How accurate are the OE weather station temperature guages?

[SUBYDAZZ]

Just noticed today that I had an abnormally high temperature reading compared to the official government site (about 5km away). My high temp was about 8oC warmer today. There were no abnormal spikes in my data / curve, all looked fine. I've noticed over the last 18 months that my temperature gauge always reads a little higher, but it's usually well within 2oC of the official one, so not sure if this is just a once off abnormality or convenient microclimate today or if the accuracy is actually getting a lot worse. Minimum temps are very similar to nearby stations.

I know a lot of people build their own Stevenson screen type thingies but surely the original device has a small one around the temperature sensor anyway. How bad can it be?

Do you need a fan with all the homebuilt type Stevenson screens?

I have a Digitech XC0348 type station that looks like this one (not mine): .

My temp/pressure sensor is mounted on a pole above the edge of my roof by about 2m so as not to pick up much if any reflected heat from the tiles.

My Station's homepage: http://users.on.net/~dazznsarah/pws/splash.html

[uncle_bob]

If your temp sensor is mounted up in direct sunlight, then on sunny calm days it's going to read higher due to the sensor heating up by direct solar radiation .

[mcrossley]

and being mounted above the roof I would say it is bound to get warm air rising up from the tiles every so often, especially when the wind speeds are low.

[MickinMoulden]

Yes, I think you would need a fan if your screen is also getting direct light, although a screen by itself without a fan would be better than having the standard sheild in sunlight. The fan would bring in more air that is at ambient temperature and expel any air that has heated up from contact with the inside of the screen, this is IMO.
Also, make sure that it isn't in an area that will heat up the air and to have that air pass over the sensor. ie not over concrete or, like in the picture (not yours, as you said), not over a hot roof.

[MickinMoulden]

I will also add, I reckon the temp gauge in the picture would read at least 10% too high in that installation.

[SUBYDAZZ]

So is it just because of the size of the OE screen compared to the sensor itself that causes the issues with high readings, because with the bigger Stevenson screens and other homebuilt ones, many of those don't have fans anyway?

My Station is installed above the roof but it's mounted well above the gutter so that tiles aren't directly under it and it's about 2m above them, so I think it would clearly not be as affected had it been mounted much lower above the tiles, although obviously not a perfect setup.

[MickinMoulden]

So is it just because of the size of the OE screen compared to the sensor itself that causes the issues with high readings, because with the bigger Stevenson screens and other homebuilt ones, many of those don't have fans anyway?

Yes, sorry but, size counts

My Station is installed above the roof but it's mounted well above the gutter so that tiles aren't directly under it and it's about 2m above them, so I think it would clearly not be as affected had it been mounted much lower above the tiles, although obviously not a perfect setup.

Your right, the 2m will be better than lower. Low down, IMO, will pick up radiated heat, but above you will still get the heated updraft. My unit was originally 3m up. No direct sunlight with no radiated heat from above as it was under the eave of my house, but was reading 5-10% too high as it was picking up the heated air from the driveway.
It's what you do with it that really counts

[MickinMoulden]

I also see you had 13% jump yesterday. At about 1pm 6/10/11, it went from about 21 to 23.8C. IMO, that's reading the updrafts. Sorry, not what you wanted to read. I had the same problem, but in the end, when I was telling work collegues "it was 39 C today" and them saying that it didn't feel that hot, it felt like I was telling lies. I thought it be best to give reputable readings so I could feel honest again. So I moved it. I had to buy 2 lots of 15 meter cable and 2 connectors to do it, but it is worth it. Is that sad or what?
Are you able to send a photo of your set up? You don't have to, I've got a pretty good idea.
PS I have seen units go up 8C in an hour, so weather is unpredictable. Maybe that one also was not reading pure air temp?

[AllyCat]

Hi,

The FO screen seems particularly poor in direct sunlight perhaps because the transmitter box actually touches the screen discs. It depends if you're working in degrees F or C, but IMHO the error can be much worse than +10%.

Temporarily (as a test), I've used a "quick and dirty" mounting beside a wall facing north-east (in the northern hemisphere), so the transmitter unit (inside the screen) can be in direct sunlight until about 10.30am (when it falls into the shade). Here is graph of the temperature a few days ago (when we had our "Indian Summer" with a daytime maximum of 30 degrees C). In the morning the false peak rises more than 6 degrees higher than the probable temperature (which I've marked with a white line).

Cheers, Alan.

[Gina]

The supplied screen allows 2 or 3 degrees C temperature rise from direct sunlight. My part wood and part uPVC Stevenson Screen reduces the rise to under 0.5C. The temperature sensor itself is within a degree C. I also have a 1-wire sensor guaranteed to within 0.5C which confirms it. I do not have a fan. My screen is fixed to the mast at about 5ft above ground level, the wind sensors are at 20ft height and the rain gauge a couple of feet away on it's own framework mounted 2ft above ground.