Aspirated DIY Stevenson's screen - inside colour

[h2005__uk__]

Hi

What inside colour would people recommend for a small, wooden, double-louvred aspirated Stevenson's screen? I've read black is best as it'd prevent any light being reflected onto the sensors, but I'm worried the black surface would store heat.

Also, should the fan (for aspiration) blow air onto the sensor or away from it?

Thanks

[Terminal Moraine]

White was always the traditional colour for the inside ( and outside ) of a Stevenson's screen but the Met Office conducted a series of tests and found that black was better on the inside of the screen; certainly the plastic screens now used by the Met Office all have black interiors.

There should be no problem with the black surface absorbing heat provided it isn't subject to a direct heat source. Any screen, be it black or white on the inside, would give erroneous readings if it were too near a secondary heat source such as a wall exposed to the sun, or an area of concrete or tarmac but away from such sources the internal colour shouldn't make any difference.

With regard to the fan I'd suggest it should blow onto the sensor as the idea is to replicate a screen which is well ventilated due to a constant breeze. The largest errors regarding Stevenson's screens occur in warm, sunny weather with very little wind which allows secondary radiation from the screen itself to warm the air inside it to a level higher than that of the ambient air in the shade outside.

I've been using a wooden Stevenson's screen painted gloss white inside and out, and also a Met Office plastic screen with a black interior simultaneously for the last 8 years and have found no discernible difference between the two sets of readings other than minor differences due to the differing positions of the screens.

 

[h2005__uk__]

Thanks for reply.

I've attached pics of my screen - it's about 25x25x25cm in size. I've not yet painted the inside, but I have fitted a fan at the (north-facing) back, which blows air from outside to the inside. I think it's helped and the screen seems to perform better than my old screen, but the readings are still a bit 'spiky' when the sun comes out.

Anyone got any further advice?

Thanks!

[Terminal Moraine]

I think your temperature spikes could possibly be do to the relatively small size of your screen and the relatively large volume of wood it's made of, or your temperature sensor, or a combination of the two. Looking at your sensor in the photo' it appears to have an inbuilt thermistor ( rather than a thermistor on the end of a wire). Generally speaking you have to pay quite a lot of money ( £150- £200 ish) for a really accurate temperature sensor with an inbuilt sensor as the models at the budget end of the market tend to be accurate only to 0.5 to 1.0c and also tend to drift with time.

Also, you have your sensor leaning against the woodwork of the screen which would make it more liable to be recording the effects of  secondary radiation  from the screen itself. Small screens containing a large volume of wood are susceptible to greater errors than larger screens with a relatively small volume of wood due to the screen absorbing radiation and re- radiating it into the air of the screen. The same also applies to the mini screens made of plastic although the 'pile of plates' type construction was found to be better than the mini Stevenson screen type in tests conducted by the Met Office.

As a first step I'd suggest suspending your sensor in the free air in the middle of your screen, with the aspirating fan going, to see if that makes any difference. If it doesn't then it might have to be a case of living with the errors or constructing a larger screen out of less substantial timber.   

[h2005__uk__]

Thanks, that's interesting - I'd also wondered whether there was too much 'mass' in the screen which was contributing to overheating.

However, I painted the inside white and this, along with the fan running in the sunniest conditions, seems to have removed the spikes when the sun is out. The temperatures I record are very close - almost identical sometimes - to the local MetO stations (and closer than other unofficial stations).

For those interested, I used the 1937 Met Office Stevenson screen spec (and used option #4 for sheathed instruments) and slightly scaled down the measurements. I think the link was originally posted here ages ago but I've re-uploaded it - https://www.dropbox.com/s/8tmm78to11a4pdk/Met Office Stevenson screen.pdf?dl=0

I also created a SketchUp model which I'll post a link to as well.