Weathering and Thermometer Shelters

Former Virginia State Climatologist Patrick J. Michaels wrote an op-ed about his paper with Ross McKitrick from Canada’s University of Guelph in an American Spectator column today about the surface temperature record. This paragraph really caught my eye: “Weather equipment is very high-maintenance. The standard temperature shelter is painted white. If the paint wears or discolors, the shelter absorbs more of the sun’s heat and the thermometer inside will read artificially high. But keeping temperature stations well painted probably isn’t the highest priority in a poor country.”

The Stevenson Screen experiment that I had setup this summer is living proof of this.

Compare the photo of the whitewash paint screen on 7/13/07 when it was new with one taken today on 12/27/07. No wonder the NWS dumped whitewash as the spec in the 70’s in favor of latex paint. Notice that the Latex painted shelter still looks good today while the Whitewashed shelter is already deteriorating.

Click for full sized image

Click image for larger view

Click for full sized image

Click image for larger view


Whitewashed Screen on 7/13/07


Whitewashed Screen on 12/27/07

The whitewash coating I used was from a formula and method provided to me by a chemist at the US Lime Corporation, who is an expert on whitewash. He said the formula was true to historical records of the time when whitewash was used on the shelters. I was amazed to find that after just a few short months, my whitewash coating had lost about 40-50% of it’s surface area. Perhaps there was a mistake in the formula, or perhaps whitewash really is this bad at withstanding weathering.

In any event the statement of Patrick Michaels “Weather equipment is very high-maintenance. The standard temperature shelter is painted white. If the paint wears or discolors, the shelter absorbs more of the sun’s heat and the thermometer inside will read artificially high.” seems like a realistic statement in light of the photos above. The magnitude of the effect in the surface temperature record has yet to be determined, but it seems clear that shelter maintenance, or lack thereof, is a significant micro-site bias factor that has not been adequately investigated nor accounted for in the historical temperature record.

I’ll have more on this experiment soon including temperature time series graphs showing the difference between bare wood, latex painted, and whitewashed shelters.

0 0 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
George M
December 27, 2007 5:16 pm

You may well have waited too long to find any accurate information on whitewashing of the historical past. For one thing, Tom Sawyer is dead. But, given the government’s propensity for writing specifications on all things useless, is there perhaps a MIL-SPEC on whitewash preparation buried in the vast DC bureaucracy somewhere?
Hope you had a good Christmas.

December 27, 2007 8:09 pm

Are you going to publish this (negative) trial for bias in the peer-reviewed literature?
Reply: I haven’t decided, as I have not begun to analyse the data from the NIST datalogger yet, that’s coming up.
I’ll probably strip the shelter of whitewash, rewash it, repaint the latex shelter, and run it all again this spring/summer so that I have repeatability. One run of time/temps isn’t enough I don’t think.

December 27, 2007 9:25 pm

You should set up falsifiable hypotheses, test them, and publish the results whether they support your suspicions of bias or not.
Please provide the temp records for this run. (you seem willing to share in process and early information when it goes one way, but not the other…seem a lot like Steve McIntyre’s image of a selective report of drill results, mining promoter. Actually he seems a lot like a mining promoter also, especially given his work with penny stock Canadian companies, that are practically “shells”.
REPLY: Thanks for the suggestion, I will but I’d point out that you seem to be confusing “waiting for data runs and data collation to be completed” with “TCO wants it now”. I’ll provide the data when it’s ready, right now I’m collating four months of data logging into comma delimited files so that I can import it into Excel. The datalogger doesn’t write in an easy to import format.
I’ll have a complete history of the experiment written up as well, today’s post was just a quick one due to aritcle in American Spectator being relevant.
As I said in the other post. I’m doing this on my own time on my own money. So I have to balance this project with the project, my blog, my daily radio program, my business, and my home life. I ended the data run I started on July 20th in mid December, so far until today, I haven’t had any time for data analysis.
Of course you are welcome to replicate the experiment yourself if you are unhappy with the schedule so far.
BTW, your website, says nothing about you. If you want trust and respect and for people to follow suggestions you make, my suggestion is that putting out information is a two way street. For now, you are just another anonymous net phantom called “TCO”.

December 27, 2007 9:37 pm

I know that it’s easy to criticize. And I give you credit for doing something…rather than being so overwhelmed by experiment design that you do nothing.
That said, I think there is a fundamental flaw in your experiment design in that you only have one screen/thermometer of each type. You need to have at LEAST 2 (preferabbly more) of the latex and white wash machines…so that you can resolve inherent “box to box” or thermometer to thermometer variance from latex to white wash. It’s actually not JUST that this is a flaw on this experiment, but that I think you don’t really think about experiment design and statistics well. Saw this same problem with your (much too much touted) data curve for the painted wood samples initially. You seem to confound multiple data from following a time series versus from having multiple sample points.
REPLY: Well its a good idea to have 2 more setups to weed out such issues, but the big problem here is that I’m doing this entirely out of my own pocket, and so far I have about $5000 invested in the new screens, shipping, datalogger, NIST calibration of the datalogger, probes, and NIST calibration of the probes.
So if you or others wish to help fund duplicate setups, I’ll gladly do it. For now, this is the best I can do as an individual funding it myself. Perhaps the real value of this work will be to find enough things of interest to have somebody write up a grant proposal to do a more elaborate test.

Evan Jones
December 27, 2007 10:34 pm

TCO, The Rev seems to be very straight-up about his experiments.
I’m sure he’ll report any and all biases he finds, in whatever direction they may be. I’m entirely confident he will provide his methodology and data. If he had used andy code, he’d gladly provide that AND full procedures and operating manuals.
Because THAT’S Science. As opposed to the alchemy-and-witchcraft-meets-the-Freemasons For Worthy Eyes Only approach to routine falsifiability we get from SOME people! A wall full of PhDs and a cocktail party full of tenured sinecures can’t make that sort of hocus-pocus, hierophant pum-pum a real scientiist–they just play scientist on TV. If you require a Freedom of Information Act lawsuit to make you cough up data and methods, whatever you have bubbling in your cauldron ain’t science!
The Rev is a true scientist who has always operated in the open, online, with experiments discussed before he did them and verifiable afterwards. As you, yourself, have observed, he has posted his results, with photographs, for all to see. He does not “mine”, he does not “pick”. And those who so presume merely reveal themselves.
I look forward to those comparative temp series. #B^1

December 28, 2007 6:03 am

To make shure that any bias is in fact not inherent in the instrumentation, just switch the instruments inside the screens every month or so. If a bias is due to the instruments it will show up in the time series as overlayed step functions. You might want to run a full year first and thereafter start to rotate the instruments.
Thanks for your efforts!

Stan Needham
December 28, 2007 6:59 am

Anthony, OT, but have you been contacted by a Boy Scout troop in the Lawrence, KS area WRT doing surveys in Kansas? My daughter decided that surface station surveys were a little beyond the capabilities (or attention span) of 6-8 year-old Cub Scouts, and she passed the information on to the scoutmaster of a troop that their Cub Pack feeds into.

Larry Sheldon
December 28, 2007 9:15 am

I looked (briefly) at and could not find anything but severe eyestrain.
My suggestion in the matter is to expend no more scarce energy on “TCO” beyong banning her if she becomes enough of a nuisance to justify the effort.

December 28, 2007 9:58 am

AW: That’s fine, but realize that your results and any conclusions that you draw from them will be limited by not addressing this issue. You might want to read the Feynman essay (it’s in his second essay collection, get it via ILL at your public library) where he talks about the man who wrote the best “rat running through maze” paper ever. The point of that paper was not anything to do with intelligence or learning, but to do with all the confounding variables which can sneak into an experimental design. I also recommend the 1950s book by Wilson (Dover paperback).
Evan: I give AW credit for showing the pictures and the problems with the experiment. Would have been more impressed if I did not have to call him on this (to share it). More impressed if the problem pictures were revealed with the same speed that the initial temp graphs for painted wood were revealed.
AW: Please respect my privacy.
REPLY: TCO, thank you for the book suggestions. I agree that lots of entropy can creep into any experiment. Let me try to explain this again, because I think you’ve missed the time-line, and are making assumptions without that knowledge.
I started with the paint experiment in May. I published my simple wood slats experiment right away to determine if a large scale experiment was even worthwhile as there was no point in my spending my own money unless I could observe differences in temp versus coatings on a small scale.
I surveyed nearby stations to determine in the field paint characteristics, and found siting issues.
I looked for a database of screen placement/siting issues, found none, started corresponding with Pielke.
The idea gelled for a national survey, I put the paint experiment on hold while I launched in June.
I revisited the paint experiment in late June/July, had equipment issues and failures which had to be solved. Datalogging commenced on July 20th. I had a communications link failure, which was fixed a week later, on the 27th, and my first full day of data is from July 28th.
My original intent was to show the paint results at Pielke’s conference in August, I had actually submitted two abstracts, one on station surveys, the other on screens/paint. You can confirm this with Pielke.
The paint experiment delays made it difficult to collect data and write a paper in the short time I had left. I had originally spec’d a month as the time for a data run, but I would have only two weeks worth of data at Pielke’s conference (needed two weeks of time to analyse and write up a presentation also) and I decided that presenting such a short data series would draw a lot of criticism, so opted to not present at the conference. In the meantime, the experiment continued to log data.
Since the opportunity was gone to present at Pielke’s conference, I opted for a long data run of four months so that I would have seasonal variation in the dataset.
I just took the datalogger offline a couple of weeks ago. I’m now in the data collating phase and soon to be in analysis phase.
I plan on setting up the datalogger again in the spring and plan to let in run though summer.
TCO you did me a favor when you started making claims that I was holding back data because it reminded me that there were many people that still wanted an update, which I’m doing on the blog. But, the intent to release results has always been there. As the operator and funder of the experiment I have the ability to choose my timeline, you do not. Sure I’ll get all sorts of flack for perceived slights or irrational judgements, but the difference remains that instead of just carping about things on blogs, as the majority of Internet phantoms such as yourself do, I’m actually doing the work. I don’t appreciate the judgemental statement you are making that “I have to call him on this (to share it).” which is in error. I find it comical though that an anonymous person sits in judgment of another but contributes nothing beyond criticisms. Such is the Internet today though, and many practice anonymous slings and arrows with no fear.

December 28, 2007 12:22 pm

I would point out the “respect my privacy” has long since been stripped from the public conversation when AGW proponents decided that “in the pay of special interests” was a legitimate way to dismiss objections.
AW, there might be a way to parlay you considerable effort and expense (thank you) into a profit and great benefit to the field. I’m thinking we can design a Stevenson Screen Mark II from modern materials with stable characteristics and otherwise similar performance. Market them for a few hundred bucks and within a few years the Watts Screen could become the reliable standard.

Bill in Vigo
December 28, 2007 12:34 pm

I find the accusations comical. It is obvious that some do not actually read or study the material released from this blog. Nor do they appreciate the amount of work and expense required to complete the field work. Always for me the field work was fun, the colation and analysis was work. Perhaps TCO is well funded by outside sorces and dosen’t consider the financial responsibility side of the experiment.
By the way he is not alone many only look at the pictures and criticize. ( they remind me of my grandchildren) not realizing that the full story is in the text.
My statement is to read (THE REST OF THE STORY).
Thank you for work well done.

Evan Jones
December 28, 2007 8:36 pm

“I give AW credit for showing the pictures and the problems with the experiment. Would have been more impressed if I did not have to call him on this (to share it). More impressed if the problem pictures were revealed with the same speed that the initial temp graphs for painted wood were revealed.”
So you did, which I acknowledge.
But he’s very open about his data (And I admit I get defensive about him). Consider that both NASA and IPCC and other official agancies which should know better have outright refused to share data or only reluctantly gone halfgway under legal threat. Which is not a very scientific attitude, to put it much more mildly than I did earlier.
The initial quick-tests were made just preliminary. He then engaged in a longer, more comprehenisive, long term test, and did not release it until it was done. Remember, the Rev resisted many eager appeals to release the surface station data until at least a third of it was completed.

December 29, 2007 1:21 am

Another quick question, re-post in appropriate thread:
Of the 460 or so surveyed stations, how many are “rural”? According to Hansen’s papers, we only needed about 250 rural stations to effectively cover the US.
I know, I’m asking for another map, showing quality of the “rural” stations.
If you (or your mapmaker ) has the time…

Erichsen Kollmar
December 29, 2007 9:39 am

Having used whitewash for the interior walls of my house (they’re poured concrete) This is the formula I was told to use by a guy who was an oldtimer, Basically one part lime, one part white cement, and enough water to make it workable with a heavy bristle brush. This mix is more durable, but I can’t say that it was the mix originally used on the stephenson screens. It may also have different thermal characteristics as compared to the straight lime wash that you have been using in your experiments. I posted this recipe a while back, but I recognize that this not an official recipe.

Rich Godbey
January 2, 2008 1:36 pm

Anthony approached me last year regarding a formula for mixing limewash (whitewash). I sent him a copy of the paper that I co-authored with Mr. Peter Mold. This paper was written for the 2005 International Building Lime Symposium held in Orlando, Florida. If interested, please take a look here to read the paper:
Spring wood and winter wood and species differentiation all have an effect on the mechanical keying necessary for successful whitewashing on wood. Whitewash must be put on very thinly and built up over the course of several applications to a thickness that provides opacity. Over time, whitewash will carbonate and good practitioners will encourage this reaction through a series of wetting and drying cycles. Carbonated whitewash is not only durable, it is stable.
I want to thank Anthony for exploring the issue of whitewashing Stevenson Screens. I think he is on to something and I look forward to the results of his continued work…

John G. Miles
January 2, 2008 2:54 pm

From TCO: “Are you going to publish this (negative) trial for bias in the peer-reviewed literature?”
Heaven forbid that the obvious “negative” bias that results from darker materials absorbing more sunlight be confirmed. Otherwise we’d have to ignore all the alarmists claims about spiraling-out-of-control positive feedback for AGW with respect to snow and ice albedo. The only real question is “how large is the bias?” or can ventilation compensate for most of the to-be-expected warming.

May 12, 2008 3:15 am

Is it possible to look at the graph of the test results somewhere? Very much would interest the final results!
REPLY: Just do a search on “stevenson screens” and you’ll find it

%d bloggers like this: