A CA reader has brought to our attention an interesting new article by Pielke Sr et al, Unresolved Issues with the Assessment of Multi-Decadal Global Land-Surface Temperature Trends, which disucsses microsite problems. There is a lengthy roster of co-authors: Roger A. Pielke Sr., Christopher A. Davey, Dev Niyogi, Souleymane Fall, and Jesse Steinweg-Woods, Ken Hubbard and Xiaomao Lin, Ming Cai, Young-Kwon Lim, Hong Li, John Nielsen-Gammon, Kevin Gallo, Robert Hale, Rezaul Mahmood and Stuart Foster, Richard T. McNider, Peter Blanken. (Note: also see discussion at Pielke Sr blog here )
They observe:
The integrity of climatological observations is often compromised by poor environmental exposure of instruments. Examples of poor exposure are provided by three United States HCN (USHCN) sites in Kentucky (see figures for Greensburg [153430] (Figure 5), Leitchfield_2_N [154703] Figure 6), and Hopkinsville [153994] (Figure 7). In each case, a combination of anthropogenic (e.g., asphalt and concrete surfaces, buildings) and natural features (e.g., trees and shrubs, slopes) of the microscale environment create forcings that are not representative of the larger mesoscale environment.
Below is their Figure 7, shows another site, which is in grotesque non-compliance with WMo standards. Note what appears to be a barbecue conveniently located below the temperature sensor. So the problems are not limited to northern California. Obviously Karl and Hansen have not performed the most minimal due diligence to ensure that the USHCN sites that they rely on meet the WMO standards that they presume.
Figure 7 from Pielke et al: Hopkinsville.
The article mentions Peterson 2006, which apparently says that these grotesque problems don’t “matter” and a response from Pielke et al 2007, which says that they do. .
Peterson [2006] concluded that any biases associated with the poor siting in eastern Colorado, when adjusted, did not affect estimates of regional temperature trends. However, in a response to the Peterson article, Pielke et al. [2007] pointed out several issues which Peterson did not adequately investigate, including often undocumented station changes, ignored uncertainties in the adjustments, and land-use/land-cover change issues associated with climate station adjustments.
The PEterson reference is: Peterson, T.C., Examination of potential biases in air temperature caused by poor station locations. Bull. Amer. Meteor. Soc., 87, 1073-1089. 2006.
Climate Audit 
76 Comments
Lets see if we can score this:
0= perfect site
9= useless site
BBQ= 2 (produces a lot of heat, proximity, but transient)
Building/Brick wall = 1 (retains and radiates heat in proximity)
Asphalt Driveway = 1 (retains and radiates heat in proximity)
Air conditioning unit = 3 (produces a lot of heat, proximity, may run regularly)
Bush = 1 (retains moisture, higher night temps)
Score=8
Don’t know if its a lights=0 station, but given the domestic issues, likely not, so maybe add a score of 1 for lights = X > 0
Oh wait, forgot one,
Fireplace behind brick = 0.5 insulated from wall, likely highly transient
Down amongst the bins? Perfect site for a BBQ. Surly it’s more likely its just being stored there.
Is there an edict going out that makes putting a GISS unit near an air conditioning unit mandatory?
I notice from Eli’s site that Roger Pielke’s station has some problems of its
own. People who live in glass houses, and all that.
Don’t forget the trees, which progressively reduce wind (mixing) and partially block IR from escaping.
I continue to wonder if things deteriorated when the US began its switch to the cylindrical sensors, which require a run of cable to an indoor display.
I think the situation is even worse than is shown because many of the country sites are surrounded by trees.
I surveyed the Brogdale Site at Faversham in England: Google Earth 51 17’49.20″N 0 52’40.30″E. This is the site that recorded the highest temperature ever recorded in England.
The site is in Orchard land crossed by windbreaks with a conifer windbreak on the prevailing wind side 13 metres from the Stevenson Cabinet. If you google in “Micro climate in Orchards” you will find a lot of research has been done to increase crop yields by raising the micro climate temperature because of less air movement, raising the temperature by “several degrees” both during the day and night. The night temperature is raised because the higher day temperatures warm the soil more, which then radiates upwards as the air temperature cools. Paradoxically it can also cause frost problems because the cold air near dawn gets trapped within the windbreak area being heavier air than that surrounding it, unless it can drain away down a slope.
And this remember is in countryside without buildings or roads nearby. This may cause a major warming effect over time as the windbreaks are planted and then grow taller, and have thicker foliage causing a gradual increase in temperature.
This would also possibly occur if the site is near woodland or any other changing landscape.
So bigcity, are you claiming that the problems found by Pielke don’t exist? Or are you just trying to change the subject?
#7
In California, they used to plant rows of Eucalyptus trees on the edges of orange groves. They also used smudge pots and later wind machines if it got really cold.
I am suggesting Roger might want to:
1) explain how his station got so crap.
2) explain why it really isn’t crap.
…which second point might lead to extending a principle of charity to the other stations pictured at surfacestations. For instance, most barbecues have wheels and though you might see a BBQ in one location this does not mean
that a BBQ took place at that location.
After all, two weeks ago I would wager that most of the readers of this site wouldn’t have known a
weather station if it had run up to them and bit them in the ass, screaming “I am a weather station!”
I still doubt whether they would constitute experts as to what constitutes a reasonably maintained one.
Just as I thought. You are merely trying to change the subject.
I also notice the typical liberal attitude towards anyone who isn’t smart enough to agree with them.
Re #10
It would appear that bigcitylib has turned into one of the people he despises the most, a (micro)climate warming denier.
Surely the objective of this exercise is to establish the facts (to the extent we can). I don’t see how slurring people contributes.
Yet another climate science innovation (end scarcasm).
Well forgive people for bringing that circumstantial evidence to attention. Maybe now it can be investigated how often BBQs took place at that location. Or you can just bury your head in the sand.
When displaying this aggression towards humans (which would be very interesting photo documentation if a site reader could come upon it!), might this activity generate some heat? Wouldn’t it be prudent to use domesticated weather stations to minimize this tendency to attack?
bigcitylib,
Do you actually do any due diligence?
Two minutes of searching finds that Pielke has indeed closely studied the Fort Collins site (which obviously is not “his”, despite your efforts to switch the discussion from the facts to a person):
Pielke Sr., R.A., T. Stohlgren, L. Schell, W. Parton, N. Doesken, K. Redmond, J. Moeny, T. McKee, and T.G.F. Kittel, 2002: Problems in evaluating regional and local trends in temperature: An example from eastern Colorado, USA. Int. J. Climatol., 22, 421-434.
Click to access R-234.pdf
There you’ll find the following:
“Fort Collins, a rapidly urbanizing site, warmed by 4.8 °C in minimum temperature from 1948 to 1998 (using original, unadjusted data), three times as great as the average slopes of the other sites.”
Re#15,
He was just swallowing the Gospel According to Eli.
The homeowner apparently had a better site…see these “before” pix:
And the plot…which has large breaks…maybe when the home changed hands or the homeowner was doing “improvements”?
The problem with these sites is that they are often volunteer. The old adage “you get what you pay for” certainly can apply to some, though many CWO’s take the job seriously.
ooops typo in url above here is pix1
The manager of the CRN website just sent me an email offering his assitance for all the CRN and HCN sites and surveys he had, including the ones that were rejected.
I’m working up a method with him to get them all imported so that we’ll truly have a one-stop shop for all weather station pix and surveys.
One the road today doing more site surveys. Will have pix posted tonight.
Where has the Pielke et al paper been submitted for publication? Is it the same/similar article to the one submitted to Bull. Am. Meterol. Soc. ? Seems to me that the comments from reviewers of either paper should provide some interesting reading!
Being located at about a meter and a half to two meters above the suface, how are these instruments measuring the temperature of the well-mixed atmosphere? Even the well-sited instruments will not be measuring such a quantity. How is the vertical location of the instruments related to the center of the mass/energy cell nearest the surface in GCM models/codes/calculations? I suspect that all the instruments are located within the region of the physical domain in which all the physical phenomena and processes are chracaterized by use of parameterizations. Additionally, I suspect the parameterizations cannot accurately reflect the complexity of the important physical phenomena and processes occurring near the surface-atmosphere interface. I’m certain that the parameterizations cannot reflect the micro-site effects that have already been discovered.
RE: #1 – The white plastic dome is either a Dogloo or a composter, most likely the latter. Minus 1 if the latter.
RE: #21 – Good point. The surface layer does not necessarily reflect the lower troposphere en masse.
#22,
Why a minus one if it’s a composter? Wouldn’t a composter be giving off heat?
Regardless, I don’t think it’s either of those. I think it’s just a white disk. It’s squeezed in between the AC unit and the wall. That area looks to me to be only a matter of inches deep.
Re #23, 24
My guess is it’s a table with folding legs, to be used when the barbie is fired up.
#19
Anthony:
That is very good news and a sign that there are plenty of open-minded scientists out there. Eli will be particularly pleased, not to mention Dano.
RE: On a second look I’d have to agree it’s not a dome and therefore N/A.
It’s interesting to note that the avg. annual mean temperature in Hopkinsville is actually trending downward How can this be happening with CO2 on the rise?
Page 10 has a reference to Table 1, but the pdf contains no tables. I also noticed that the bookmarks on the left do not work. Is that May 11, 2007 pdf file incomplete? — John M Reynolds
I am planned on visiting the CUYAMACA site in the next week or two. While trying to locate it, I noticed that the GPS coordinates place it next to the Cuyamaca Dam. In addition to the lakes area being very sensitive to precipitation and water demand, Lake Cuyamaca has a history of being completely drained for extended periods of time. In the hot dry San Diego summers, the lake is likely to have a significant cooling effect.
#30
Distance from the lake will be important as will water level. Very likely that lake level
will be recorded… I’ll check california records.
Thermal imagining would be interesting (but expensive) or even a simple $50 IR Thermometer might be interesting to try out at these sites. I ‘ve never used one but if they work well they could help identify heat sources.
additional comment to #32
Arial thermal imaging of these sites would prove interesting too, but I don’t know of publicly available photos.
Arial thermal imaging and remote sensing is being used more and more. My brother uses it irrigation and for crop insurance. I used to know a guy that does thermal imaging photograph, they take arial thermal pictures of the sugar beet piles and by the temperature they know what part of the pile to process into sugar and when. (Sorry, if that’s getting off topic.)
People interested in visiting sites should consult the Station Locator here:
http://www.ncdc.noaa.gov/oa/climate/stationlocator.html
I selected Marysville CA
http://www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwDI~StnSrch~StnID~10100105
Asked for Station History
http://www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwDI~StnSrch~StnID~10100105#HISTORY
Asked for More MMS
http://mi3.ncdc.noaa.gov/mi3qry/identityGrid.cfm?fid=2982&PleaseWait=OK
This returned 5-digit lat and longs. I’ll check with Anthony as to how these compare to actuals. If these are accurate, then this might work for other sites. Nicholas could probably figure how to scrape the information off the website for USHCN sites, if there is no existing collation.
Latitude: 39.14583 (39°08’44.988″N)
Longitude: -121.58528 (121°35’07.008″W)
32.98972 -116.58722 are shown as coordinates for Cuyamaca in the above process as compared to the ones that you used: 32° 59′ 24.00″, -116° 35′ 24.00″ . They are even closer to the dam than the ones you used.
#31,
Distance to the lake will vary depending on lake level. Especially if the slope of the ground is shallow.
Additionally, when the lake is low, it will heat up faster.
Being close to a lake is bad. Being close to one that varies in level dramatically is worse.
#28. The downward slope is seen because it hasn’t been adjusted yet. Or a faulty adjustment was made. However, the most likely cause is the warm decades from 1910 to 1950. Those measurements are too high; that science is settled.
It may also be an example of the rare ‘Perceived Cooling’, aka PC, which is not actually cooling but has all the attributes of cooling.
Amateurs might guess that were probably changes in trees, irrigation, lawn watering, and lake surfaces in the area over a century.
#36,
Since the site is located next to the dam (the low point in the lake), distance to the lake is nearly a constant.
Re:#38
As the lake level varies, certainly the vertical distance between the station and lake will vary, perhaps by more than 100 feet (depending on dam height above the bed). Also, if the station is not closest to the deepest part of the bed, the horizontal distance to the water will also vary with lake level, as noted above.
#35,
If that is a more accurate coordinate then it’s in a rather interesting location. It doesn’t look accessable to the public by either foot or boat but I should be able to get close enough to identify and photo it.
Hybrid map
Ed
#34 The accuracy seems to depend on the site. Some have higher resolution GPS parameters than others. I’ve used google earth to scope out a number of sites I will try to visit this weekend to get an idea of what I am dealing with in terms of location. Based on the coordinates I entered I would guess a couple are close to dead on and the other two are probably off a couple hundred yards (one was in a reservoir).
R. Pielke Sr. is a past chairman of the Assoc. of State Meteorologists. I asked him about a yera ago if he could have each state meteorologist identify at least 2 sites that had not suffered significant land use change in the last 50 odd years and then use these sites to check the average temperature chanege. He replied that he didn’y have time. I suggested a graduate student assignment, but got no reply. Pity. Murray
Since I learned about the method to get the more-exact coordinates, I’m 1 for 2. Fort Stockton, Texas was dead-on accurate (and I logged it, and the site looks pretty good); Boerne, TX was not. (and I ultimately gave up trying to find the sensor – neither city hall nor the fire department were of much help.)
#32. If you get a IR sensor I suggest you calibrate. My $100 version was about 10 deg F off.
On another note I could not find records of lake levels ( daily, monthly etc like most California
Resevoirs in Norcal have ) But I did find anecdotal histories that indicated the lake has been
emptied periodically throughout history. http://www.lakecuyamaca.org/history.htm
have a look. I’ll check some more
stuff. I think the lake is very shallow, like 10 ft.
Oh, they got an award for 50 years of service from NOAA for their climate station
http://www.hwd.com/news/noaa.htm
A couple of interesting excerpts from the article:
and
I bet a virtual dollar/euro that the MMTS’ need for a cable to connect the sensor to the display/recorder played a role in some of the relocations. Closer to the house = shorter and easier cable runs.
The University of Illinois offers some quantification of microsite changes, including an urbanization effect of about 0.8C.
#44, The 50 years of service award link was quite helpful since it identifies the Helix Water district as the owner of the sensor. I’ll know whose “posted” signs to look for :>).
RE: #46 – I’ll bet that it’s out on the end of the dock / structure off the dam, which looks like a height gage for the water. They probably stuck it up on top. If so, indeed an interesting location especially late in the day when the dam starts to heavily radiate all the stored heat from having the sun beat down on it all day.
#34
I would add another useful mapping site (http://mapper.acme.com/) as a tool for location of the stations. It draws from Google maps but also includes high-res USGS topo maps. I found the topos useful for spot-checking the ground surface elevation for correlation to Anthony’s master station list.
This helped me on locating sites within cities/town with multiple listings in the NCDC list that Steve linked to. No use wasting time and gas $$ locating the non-HCN sites.
Since it’s becoming a theme at some of the stations, I’ll plan on packing the Charcoal grill for a hot lunch during my visit…….
Thanks for a great site, Steve…definitely eye-opening.
re #47.
I don’t know how many folks appreciate how hot damns can get. I would suspect
that the station would be loacted close to some structure
It appears that the move to MMTS has drawn the stations close to structures, for
cabling requirements. ( open question of course )
It might be instructive to have a look at the cable they use and the transmission
protocal. I would expect that the automated stations are connected via a standard
instrument low bandwith interface ( given the age I’d guess RS232 or something like that)
I’ll go see if I can ferret out naything in that regard.
If we knew the interface we could almost dictate how close to a structure ( computer)
it would have to be.
This is an example of the address format for using lon./lat. with maps.live.com. Birds eyeview here shows dam and surrounds. Not a real close up. In eastern MA you can see trampolines in people’s yards…just in case you told your insurance co. you didn’t have one.
http://maps.live.com/default.aspx?cp=32.98972~-116.58722&lvl=14&style=r&v=2
maps.google.com address format for lat./lon.
http://maps.google.com/maps?q=37.771008,+-122.41175&iwloc=A&hl=en
Hmmm…I am wondering how I tell which station is which. The Poughkeepsie, NY USHCN station (ID 306820) seems to be represented by two stations, one here
http://www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwDI~StnSrch~StnID~10100288
and the other here:
http://www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwDI~StnSrch~StnID~10008479
For the second link, scroll to the station history to see the station ID.
Both stations are operating today and are some distance apart. Interestingly, NCDC posted pictures of the site in link 2, which is located at an airport. A total of 209 NCDC airport and heliport sites have photos posted.
Any thoughts on how to tell which 306820 is which?
Note that you can just type `latitude, longitude’ coordinates into the “Search the map” box on maps.google.com (e.g., `51.47, 0.0′ for Greenwich).
Caveat Beware that google maps has an error of several hundred meters: try typing the exact coordinate of the greenwich observatory: 51° 28′ 38″ N, 0° 00′ 00″ W in
the http://maps.google.com/ search box.
The green arrow won’t point to the observatory…
Has to do with geodetic datum shifts.
#54
Possibly Google maps use WGS84 datum
http://www.nmm.ac.uk/server/show/conWebDoc.416
#49 The limit for RS232 is around 25′. At that time we were using 20ma current loop (as used in the old teletype terminals) for similar applications. 20ma current loop has a limit of around 1000′.
#38,
Unless the sensor is located at the middle of the dam, and unless the dam is miles long, then lake height will vary how much water is near the sensor.
In this case, distance to the water will still vary as the vertical height of the water changes.
If the sensor is located on dry ground, near the end of the dam, then the average distance to the water will still vary depending on lake height.
also see discussion at Pielke Sr blog here )
re #58.
Thanks SteveM, I found the peterson aricle last night and posted a link to it over on another thread
( surfacestation). Also found an interesting bit on the impact snow cover has on the MMTS.
posted in surafcestation comment 104
RE 56, I have the manual for MMTS back at the office and it has the cable info, will post it when I get back from site surveys.
http://www.campbellsci.com/air-temperature
You can get info on temperature sensors here.
I’m still wondering how to discern which of two sites is used the USHCN (see #52). After all, I don’t want to post and later be responsible for the entire site being discredited because I went to the wrong site.
GISS puts the station at 41.6N, 73.9W with operating dates of 1/1890-5/1900 and 4/1928-present with intermittent outages as late as 2/2004.
The USHCN refines the location at 41.63N, 73.92W.
The NCDC MMS pinpoints one of the two stations in question at 41.63333N, 73.91667W. The second station is located at 41.62667N, 73.88417W.
The two sites in question share the same coop ID and are operating today. The first station is located in the vicinity of a mall and has records going back to 1993. The second station is at a municipal airport and has records going back to late 1932. Pictures of the second site are posted on the NCDC website.
I am guessing that the station referred to by GISS and USHCN is in fact the second site, but want to be sure given the operating dates do not match up and the close proximity of a second, operational site with the same coop ID.
Any suggestions?
Anthony,
I registered for the site. Great work!
It might be interesting on the front page to show
1. Total sites in each network
2. # of sites visited.
3. # of sites with “issues”
Ideally, you could have a map where folks could click on their state and get a list of unvisited sites
Have fun
#62. ask the National Weather Service at NOAA for the B-44 forms. That’s what Karl et al say to do – see my recent post.
Try talking to the station operator which is listed as EDWARD H SOMMERFIELD. I’m sure he could shed some light on your dilemma.
A complete list of stations and who is operating them can be found here:
http://cdiac.ornl.gov/ftp/ndp070/history.txt
#52. The first one started in 1993; the second one in 1932. So the second one is the historical site.
Sorry, the last post was in response to #62
Re: #63
An addition to steven’s list:
It would be nice to have a quick link to a list of ‘recently added sites’ on Anthony’s surfacestation.org site.
Thanks Steve (#64 and &66) and Barry B. (#65), I will contact them to clarify. Further reading of information on both sites found on MMS seems to clarify things but I need to verify.
Here is what I pieced together, and it represents the problems I have always had with the historical data.
The airport location held the coop identity until February 1993. The location with topographic details: TOPO-STATION N WIDE FLAT VALLEY SURROUNDED BY HILLS ALL DIRECTIONS. Pictures on the NCDC website show the current station is unobstructed on all sides, as one would expect with a station in the middle of a municipal airport. The observation time is given as 2400.
For the newer (residential) site, the B44 says it was established in November 1993: STATION REACTIVATION, RELOCATION (NO LONGER AT AIRPORT), NEW OBSERVATION TIME. Relocation was 1.5 miles W to location with topographic details: HEAVILY TREED RESIDENTIAL AREA, MAINLY FLAT. The new observation time is 0700.
This relocation corresponds to the absence of monthly GISTEMP data from 1/1993 through 12/1993.
So it appears the station underwent a significant change in 1993, but as I said, I will verify directly with NOAA. If true, the nice thing is we have pictures representative of the original site (it not the original equipment) already on the NCDC website, and I can try to get pictures of the new site.
45:
Whoa, here. If I understand this correctly, there was a dramatic shift in the way temperatures were recorded in the mid-late 80s. Prior to then, someone simply read the thermometer twice or several times per day, obtaining instantaneous temperature measurements, which were not necessarily the highs or lows for the day. Now, they are recording the highs and lows? If this is true, it amounts to a huge inhomogenity. Am I missing something?
jae,
Another poster has written in that the older style sensors had a mechanical mechanism for recording the high and low. For the high, it was a plug that was pushed up by the mercury, I don’t know what the means was for the low.
Seems to me that the plug marking the high could settle if there was enough vibration, or even if a temperature spike caused the glass to expand enough. Or if years of slidding up and down the glass tube had caused enough wear that the plug no longer fit snuggly inside the tube.
70. Oh. Thanks.
69#, The thermometers used mechanical means to capture the min and max temperatures. Try goggling “min max mercury thermometer”.
If anyone is interested, re comment 29 by me, I sent Dr Pielke an email. The reply was that a new version will be uploaded in the next couple of days. — John M Reynolds
I’ve been spurred to seek out some Canadian sites in the Southern and Central Alberta vicinity. Can anyone tell me why most Canadian sites stopped recording around 1989-1990 and the ones that do have data show a strong discontinuity (i.e. jump) in the mid-late 90’s? The GISS record in central Alberta has only one post-1990, Edmonton Muni Airport to 2001, in the heart of the UHI. Neither Edmonton Int’l nor Calgary Int’l Airports are current. Oddly, there are many more current sites in the Arctic.
Baker Lake is one such site.
i wonder if ed made his cuyamaca trip and if someone has done the survey there iaw surfacestations criteria?
i found this image that is arguably a photo of the station.
i found its link on this nws page
you can see a paved surface at the bottom of the pic. in the near background you can see where there’s a road with power poles along it. and the station is within a few yards of the road. is anyone planning to go there? i plan to.
As I have seen several comments about the portability of grills, I would like to use the Hopkinsville, Pielke,to illustrate the value of looking at mico-site information versus making assumptions. I grill every Friday night or better for the past 21 years. I have always periodically grilled since I was 8. Although, grills often have wheels they are cumberson to move. It is easy to roll them on a cement surface especially if you limit the movement to only a step or two. After that, balance, integrity of the wheels, and other factors determine how easy it is to move any particular grill. Expierenced grillers store the grill near their normal cooking spot. You can make a large mess or damage your grill moving it around too much or too far. The problem with anyone assuming that someone would move a grill far from its storage spot fails to understand grilling as shown here. It is apparently a charcoal grill. In fact it appears to be the same model I use. The problem assuming a move is that you would have to move your grill, your charcoal, your utensils, and yes the big white round object appears to be a fold-up table. I have used several in my lifetime. So your griller will now have to move all this, plus what he is cooking, and do not forget the BBQ suace for the meat or the alcohol for the griller, chairs and coolers. Nor is this the only concern. The close proximity of the unit to the shown wall means that any grilling nearby will mean that the building’s impact on heat and wind will draw the hot air towards the unit, whether it is from the cement or from the grill. Finally as a humurous note, a commenter in satirizing the grill comments talked about grilling in a huricane. Been there, done that. It was Hugo that thankfully missed us. I grilled that night to make sure that if we lost electricity we would have something cooked we could eat without heating. I also got an extra bag of charcoal so if we lost electricity, we could still have a cooked meal each day.
One Trackback
[…] ⟨ttp://www.climateaudit.org/?p=1657 […]