Temperature Inconsistencies

[Flosex]

140318 Temperature Inconsistencies.pdf

I have been running a Fine-Offset 1081 since 7 January 2014 and a Fine-Offset 3080 since 20 February, and have been trying to assess the accuracy of their temperature sensors. The attached pdf table shows the high and low temperature recordings for each of them in fahrenheit over a recent four day period, together with those recorded from my closest five stations (as shown on Wunderground).

Can more experienced members please tell me if the variations in temperature recordings are typical? Whereas I realise the differences between my 1081 and 3080 can be put down to an inaccurate sensor in one of them, I am very surprised at the differences in the five nearest stations to me which are all withing 15 miles.

The Franklyns Barn station is running a Davis Vantage Pro2, the Sullington station a Davis Vantage Pro, the Ashington station a Davis Vanyage Pro2 + FARS, the Churchill Lane station an OWW and the Billingshurst station a Maplin N96GY (Fine Offset WH1080).

Thanks for any comments.

Chris

[philcdav]

Hi Chris.

The variations seem excessive at first sight but a i'll bet the units and locations are all quite different.

A better 'view' of the data would be over several hours.

I ran 2 FOSH units side by side for a while. One had a white painted flower pot over the temp sensor.

Overnight temps were within 1 degree but daytime/daylight temps drifted further apart with the 'shielded unit always lower.

That's why the 'proper' stations use Stevenson screens

BTW the only way to get accurate results is with a calibrated device. Expensive and hard to maintain. Plus you would need a screen anyway.

Hope this helps

[beteljuice]

As phildavc says, a tenperature sensor with a decent screen / location is an important critera.

It doesn't matter whose kit you've got. - If you have a Davis unit mounted on the roof that is NOT the location for a thermometer !

In a similar vein a cheapie housing at the correct height but in full sun will also be suspect.