Does a home-made radiation shield work?
A thermometer in direct sun measures the temperature of the thermometer. The usual fix is a stack of white plates that blocks sunlight and lets air through. Bought shields cost more than my sensor, so for two years I used eight white plant saucers on threaded rod.
The comparison
This spring I borrowed a commercial shield and ran both for six weeks on the same pole, half a metre apart, with sensors from the same batch. Before the test the two sensors sat together in a closed box overnight and agreed within 0.1 °C.
| Conditions | Saucers minus commercial |
|---|---|
| Night, any wind | 0.0 °C |
| Sunny, wind above 2 m/s | +0.2 °C |
| Sunny, wind below 1 m/s | +0.9 °C |
| Worst single reading | +1.6 °C |
What I take from it
With any breeze the saucers are fine. On still, bright afternoons they read almost a degree high, and those are exactly the afternoons that set the daily maximum. So my summer maxima for the last two years are probably a little too warm.
The saucers have small gaps and a flat underside, so warm air collects under each plate when nothing pushes it out. I widened the gaps from 10 mm to 18 mm and painted the undersides black, as the commercial one has. A second short test put the calm-weather difference at about +0.4 °C. Better, not solved. A small fan would solve it and add something else to break.