Bad Paint Job = Rising Surface Temperatures Part2

A test of wood surface temperatures related to paint

NOTE: This is not research quality experiment, but a simple test to point me towards others tests and experiments. IR thermometers have limited resolution and are subject to calibration errors. The real experiment is being conducted on 3 identical Stevenson Screens with an NIST calibrated data logger and the results will be published in a couple of months.

You may recall a blog entry where I talked about

Bad Paint Job = Rising Surface Temperatures? The premise is that the early

weather station temperature shelters called Stevenson Screens originally have

been specified to be painted with whitewash when the were designed and

comissioned in the 1890’s

cotton region shelter aka Stevenson Screen

Stevenson Screen at the NWS office in Monterey – good

paint but right next to asphalt parking lot and concrete walk – a

definite no-no!

Lately they have been repainted with lead, oil, and

latex based paints which have significantly different infra-red properties

(Pigment: Titanium Dioxide) than the Calcium Carbonbate based whitewash.

Curious to follow where the evidence leads, I decided to do a test today.

Preliminary spot test of the temperatures of bare wood, latex paint, and

lime based whitewash (CACO3)

5/17/06 by Anthony Watts, Chico, CA

Click thumbnail pictures for larger ones

Test setup in full sun. Three slats of 1×4 pine, 2 feet long each, bare wood

in the center as a control, latex on the left, whitewash on the right.

Methodology: Whitewash was mixed after conferring with Chemist Richard

Godbey of the Chemical Lime Company in Henderson NV and after reading a paper he

authored on the history and home creation of whitewash which you can read here


Pictures: Materials, Mixing Whitewash, applying whitewash coat

1, coat 4 after 24 hour curing

Slats were all cut from same plank, chosen to be as knot free as possible,

cut to exact same size, 2 ‘ long, and spaced equidistant on the frame.

Measurements were taken 3 times with IR thermometer shown below, at1:50 PM,

2:10PM, and 2:20 PM PST to be sure results were repeatable. What you see below

is the 2:20PM spot test.


Device used to measure surface Temperature, A Fluke model 561 HVAC Pro IR

thermometer, handheld with laser dot sighting of measurement target

Temperature of bare wood, this represented average of several scans. The

temperature was 102.6 degrees F

Temperature of Latex Painted wood, the leftmost slat temperature 88.2F

Temperature of Whitewashed wood, the rightmost slat temperature 82.5F


Temperature difference between bare wood and Latex painted wood was about 14

degrees F

Temperature difference between bare wood and whitewash painted wood was 20

degrees F

Temperature difference between latex painted wood and whitewash painted wood

was 5 degrees F


Next Test: time series temperature over several days measured by NIST

calibrated data logging thermometers compared to ambient aspirated air



newest oldest most voted
Notify of

Would it be difficult for you to just go ahead and get a couple stevenson screens? That seems to me like it would be the most accurate way.

David Walton

Outstanding Mr. Watts. A simple and elegant experiment. I will be watching to see how the data data develops. Maybe you can attract some grant money and expand to multiple full stations and a variety of coating materials. Be sure to mention that you wish to CONFIRM the validity of anthropogenic gw temperature data, the bucks will flow.


I have 3 on order now. 2 bare wood and one pre-painted with latex.
But, I had to do this test first to be sure that the differences I suspected existed in paint IR absorption/emissivity. Otherwise I’d be tossing a couple thousand down the drain on SS’s on a whim.
Now I have a reason. In the meantime, I have a couple weeks before they arrive, so I’ll do other testing.


Thanks for the suggestion, but I have no interest in pursuing grant money, it just gets in the way because you have to account for it’s expenditure.
Besides, I should be able to complete what I need soon, and I have funded scientists watching who can’t move as fast as I can ready to pick up the work.
My intent is to open a door, let them fill the room with paper.


What an interesting experiment. It seems like beyond the paint the locale would also have significant impact on temperatures. I know you’ve mentioned that before, have you had much opportunity to visit different sites?
Also, have you received any information about changes in temperature measurement techniques/sensors over the years?
I ran across a couple of university sites that described in detail the sensors, and they do appear to be highly accurate and make use of good averaging methods as well as accounting for electronic bias of the circuits? However, I’m under the impression that those methods are probably only recent introductions, with at best 10-20 years in the field. I would doubt that those components were changed out with older thermometers en masse.
I also sent you the “brief” of a white paper that set the temperature errors at close to +/- 0.5F for the measurement devices. Have you seen any more detailed information on that aspect of the measurements?


Anthony- Good to see you starting on the experiments. A word of caution on using the IR thermometer. It relies on temperature measurement by detecting the infrared emission from the surface being measured. If there are differences in the IR emissivity of the surfaces (which is directly related to the effect you are trying to determine in your experiment), then the temperature measurements may be quite different. I say may, because various IR thermometers use different techniques to correct for differences in emissivity. You can check this by placing the samples in a controlled temperature environment, like an air-conditioned room, so that the samples are at the same temperature, and then measuring the temperatures of the samples to see if there is a difference.


Yes, using IR thermometer, emissivity difference (due to material difference) may lead to different readings for a same temperature.

Bob Meyer

Great experiment. The Fluke that you are using can be adjusted for emissivity so you will probably want to make the measurements that paminator and Demesure described and do it in a dark or dimly lit room. You can then find out what emissivity adjustments result in the same temperature for all three cases.
I’m not familiar with that particular instrument but the manual probably describes this process well.


but what portion of the radiant temperature is due to the emissivity of each surface and what portion is due to the temperature?