Comparisons between NIWA stations and PWS's

[einzack]

I'm lucky here in Reefton because I have a NIWA station that is 460m away from my house.
I therefore routinely check my La Crosse WS2355 readings against those from the NIWA site.
I have on the whole been reasonably happy with how my station was tracking. I know I have issues with placement etc, but the restrictions with using a relatively cheap system such as mine with no modifications means it's never going to be perfect. I have a plan (and even bought the materials) to make my own Stevensons screen, but have yet to actually do it.

In early April this year I contacted NIWA as I had noticed that I was getting some significant discrepencies between the NIWA site and my own for all of March. The discrepency began quite abruptly and was reasonably consistent in that the NIWA site was reading 3-5°C higher than my own station every day.
I got a very nice reply from NIWA saying that they had also noticed an apparent discrepency with the Reefton EWS readings for the previous 6 weeks. The following week a site visit by techs showed a +3.6°C bias at ambient air temps (20/21°C).
Once they had made this adjustment, I was once again happy with the performance of my station.

The performance had remained relatively consistent since they made this adjustment in April. However, it appears the NIWA techs revisited last month, with no data recorded at the NIWA site on the 21st and 22nd of June. Since then there has been a marked difference between the temperatures readings.
The average variance from June 1st to 20th was +0.7°C and +0.4°C for daily max and min temps respectively (with my station recording lower). Since the 22nd, this has changed to -0.8°C and -1.1°C.
Just struck me as quite a significant adjustment that the NIWA techs made considering that from April to mid June, there hasn't appeared to me to be any marked change in the performance of the NIWA site compared to mine.

It has made me question the performance of my station, where previously I had been reasonably happy with it. Considering the quality of the NIWA equipment compared to mine, I feel I have to believe their data over mine, but I don't feel entirely comfortable in doing this knowing that they have made 2 calibrations in the last 3 months striving to get correct readings. Who's to say the current calibration is correct?

I've attached a graph showing the difference in daily readings for June, which clearly shows the adjustment.

• Tweet

[beteljuice]

Bottom line: the man who has one thermometer knows what the temperature is; the man with two is never sure ...

[Rwood]

It must be a little unusual for NIWA or Met. to make changes in such quick succession.

[Derek]

Bottom line: the man who has one thermometer knows what the temperature is; the man with two is never sure ...[/quote]
That is so true... as one who calibrates thermometers, this I know only too well 

I think that NIWA would be using platinum resistance thermometers or possibly very high quality thermistors, but I'd be surprised that they would actually do field calibration, normally instrumentation would simply be replaced and the suspect unit taken back to a lab for service / calibration. 

[Rwood]

Bottom line: the man who has one thermometer knows what the temperature is; the man with two is never sure ...

That is so true... as one who calibrates thermometers, this I know only too well 

I think that NIWA would be using platinum resistance thermometers or possibly very high quality thermistors, but I'd be surprised that they would actually do field calibration, normally instrumentation would simply be replaced and the suspect unit taken back to a lab for service / calibration.
[/quote]

Perhaps saves time or money? The SOE's top priority appears to be the bottom line.

[einzack]

went back to the email I got and it does state that they replaced the temperature sensors back in the April visit.
I assume they did the same again 2 weeks ago? if so, they don't last very long.....

[Derek]

Well those ones haven't lasted, but they should do.  Platinum resistance thermometers (PRT's) are so reliable and stable they are allowed 5 year intervals for recalibration - one of my reference PRT's has only shifted about + 0.002°C from its initial calibration 15 years ago...

[Weather Display]

maybe a build of moss or similar on the sensor?