren’t correlated in time. So we know the UHI/LHI exists, but it exists as thousands of step functions smeared together in the temperature record. No shock that there is not trend.

    4. Not all step functions are going to be positive. Moving that building exhaust could produce a negative step function as easily as a positive one. My guess is that as urban influences gather, they are mostly positive, but at some point they most likely saturate. There’s only room for so many buildings, parking lots, air conditioners, etc, next to one weather station. As the effect becomes saturated, any additional changes are increasingly likely to be negative, and at some point I’m guessing a rough equilibrium would occur. Think of it like four buildings, one on each side of the weather station. Each one would have been a positive step function when it was built. But once you get to four, you’ve run out of sides, so you are maxed out. After a few decades, one building burns down. Negative step function, but it shows up right in the data where we are looking for a positive trend to prove UHI. I believe Watts et al 2012 may be demonstrating exactly this effect when you look at the poorest quality stations.

    5. My recollection from BEST a while ago was that 1/3 of stations show a cooling trend in their data. OK, how do we know that it wouldn’t have been 2/3 without the influence of UHI/LHI? Or maybe none? I just don’t see how you can take a UHI effect that grows via a handful of events at each weather station, spread out over decades, averaged with thousands of other weather stations with similar events but at different points in time, and calculate by ANY means what effect UHI has (or doesn’t have) on the over all trend.

  200. Willis Eschenbach says:

    Steven Mosher says:
    April 4, 2013 at 7:38 pm

    Finally, I was curious why you didn’t use your technique to investigate things like the Oke paper or the McKitrick paper which found evidence of UHI. I’d think you could use your method and the Oke or McKitrick datasets to determine whether they were wrong or not and if so why … seems like if you’re going to implicitly claim the previous studies are wrong, you should show it directly using their data.

    #######################

    Ross’s paper is flawed because of gross data errors. Its his job to fix his errors. he is well aware of them.

    Let me tell you how he calculated population growth.

    Take the united states population in 1979: Now divide that population evenly into every 5 degree grid cell. yes Alaska and New York get the same density. DONK

    Now take the population in 2000. divide it in every cell evenly. DONK.

    basically he assume that population and population growth at a 5 degree resolution, uniformaly distributed is correct. DONK.

    Despite the donking, all that you have described so far is that he did his analysis using country level populations … so what? Yes, it makes the analysis cruder, but it certainly doesn’t invalidate the analysis. Since the issue is population GROWTH and not the raw population, I fail to see the problem. As the authors point out, there are interesting results despite the broad brush approach:

    Population and GDP density varies considerably within countries, as well as between countries. Hence national averages will not capture all the important variations that may influence the temperature data. However, the trade-off we face is between encompassing the full range of variables we want to include versus matching the grids of measurement of climatic and economic data. Since national governments bear primary responsibility for climate data collection, the nationally-defined economic measures will capture important information about the availability of resources to monitor the whole country’s climate. Also, the substantial variation among countries implies that some of the effects of interest are definitely measured by the data we have available, albeit at a more coarse resolution than we would like. In the concluding section we will discuss the possibilities for future research arising from the development of some new socioeconomic data bases at the gridcell level.

    Your next issue is a couple of errors …

    he also made the mistake of putting 56 million people in antartica and that many on St helena.

    How? those are british stations so he just used UK population. DONK. if I did that you people would have me hung. I should check solomon islands for grins.

    And you’re telling me that results for Antarctica somehow invalidate his underlying thesis? That’s a procedural issue, and one that he dealt with in the paper:

    11 cells are in Antarctica, where there is no economy to speak of, several countries share jurisdiction over different research sites, and there is an anomalously high rate of missing values, probably due to the extreme conditions in which data are collected, so these were also removed.

    So that’s Antarctica, and no, I don’t care about the island of St. Helena. I’m sure it’s wrong, all results are wrong, … but by how much? I say again, the study is looking at population GROWTH, not population, so I doubt that the growth figures for the St. Helena population are that far off … but if they are, so what? Throw it out? How does that affect his results?

    So, folks may give me grief for 500 meter data.. ross created 5 degree by 5 degree data.
    DONK.

    A coarse grained study is somehow without value? Say what? The HadCRUT results are on a 5X5 grid, and yet they’re used every day.

    Steven, that is as shameless a job of study assassination as was done by the IPCC when they gave a bunch of trivial, bogus excuses (but no less trivial or bogus than yours) for throwing out Ross’s work in the TAR. All you’ve identified so far are meaningless issues that do not affect the strength or the significance of the results. Yes, it is a large-grained study, done on a country level … so what? It still gives significant information.

    There are valid country-level studies done every day, Steven. And as the authors of this one point out, they found significant results at the country level, so you can’t just turn up your nose and claim that it’s not good enough. They find valid and significant results at the country level, so they must be doing something right. Your insistence that valid and significant results can be ignored simply because they are at a country level, DONK, just makes you look as stubborn as a … aw, never mind.

    Yes, the study should be redone at a more fine-grained level to give better answers, and no, that doesn’t make the current study useless or wrong.

    I sure hope you have more issues with the McKitrick study than you’ve listed so far, because if that’s all you got, then it sure looks like you’re just invalidating the study based on meaningless objections, and not seriously analyzing their results. The McKitrick study may have serious problems, but if all it’s got is are the problems you’ve identified above, then there’s no serious problems at all.

    w.

  201. k scott denison says:

    Assuming that Mr. Mosher will return, I think them crux is this:

    The issue isn’t urban/rural. The money question is is there a difference in trends between sites where no significant addition of man-made structures has occurred during the period of the temperature record versus stations where significant additions have occurred. I would suggest the metric for stratifying stations is something like a percentage change in the total man-made surface area within a radius of X meters of the station. Perhaps even stratifying into multiple classifications based on those where changes have been made within a narrow radius of Y versus 2Y, 3Y, etc.

    This type of metric would certainly be more accurate as it is well known, as Willis and Anthony have pointed out several times, that the most likely cause of inaccuracies is what is very local to the station, versus what is in the general vicinity. Have to look at what was literally next to the station at the beginning of the record and what has been built (looking first next to the station and then in concentric rings) around the station over time to see if that affects the trend.

    Maybe this has been done and I’m simply unaware of the results. If I were to do the analysis I would start with a small set of stations where I know changes have and have not occurred. Do the analysis as a percent of man-made surface area within 100 meter circles by year, plot it all up, analyze and see what it tells me. Wish I had the data, time and skills to do so.

  202. k scott denison says:

    BTW, occurs to me how this thread is, like the Nancy Green thread, also about the lack of resolution in data leading to improper/misleading conclusions. Seems to me the lack of resolution in terms of changes around stations is the issue here, and none of the analyses I’ve seen from BEST resolve this issue as the methods for classifying stations don’t have the resolution needed to detect the influence of man-mad structures on temperature stations. In essence, the BEST approach uses proxies for the changes in influence, when in fact the actual data should be available, albeit onerous to capture and analyze as it would have to be done manually for each site.

  203. Willis Eschenbach says:

    Paul says:
    April 5, 2013 at 2:53 am

    I think Steven Mosher is holding his own here. Not impressed by Willis renaming the UHI to avoid a challenge!

    Say what? I said the UHI was wrongly named because the effect is not essentially urban. I said it should be called the LHI, the local heat island, because it was a LOCAL effect due to LOCAL issues.

    Had nothing to do with “avoiding a challenge”, Paul, that’s all your imagination.

    Regards,

    w.

  204. k scott denison says:

    Seems to me we should know the answer to this question: “What is the threshold surface area of asphalt (or concrete or brick or…) that can be built within what distance of a temperature station without affecting the temperature measurement?”

    Could be done experimentally, no?

  205. miker613 says:

    Steve Mosher is holding his own in the obvious way. Everyone is saying studies should be done. He claims he has already done one, analyzed using many of the variables that everyone is suggesting, and that he got nothing. Fair enough. He is offering to repeat the studies using other people’s filters. Also fair enough. Until his work is scrutinized, I don’t see what else there is to say about it.

  206. k scott denison says:

    Well, miker613, Mr. Mosher has analyzed data, but not in a way that approaches the real issue: is there a trend difference between stations where no local man-made objects have been constructed versus those where they have, over the length of the temperature record.

    What Mr. Mosher has done is a different analysis, one without the resolution needed to answer this question.

  207. johanna says:

    This thread has been very educational about lots of things, including the rationale for airport weather stations (thanks, pilots).

    Anthony and Willis and commenters have done a great job over the last few years of educating readers about the ins and outs of measurement, which looks simple on the surface, but is in fact quite complicated. This reader has learned a lot, and hopefully others have, too.

    WUWT is a free university, minus boring lecturers and having to sit in an uncomfortable seat frantically scribbling notes.

  208. miker613 says:

    Well, he has made an offer: Present a filter, usable by machine (as opposed to Go there and check by hand), and he’ll try it. Sounds like he’s tried a number of them on his own.

  209. clipe says:

    Aircraft movements for Nadi, Fiji (NAN)

    http://www.ats.com.fj/mms.aspx

    Can’t find anything on HIR but haven’t really looked hard.

  210. k scott denison says:

    miker613 says:
    April 5, 2013 at 12:02 pm
    Well, he has made an offer: Present a filter, usable by machine (as opposed to Go there and check by hand), and he’ll try it. Sounds like he’s tried a number of them on his own.
    ——–
    I have presented one. Usable by machine is a convenient, but not appropriate, condition. Doing what is accurate may not be convenient. If you have a database that lists construction by location by year please let us know. Lacking that, the work Anthony has done to document the environment around stations is the best I’ve seen to date. All done through direct observation, not by proxy.

  211. Beta Blocker says:

    samsonsviews says April 5, 2013 at 3:03 am
    …… UHI is a storm in a teacup if you ask me. Real debate is on climate sensitivity & the role of feedbacks. The starting challenge would be to prove that the theory of a sensitive climate that responds to forcings is real..

    Yes, you are very definitely right about that …… all of these debates over UHI, and over what kinds of trends the most popular temperature indices are actually displaying, are all mere Kibbuke Theater.

    So what if Global Mean Temperature is generally rising faster than it did between 1940 and 1975? So what if there is a recent pause in the post 1975 upward trend? (So what if there isn’t actually any such pause?)

    The historical evidence is that temperature trends can accelerate or decelerate very quickly over short time frames, both in the presence of an upwardly trending CO2 concentration, but also in the absence of an upwardly trending CO2 concentration.

    As long as there remains no long-term declining trend in GMT which is statistically significant and which continues for 30 to 50 years, the climate science community will continue to attribute all upward trends in GMT to the effects of man-made GHGs. (“Nothing else explains it.”)

    The only thing that matters is if those in positions of political power and influence pay any real attention to what the climate scientists are saying, or they don’t.

  212. miker613 says:

    “Usable by machine is a convenient, but not appropriate, condition. Doing what is accurate may not be convenient.” It’s always nice to have good data, but usually we have to go to war(k) with the data we’ve got. Are you claiming that there is no way to use all the metadata we have to detect the signal? That seems to be Mosher’s claim – but of course it pushes in the direction that there is no signal.

  213. clipe says:

    Steven Mosher says:
    April 4, 2013 at 12:18 pm

    [...] the temperature falloff for a jet plume is rather dramatic. I’ll refer you the ground handling safety guidelines published by airline manufacturers. Just request them or you can find a few of them on the web. Also, if the blowing of exhaust onto thermometers was a regular occurance it would be visible in the one minute data from airports. Its not. if it was persistent then you’d see it by comparing CRN data to nearby airports. Again, nothing.

    This isnt to say that it cannot happen, only that the conditions are rare and not easily found in the actually data.

    So

    1. the plume falls off rapidily.
    2. if it were a common occurance it would show up in 1 minute data. it doesnt.
    3. it doesnt show up comparing the best (CRN) to nearby airports.[...]

    The “ground handling safety guidelines” are just that – guidelines. I always give idling aircraft a wider berth when downwind of the “plume”.
    35 years on the ramp at YYZ teaches a person a lot about wind direction as it applies to heat plumes.

  214. clipe says:

    oops emphasis mine

  215. k scott denison says:

    miker613 says:
    April 5, 2013 at 1:26 pm
    “Usable by machine is a convenient, but not appropriate, condition. Doing what is accurate may not be convenient.” It’s always nice to have good data, but usually we have to go to war(k) with the data we’ve got. Are you claiming that there is no way to use all the metadata we have to detect the signal? That seems to be Mosher’s claim – but of course it pushes in the direction that there is no signal.
    ———–
    What I am saying is I have not seen anyone show that the meta data, as used in the BEST process, relates to how one would expect to see temperature behave at stations where there were no man-mad structures before time T0 and man-made structures after T0. Have you seen this data?

    Or to ask an even simpler question, what % do the stations classified in BEST as “urban” had a structure built within 1000 m during their tempertaure record? What % of “rural” stations? Are trends different for those where a structure was built versus those where no structure was built?

  216. k scott denison says:

    Or, more basically mike613, should one expect that the trend would be different at a station where no changes were made versus one where an asphalt pad was built next to the station?

    I would certainly expect there to be a difference. If I did the work and found none I would remain skeptical and expand my analysis. If I did the work and found a difference I would remain skeptical and expand the analysis.

    Another way to approach it is to look for stations with very high trends and very low trends , dig into the meta data, visually observe the stations, etc. searching for any relationship between siting and trend.

    If I remember right, there were ~30% of stations that showed cooling in the BEST data. What do these have in common? Are there differences in siting between these and those that show the strongest warming trends?

  217. A. Scott says:

    Mosher/Willis … please correct me if I’m wrong, but, if I recall correctly – from what I;ve learned from past discussions on UHI, for some here an important point is missed in the discussion about UHI. That is that there are really two different issues involved with UHI.

    One is the actual Urban effect on observed temperatures. We all know, and most have experienced, the fact that temps are higher as we move from rural into urban areas. This is clear proof there clearly IS an UHI effect on actual temperatures.

    That we all know UHI is real as it affects real temperatures, makes it hard to reconcile with and understand the other context of the discussion – whether those UHI areas affect the area or global temperature trends.

    Common sense says heck yes they do. If you drop an ice cube in a bowl of water, the temp of the water will decrease. If you drop a hot object in the same bowl of water, the water temp will increase. In that context I struggled with the question how could UHI not affect mean temperature trends?

    The difference is when talking about temp trends, those all sciency type folks are talking about <anomalies and not actual temps.

    Anomalies measure the temperature trend as a deviation from a standardized average base temperature period. The simple explanation is that actual temperature varies greatly due to many influences – elevation, population, siting issues, and the like.

    Using the measured temp difference from each sites own standardized base period allows a comparison of the temp TREND – regardless of different site characteristics. Simplistically, using anomalies converts the apples, oranges, limes, pears, peaches, and lemons etc. of the disparate station’s data, into all “apples” … by comparing only the CHANGE in temps relative to a standard baseline average.

    For purposes of mean temperature change, or trend, we want to know how much each stations temps changed over time, not what the actual temp was. Using anomalies creates a simple, standard method to compare differences (or similarities) between disparate sites.

    To apply that to the UHI discussion here – while its true from a temperature standpoint that Urban areas DO exhibit warmer temps, from a temp change standpoint, an urban site and nearby rural site are very likely to show then same or very similar change in any given year. If you have a warm summer in an area it is highly likely to be similarly warm – a similar amount above average – for both a rural and a nearby urban site.

    That is what Mosher, Zeke or others are talking about in these discussions – that the temp trend at an Urban site is usually little different than the trend for the same period at a nearby rural site.

    It confused me for many years – then one day a light bulb went off and I understood the difference.

    So it makes perfect sense that while there certainly IS an Urban Heat Island effect when it comes to instant temperatures, it also makes perfect sense that the CHANGE in temps during a fixed period between nearby urban and rural sites would be similar.

    Thus the statement that guys like Mosher, Zeke et al find no UHI effect when analyzing global mean temperature change – which uses anomalies – does make perfect sense. Since nearby rural and urban sites temps change form period to period relatively uniformly, it does make sense there is no UHI effect found on mean temp change.

    Again – if I got any of that wrong, Mosher, Zeke etc. please correct.

    As an aside, to complicate matters more, the base period is not always the same for different data sets. World Meteorological Organization policy suggests using the latest decade for the 30-year average. But even a quick Google search shows different data sets using differing base periods.

    http://www.ncdc.noaa.gov/sotc/global/#introduction 1981-2010
    http://www.ncdc.noaa.gov/ghcnm/maps.php 1961-1990 (1971-2000 for SST)
    http://data.giss.nasa.gov/gistemp/ Table Data: 1951-1980 means

    This does not cause problems per se, as long as you are comparing data within the same set, however it can become highly important if you were to try and compare dtat between two different data sets.

    NCDC explanation on using anomalies:

    Why use temperature anomalies (departure from average) and not absolute temperature measurements?

    Absolute estimates of global average surface temperature are difficult to compile for several reasons. Some regions have few temperature measurement stations (e.g., the Sahara Desert) and interpolation must be made over large, data-sparse regions. In mountainous areas, most observations come from the inhabited valleys, so the effect of elevation on a region’s average temperature must be considered as well. For example, a summer month over an area may be cooler than average, both at a mountain top and in a nearby valley, but the absolute temperatures will be quite different at the two locations. The use of anomalies in this case will show that temperatures for both locations were below average.

    Using reference values computed on smaller [more local] scales over the same time period establishes a baseline from which anomalies are calculated. This effectively normalizes the data so they can be compared and combined to more accurately represent temperature patterns with respect to what is normal for different places within a region.

    For these reasons, large-area summaries incorporate anomalies, not the temperature itself. Anomalies more accurately describe climate variability over larger areas than absolute temperatures do, and they give a frame of reference that allows more meaningful comparisons between locations and more accurate calculations of temperature trends.

  218. A. Scott says:

    Several folks above have touched on the the issue that is IMO relevant and should be addressed.

    While in the short term an urban and nearby rural location are likely to experience very similar changes on an annual or similar basis, the overall discussion is about long term change. And I do not see how long term UHI effect is addressed in any of these studies.

    Over the period of years relevant to the climate change discussion (say late 1800′s to present, or even 1920 to present where we have decent temp records) urban areas have experienced significant growth.

    The premise that using anomalies allows a standard compare between two nearby urban and rural sites is correct ONLY to the extent those sites remain static – unchanged over the time period reviewed.

    But we know that is NOT the case for most. Many sites that were rural 100 years ago are still similarly rural. However, virtually all urban sites have seen huge growth and expansion during that same 100 year period.

    Because we DO know, from an actual temp standpoint, that UHI is definitively real, we know beyond doubt that the actual temperature has increased over this longer time period. That also means the deviation from the base average has also increased over this long time as urbanization has occurred.

    As the rural sites have remained physically largely the same over the last 100 years, and the urban areas have grown dramatically – causing their actual temps to increase significantly – statistically there MUST be a resultant UHI influenced increase in mean temperature change, in the overall deviation from the standard average base temps, as a result of this accelerated urban divergence.

    It could well be, with so much of the globe still rural (or oceans etc) that the signal, even despite massive urbanization the last 100 years, is too small to identify. Even that though seems highly unlikely with the amount and degree of urbanization over the last century.

    Using a smaller area with significant urbanization – say continental US – there should be a clear UHI signal. It would seem running the trends analysis on the rural and UHI stations separately for the last 100 years or so, then comparing the temp change anomalies of the two, should show Urban temp anomalies increasing faster than rural during this longer term period.

    That demonstrates with certainty in my opinion a UHI effect does exist. The question then is why it does not readily appear in the overall review. Either it is too small to be readily apparent – which seems unlikely – or there are other issues in play masking the accelerated UHI anomalies

    Mosher, Zeke, Willis … etc. – please take a look and comment …

  219. k scott denison says:

    A. Scott says:
    April 5, 2013 at 2:58 pm
    Mosher/Willis …
    ———-
    Your post is true if and only if there is no change in the vicinity of the temperature station over the history of the record. And it’s not about going from rural to urban, it’s about whether anyone put a heat source or sink nearby, for example a building, parking lot, AC exhaust, etc. at some time during the station history. How many stations do you think have been in the same environment for their entire history?

  220. A. Scott says:

    You could drill down even further to demonstrate there IS an UHI effect on mean temp trends by selecting any large city, and one, or better a handful of, nearby rural sites. Run the trend anomaly analysis on the rural sites, and separately on the urban city for the last 100 years and comapre them.

    Mosher, Zeke, Willis or … ? Can one of you plot this for us.

    One graph showing any large Urban city’s anomaly history for say last 100 years. Overlay the same anomaly history for a group of nearby representative rural sites.

    And maybe a similar paired graph on continental US (or similar).. last same graph on a global mean basis.

  221. KR says:

    A. Scott – I would wholly agree with you in terms of trends being a completely different matter than local temperature values – and I believe you have described the situation WRT anomalies very well. WRT the _changes_ in rural/urban nature, and analyzing or correcting for the effects of such changes on long term trends, there are several approaches:

    1) Use the metadata for the stations (NOAA and others) and reset the offsets when the site situation changes – consider the before/after periods to have different average offsets. This accounts for equipment and enclosure changes, although perhaps not nearby development.
    2) Breakpoint analysis (BEST), numerically looking at changes/shifts in one station that are inconsistent with those around it, on the assumption that entire groups of nearby stations don’t change at once – again, break the station data into two at that point. That can account for equipment and enclosure changes, local influences, whether a shading tree has been cut down, a nearby plot paved, etc. Interestingly, this matches the metadata corrections quite well, and perhaps indicates that local developments are not a huge influence.
    3) Mosher in his analysis took a _very_ stringent view of what a ‘rural’ station would be, with the reasonable expectation that areas that became urban remained urban. The “very rural” areas match the overall records, supporting the claim that UHI is not an issue with trends. And hence his requests for alternative methods of identifying rural areas (he having tried a number of them) as a crosscheck.

    Local heat influences are a different matter entirely from UHI. If a local influence exists, it may shift the baseline average for that station either up or down. Only a consistent bias in how LHI changes would affect long term trends, given the number of stations. I would note, however, that (a) the BEST method should identify LHI changes, and (b) since the metadata and breakpoint methods give essentially the same results, it would seem that local influences do not have a significant effect on trends.

  222. A. Scott says:

    k. scott dennison … your issue – station siting – is relevant, but completely different and separate from the UHI discussion here.

    It is also however, exactly why I suggested including a GROUP of nearby rural stations when comparing to an urban station – which would tend to filter and one station’s issues.

  223. k scott denison says:

    Here’s an idea: how about we look at individual station data and get it “right” before we rush to use surrounding stations to correct urban stations? How about we dig into the data before rushing to average and correct it? That is what is s compelling about the CET: it is a well understood record.

  224. Pamela Gray says:

    Have you ever seen a bank of tall trees all the same height and columnar in shape next to a farm complex? Why do farmers plant those trees like that? The use of such growing wind breaks can and do affect ground temperatures. So you see how vegetative growth and changes surrounding a rural station can affect its temperature trend over a long period of time. Rural stations situated near a large lake or reservoir can undergo similar changes in its environment. Rural stations located in the direct path of El Nino and La Nina long term oscillations can fluctuate wildly as opposed to one in the protective (and warm) surrounds of an urban site.

    Calibrate Mosher, calibrate. Go back and do what you should have done before sending the article to print. Do a random study of the two extremes you have chosen to study but at ground level and with the historical changing environment in mind. Impinging trees can be every bit as good at changing ground temperatures over time as a growing asphalt meadow.

    The greater question is this: With such a long list of co-authors, how come no one thought of this most basic quality control measure? Hmmmm?

  225. Bill H says:

    plazaeme says:
    April 5, 2013 at 1:13 am

    So, let’s try to follow Willis rationale (which I like). Local Heat Island. Of course, there could also be Local Cooling Islands. But we can think, what would be more probable / abundant, cooling or heating islands, compared to areas not affected by human activity? I would say heating ones, and by far. And that makes the point for Willis idea on Local Heating Islands. The more human activity you have, the more heating (wherever). Because (I guess) there are far more human heating activities than human cooling activities. Unless I am wrong.

    ======================================================

    Unfortunately you would be wrong.

    Natural status of the surroundings (or as close to natural as one can get) is the the base line. I do not see many people running their conditioners to cool the outside. Watering the lawn or close area might cause a drop if done in the heat of the day but beyond that if the thermometer is properly covered the ambient air temp will be steady.

    Station siting criteria are important because you need to limit external biases. This is something that escaped our scientific community along with its importance to policies which govern the populace. Using garbage data to drive agenda creates GARBAGE POLICIES and AGENDA.

  226. Scott says:
    April 5, 2013 at 5:59 am

    Around here in my semi-urban area, I’ve noticed homeowners are constantly installing groundwater drain tiles to quickly remove surface water after a rain and route it to the storm drains.

    There have been a number of large scale changes in agricultural practices over the last 70 years, one of which is field drainage in wetter areas like northern and western Europe.

    The implicit assumption that local anthropogenic influences on temperatures are restricted to urban areas is entirely false, and because agricultural areas are much larger in extent than urban areas, changes in agricultural practices likely have a larger effect on global average temperatures than urbanization.

    I have previously argue that temperatures measured at lighthouses would be the best uncontaminated global temperature change measure. As measurements only start when the lighthouse has been built, and agriculture and urbanization is rare in the vicinity of a lighthouse.

  227. Chuck Nolan says:

    The baseline needs to be established using only long term sites which are, stable in location and fit for purpose. My guess is they are few and far between.
    Since they insist the end of the world is resting on this data we gotta go look.
    And no fudging.
    cn

  228. Chuck Nolan says:

    Bill H says:
    April 5, 2013 at 5:30 pm
    plazaeme says:
    April 5, 2013 at 1:13 am
    —————–
    What about Palm Springs?
    Long term LCI effect?
    cn

  229. A. Scott says:

    k scott denison says:
    April 5, 2013 at 4:43 pm
    Here’s an idea: how about we look at individual station data and get it “right” before we rush to use surrounding stations to correct urban stations?

    Hmmm … I don’t seem to recall mentioning ANYTHING about correcting urban stations. At all.

    I DID suggest, in trying to look at the UHI trend effect between rural and urban, that we uses several rural stations compared to the urban station. This does nothing to correct the urban station in any way. It simply helps moderate any site specific issues with the rural ones.

  230. Bill H says:

    Chuck Nolan says:
    April 5, 2013 at 7:34 pm

    What about Palm Springs?
    Long term LCI effect?
    cn

    =========================

    Long term LCI? Nope! Natural variation due to natural occurrence.. the next question (or should i say first is which Palm Springs Florida or Nevada or California or…)

  231. Frank says:

    Steve Mosher made the following challenge:
    1. Define Urban or rural ex ante in a way that is objectively measureable.
    2. i will divide stations into urban and rural per your definition.
    3. i will compute the difference.
    A cookie for anyone who can find the signal.

    Steve, that’s a safe bet that skirts the real issue. UHI’s clearly exist and have the potential to bias surface temperature trends. So does poor station siting, which is sometimes included in discussion of UHI. However, biases are only introduced into the trend when the amount of UHI or the quality of siting CHANGES appreciably over the period in question. Looking at stations TODAY tells us little about how much CHANGE occurred in the past and when it occurred. Did you have a good, private laugh when you posed this challenge? The DIFFERENCE in trend between “urban” and “rural” stations doesn’t tell us much about UHI biases unless you’ve got a group of urban stations whose UHI has increased and you’ve successfully excluded anthropogenic biases from the rural group. That’s unlikely with the limited existing information. The trend from a carefully selected group of rural stations is meaningful. BEST should simply report only the trend from the most rural stations they can find and stop suggesting that the difference between their rural and urban groups proves that UHI isn’t important.

    I left some related ideas at your website.

  232. Jon says:

    “Say what? I said the UHI was wrongly named because the effect is not essentially urban. I said it should be called the LHI, the local heat island, because it was a LOCAL effect due to LOCAL issues.”
    Anthropogenic Heat Island?

  233. Juraj V. says:

    Is it such a problem, if WMO asks local staff to categorize their stations used for global data according to some rules? They can have a new categorization overnight.

  234. Pamela Gray says:

    Mosher says:
    “2. Part of the problem is exactly what Willis suggested. The methodology cannot capture based on any definition of rural vs urban which weather stations have an artificial influence and which not. It has to be done station by station. But I think there are additional factors.”

    Then you admit you did not do any kind of station calibration of your station selection methods with visual on the ground inspection so you could determine artificial (or natural vegatative) influence. Good to know.

    By your own on the record doubts, the report is therefore invalid and should be removed from publication for lack of basic attention to standard research methodology.

  235. Pamela Gray says:

    Mosher: My hunch, with all these science beagles reading this thread, is that someone will do this quality calibration for you (if you release the station names used in your study) unless you and your co-authors get off the couch and do it yourselves ASAP (and describe your “random” selection method of course). Don’t want to? Request for station names from the researchers anyone? Do you really need someone to do a FOIA on you?

  236. Lars P. says:

    It is enough to have a look at their picture 1 to see clearly the effect of UHI on temperature and trends.
    The Tokyo temperature would not have run the same way without the human population accumulating there.

    Some decades ago the UHI was a recognized effect on trend and climatologists were naming a logarithmic dependency between the city population and the temperature.
    What GISS does is to subtract 0.01 per century. This is laughable.
    The cities grew, human population increased from 1 billion to 7 billions in the time when Berkeley make their temperature graph. And of course this effect is mostly visible in the higher latitudes where it is colder, there the difference – the resulted UHI is greater.

    In their Fig 4 – trends for 70 years – one can very easily identify the urban agglomerations as red spots. Furthermore it is interesting to see that the warming is more in the North, where the UHI effect would be greater.

    Looking back at Fig 1, one can see that Tokyo temperature stabilize at the end. Well, the cities are not growing at the pace they were growing in the first half of the century, so the delta UHI is not increasing as such.
    And this is also the explanation for their negative trend after 1950 in cities. Europe, North America, Russia ceased to grow in population or grew in a moderate way.
    Also the UHI increase is logarithmic. The delta UHI is not as significant any more once the cities are already developed. The time of exponential population growth is gone, and with it also the exponential increase in measured temperatures.

    As a personal estimation I would think that at least half of their temperature trend comes from UHI and LHI. But it does not matter so much for the future, as we see the trend for cities gets “cooler and cooler” as the cities will grow less and the effect is logarithmic.
    Oh yes, soon they will try to adjust for negative UHI due to cities.

  237. Mark says:

    Gene Selkov says:

    You can estimate the magnitude of the effect using this paper outlining the temperature contours of jet exhaust at different throttle positions:
    http://www.boeing.com/commercial/airports/acaps/7471sec6.pdf

    This for a B747-100. More typical for Honiara International Airport would be a B737 or A320 (as shown in the pictures). These are much smaller aircraft and lower to the ground.

  238. Mark says:

    Curt says:

    Airport sites are particularly problematic, as the airports are often placed initially in rural areas where land is cheap and there are few neighbors to annoy. But there is an almost inevitable development of supporting infrastructure around the airport over the years — but since no one lives in this infrastructure, population estimates of UHI don’t catch this (though MODIS-type estimates could).

    It’s also possible to get changes in the patterns of traffic, both aircraft and ground vehicles, without the infrastructure changing much at all. e.g. more small aircraft replacing fewer large ones. With exhaust from an E-170 being more likely to affect a thermometer near a runway or taxiway than that from an A380. Simply due to the former being nearer to the ground than the latter.

  239. Pamela Gray says:

    Tick tock tick tock tick tock tick tock tick tock tick…

  240. Pamela Gray says:

    When I did my research on the auditory brainstem response to high frequency tone bursts, we used brand new electrodes with known impendence (listed on the bubble pachage they came in) when place on abraided human skin. Guess what? Didn’t care what the package said about the impedence value. Thanks to sage advice from an experienced lab worker and mentor, I calibrated those electrodes EACH AND EVERY TIME I used them so that I could compare package values with each-time-I-used-them values. Mosher?

  241. Pamela Gray says:

    “package” Pam “package”…damed Sunday morning Irish Cream Coffee

  242. Steven Mosher says:

    Pamela Gray says:
    April 6, 2013 at 9:05 am
    Mosher: My hunch, with all these science beagles reading this thread, is that someone will do this quality calibration for you (if you release the station names used in your study) unless you and your co-authors get off the couch and do it yourselves ASAP (and describe your “random” selection method of course). Don’t want to? Request for station names from the researchers anyone? Do you really need someone to do a FOIA on you?

    #################

    The stations were released long ago on climate audit. before I joined the party. So have at them

  243. Steven Mosher says:

    Chuck Nolan says:
    April 5, 2013 at 7:32 pm
    The baseline needs to be established using only long term sites which are, stable in location and fit for purpose. My guess is they are few and far between.
    Since they insist the end of the world is resting on this data we gotta go look.
    And no fudging.
    cn

    ###################################
    wrong. there are plenty of long term stations. See my posts

  244. Steven Mosher says:

    clipe

    “So

    1. the plume falls off rapidily.
    2. if it were a common occurance it would show up in 1 minute data. it doesnt.
    3. it doesnt show up comparing the best (CRN) to nearby airports.[...]

    The “ground handling safety guidelines” are just that – guidelines. I always give idling aircraft a wider berth when downwind of the “plume”.
    35 years on the ramp at YYZ teaches a person a lot about wind direction as it applies to heat plumes.

    #########################

    having spent some time on the ramp and tarmack then you know that the picture in the post above is meaningless without wind informaton

  245. Steven Mosher says:

    Paul Homewood says:
    April 4, 2013 at 12:29 pm
    According to Richard Muller, himself,

    “Urban areas are heavily overrepresented in the siting of temperature stations: less than 1% of the globe is urban but 27% of the Global Historical Climatology Network Monthly stations are located in cities with a population greater than 50000.”

    http://notalotofpeopleknowthat.wordpress.com/2011/10/23/mullers-problem-with-uhi/

    How many more would there be if they took the population down to, say, 5000? And excluded airports?

    If BEST really want to exclude UHI, let them ignore all of these sites completely and give us figures based on only on sites that have been guaranteed as reliable by the local Met Offices. If such guarantees cannot be given, that country’s sites should be ignored completely.

    ##############################################################

    The population of the very rural, as we pointed out in the paper had a median of 3 people per sq km.

    At some point I would expect people to understand that Anthony and Willis have diagnosed the real issue. not UHI but the siting. UHI is so varied, and rural as I have shown can be warmer than urban. Its all down to the siting

  246. Steven Mosher says:

    “So, even if Geoff’s finding are correct, they can still be dismissed because Australia is the “Bermuda Triangle” of climate? Is that what you are trying to imply? Seems like you’ve already come up with an excuse without even looking at Geoff’s work. Your confirmation bias is painfully obvious.
    ######################
    Geoff sent me his work a long time ago. I believed I asked the same questions and we got no where. your assumptions are wrong

  247. Steven Mosher says:

    richard verney says:
    April 5, 2013 at 1:20 am
    Willis Eschenbach says:

    April 5, 2013 at 12:26 am
    //////////////////////////////////////////

    What dom the data sets say as to anomaly trends when only Class 1 sites ared used?

    Ditto, when only Class 2 sites are used?

    Ditto, when only Class 1 and 2 sites are used?

    With powerful computers, it should be easy to output these data series (at least as faer as US data is concerned).

    I would have thought that BEST would have looked at the classification of sites, and examined each catagory seperately and jointly to assess to what extent (if any) good siting makes a difference..

    ##############################

    1. Anthony did this and found nothing.
    2. we used his dataset and found nothing.

    3. he now has a new dataset. Thats good because there are errors in his old one which are easily found using my methods. Perhaps I shall write that up

  248. Steven Mosher says:

    ‘Despite the donking, all that you have described so far is that he did his analysis using country level populations … so what? Yes, it makes the analysis cruder, but it certainly doesn’t invalidate the analysis. Since the issue is population GROWTH and not the raw population, I fail to see the problem. As the authors point out, there are interesting results despite the broad brush approach:”

    ##################

    now you sound like climate science Tm, saying that errors dont matter, when you havent tested whether they matter or not.

    “Yes, the study should be redone at a more fine-grained level to give better answers, and no, that doesn’t make the current study useless or wrong.

    I sure hope you have more issues with the McKitrick study than you’ve listed so far, because if that’s all you got, then it sure looks like you’re just invalidating the study based on meaningless objections, and not seriously analyzing their results. The McKitrick study may have serious problems, but if all it’s got is are the problems you’ve identified above, then there’s no serious problems at all.”

    ###################################################

    Look, the reviewers asked us to comment on Mckittrick. My answer is simple. There are gross assumptions in mckitrick about his key variables. There are gross data errors. I have made him aware of these errors some time ago and pointed him at better data. To date he and others have said, ‘Fix my errors for me. prove my errors matter.” Sounds so much like Mann I want to puke.
    So, I answered the reviewer. An analysis of the errors of Mckittrick are beyond the scope of this paper. I am under no obligation to prove his errors matter or dont matter. The errors prevent any meaningful comparison of the results, any meaningful comparison of the methods. Further, the whole regression is based on a non physical model and is dimensionally incorrect.
    That said, there is merit to looking at changes over time. We do that by selecting stations have have not changed ( rate of change zero ) and comparing to stations that have changed.

  249. Steven Mosher says:

    k scott denison says:
    April 5, 2013 at 10:22 am
    Assuming that Mr. Mosher will return, I think them crux is this:

    ############################################################

    Lets see. The main post seems to make a claim that this location would have been classified by us as very rural. Well, we classified it as Urban? Is there a correction to the main post or does this error not matter? we will see

    ####################################

    The issue isn’t urban/rural. The money question is is there a difference in trends between sites where no significant addition of man-made structures has occurred during the period of the temperature record versus stations where significant additions have occurred. I would suggest the metric for stratifying stations is something like a percentage change in the total man-made surface area within a radius of X meters of the station. Perhaps even stratifying into multiple classifications based on those where changes have been made within a narrow radius of Y versus 2Y, 3Y, etc.

    ######################################

    1. I can look at changes in man made surface between 1992 and 2006 or between 2001 and 2006.

    2. understand that if these changes happen “at once” change point analysis will find them
    OR the change is too small.

    3, Gradual small changes are the challenge. Like trees growing.
    #################################

    This type of metric would certainly be more accurate as it is well known, as Willis and Anthony have pointed out several times, that the most likely cause of inaccuracies is what is very local to the station, versus what is in the general vicinity. Have to look at what was literally next to the station at the beginning of the record and what has been built (looking first next to the station and then in concentric rings) around the station over time to see if that affects the trend.”

    1. The sensor does not see what is very close to it unless you have windless condtions.
    2, Within 500m is more relevant than within 30m

    ##################

    Maybe this has been done and I’m simply unaware of the results. If I were to do the analysis I would start with a small set of stations where I know changes have and have not occurred. Do the analysis as a percent of man-made surface area within 100 meter circles by year, plot it all up, analyze and see what it tells me. Wish I had the data, time and skills to do so.

    1 . the data exists.
    2. it doesnt take much skill. I do it.

  250. Steven Mosher says:

    “If I remember right, there were ~30% of stations that showed cooling in the BEST data. What do these have in common? Are there differences in siting between these and those that show the strongest warming trends?”

    See my analysis of cooling stations started on my blog.

  251. Steven Mosher says:

    Puppet_Master_Blaster_Master says:
    April 4, 2013 at 6:05 pm
    Obfuscation.

    The people at BEST are really trying hard to ‘smoke screne’ the world. Guess they will soon start offering online casino betting plans, bitcoin investment options and ‘life’ insurance just like Al Capone’s ‘companies’ did.
    #################################
    conspiratorial ideation

  252. Steven Mosher says:

    samsonsviews says:
    April 5, 2013 at 3:03 am
    I think Steven Mosher is holding his own here. Not impressed by Willis’s renaming the UHI to avoid a challenge.

    ##################################

    haha neither am I, especially when the title infers the opposite.
    And when the site he picked was actually classifed as poorly sited.

  253. Pamela Gray says:

    Let me be clear Mosher. You re-defined your selection method. So bring the list of stations your method designated extremely “rural” for the purposes of this present study. The list of all the stations is not what I am asking for and you know that. You obviously assume that the rural stations in the study are somehow devoid of spurious artifact. How did you make that determination without calibrating a randam sample. So let me ask AGAIN!!!! Did you randomely sample that rural list for station siting artifact? A simple yes or no would suffice. And I am sure as hell not going to do for you what you should have done and reported in your methodology section of the paper.

  254. Steven Mosher says:

    Pamela Gray says:
    April 7, 2013 at 10:26 am
    Let me be clear Mosher. You re-defined your selection method. So bring the list of stations your method designated extremely “rural” for the purposes of this present study.
    #############################
    already posted on climate audit and refered to above.

    The list of all the stations is not what I am asking for and you know that.

    ###################

    seriously I thought that is what you are asking for so I supplied it.

    “You obviously assume that the rural stations in the study are somehow devoid of spurious artifact. How did you make that determination without calibrating a randam sample. So let me ask AGAIN!!!! Did you randomely sample that rural list for station siting artifact? A simple yes or no would suffice. And I am sure as hell not going to do for you what you should have done and reported in your methodology section of the paper.”

    I’m unclear what you mean by “spurious artifact”

    When I went through the list on climate audit. I found artifacts. Others found artifacts.

    http://climateaudit.org/2011/12/20/berkeley-very-rural-data/#comment-317667

    http://climateaudit.org/2011/12/20/berkeley-very-rural-data/#comment-318011

    I suspect somebody alerted Berkeley and I was invited to attend meetings.

    The first artifact I found was due to the use of 0.1 degrees as a screen. This changes the distance of the screen as latitude changes. So, I figured out how many rural were mis classified.

    Next artifact I found was a latitude based artifact. This owing to the time that Modis was constructed ( feb ) so I applied a filter for that.

    Next artifact I was concerned about was the airport artifact ( see the squares willis draws )
    airports will show up as rural because of the 500 meter and 1km rule. So I applied a filter for that.

    Next artifact was the high density industrial artifact. Small areas with big industry but sun 1sq km. So I applied a filter for that

    Next artifact was the 1km issue willis mentioned. So I went back to modis Source data.

    So, I did not randomly sample the rural stations for artifacts. I exhaustively looked at them all.
    I diagnosed characteristic flaws of urban categorization.

    Using this improved filter I reclassified. It

    http://climateaudit.org/2011/12/20/berkeley-very-rural-data/#comment-317667

    I attended my first meeting at Berkeley. I explained how I had gone through the stations as classified and accounted for artifacts I had found. I gave them my “improved’ version.
    Not randomly sample, exhaustively audited.

    They ran the test with my stations….

    No change in the results. So, I walked in ready to demolish the findings. I had publically called out mistakes I found ( long before willis trots out a BOGUS sample of mis classificaton) and I got involved to improve things. My improvement FAILED to find any difference. That was about 6 months of work I put into trying to build a classifier that would handle every situation and artifact could find by looking at thousands of sites over the past 5 years.
    That said, my classifier is able to find errors in fall 2011 classification. do they matter?
    dunno.

  255. Steven Mosher says:

    Here is a trick test for folks still on this thread.

    How would you classify these stations. would you call them CRN1?

    [googlemaps https://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=%0943.10917,%09-76.10333&aq=&sll=40.720364,-114.0274&sspn=0.021955,0.052314&t=h&ie=UTF8&z=14&ll=43.10917,-76.10333&output=embed&w=425&h=350%5D

    [googlemaps https://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=%0932.01667,%09-81.13333&aq=&sll=43.10917,-76.10333&sspn=0.010574,0.026157&t=h&ie=UTF8&z=14&ll=32.01667,-81.13333&output=embed&w=425&h=350%5D

    [googlemaps https://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=40.72056,%09-114.03583&aq=&sll=32.01667,-81.13333&sspn=0.012281,0.026157&t=h&ie=UTF8&z=14&ll=40.72056,-114.03583&output=embed&w=425&h=350%5D

    If you wonder where these stations come from. Well, here is a hint.
    11 of 13 are airports.. CRN 1. but check out the 3 above.. hmm WUWT?

    Lat Lon COOP. CRN Station
    1 34.70417 -118.42750 42941 1 FAIRMONT
    2 38.81780 -102.36080 51564 1 CHEYENNE WELLS
    3 29.72583 -85.02056 80211 1 APALACHICOLA WSO AP
    4 32.01667 -81.13333 97847 1 SAVANNAH WSO AP
    5 46.42667 -105.88250 245690 1 MILES CITY FCWOS
    6 43.10917 -76.10333 308383 1 SYRACUSE WSO AP
    7 34.98944 -99.05250 344204 1 HOBART FAA AP
    8 44.84280 -117.80860 350412 1 BAKER FAA AP
    9 44.38139 -100.28556 396597 1 PIERRE FAA AP
    10 27.76667 -97.45000 412015 1 CORPUS CHRISTI WSO AP
    11 29.53330 -98.47000 417945 1 SAN ANTONIO WSFO
    12 40.72056 -114.03583 429382 1 WENDOVER AWOS
    13 41.31250 -105.67444 485415 1 LARAMIE AP

  256. A. Scott says:

    KR says: April 5, 2013 at 3:52 pm
    A. Scott – WRT the _changes_ in rural/urban nature, and analyzing or correcting for the effects of such changes on long term trends, there are several approaches:

    KR – you note that identifiable changes in site situation and short term local heat island (LHI) changes are simply adjusted for. Which, if clearly identifiable as to cause (or which show up as a one time “step” change as opposed to a change over time) , makes sense.

    What does not make sense is the long term trend changes being adjusted out similarly. The ongoing long term effects of urbanization and growth should be apparent in each urban temp record. We know beyond doubt from the data that urbanization causes increased temps.

    At one point both the urban site, and the rural sites around it, were all rural. As the urban site was built out and population increased, its average temps increased. The surrounding rural sites did not show the urbanization effect. Both Urban and rural sites should show any overall warming or cooling similarly.

    You cannot, and should not, adjust out the urbanization effect on the urban sites. It is an important part of the equation.

    And the difference in trend between rural and urban sites should be readily apparent over the last 100 years or so as many urban sites grew dramatically.

    That is why I suggested a simple test to start. Pick and urban site that we know experienced significant growth over the last 100 years. Then choose a group of adjacent rural sites that surround the urban site – the more available the better.

    Plot the temp anomaly trend for the rural sites over last 100 years as one graph, and overlay that on the same plot of the urban site they surround. To make it more interesting, also plot the raw uncorrected, un-adjusted data for these same sites as well.

    Now do the same for all rural sites vs all urban sites in the entire continental US – overlay the 100 year trends for each. And last, same thing for the world.

    This will help visualize the data and difference in trends for us common folks. And provide a foundation for taking the discussion to the next step.

  257. Steve Mosher, its good you came and addressed the points people made, but you didn’t address the major flaw in this paper

    UHI results from urbanization (horizontal and vertical) over time.

    Whether a site is rural, or not, or how rural it is, at any particular time is irrelevant.

Comments are closed.