Thank you, I asked this simple question and you are the only one who gave me the down and dirty answer and it was a short answer!!!. Thank you
Hi @brad1 . Glad you’re asking. Here are direct links to official (from the WF staff) explanations of the CL system:
CL for Rain is not yet deployed, but you can gain some insights here: Rain Gauge Reading Seem Off (auto-calibration system not yet deployed)
but it still is a bit fuzzy system and I still have mixed feelings. If there is a big tree next to the sensor, it will measure less UV. The CL auto-callibration system doesn’t know that, and will adjust the reported values so it matches the other trusted sensors (or the predicted value from location, time of the year, ozon, clear sky etc). At first that looks like a good thing, but is it?
If the same thing would be done to temperature, I think we have a problem as I live in a city and temperatures are higher on average in town, but there is no good reference measurement in town to match it to. I expect my temperature measurement to stay higher.
the reported pressure now matches the airport 15 km to the south exactly. So good that one almost could start to wonder why we measure it in the first place. The only interesting part is that I now get real time values for pressure. But I wouldn’t be surprised if the airport would provide that as well.
Having said all that, I’m still happy that some calibration can be done, as the factory calibration of some sensors was pretty bad. If factory calibration would have been perfect, a yearly re-calibration could have been done. But then the system could know (learn) that my temperatures are higher and should be higher, and keep it that way. But that is not the case.
I would still love to see an option for user calibration. If I look at barometric pressure, there are times during the day it’s spot on, but then it will invariably begin to drift away from my VP2 which I know is correct. Eventually it comes back to being close to the Davis, but is guaranteed to drift away again. It’s almost as if the auto calibration procedure is making things worse, but to be honest, I don’t know if it’s the auto calibration that’s making it better for a period or worse. It’s frustrating because at times I’m thinking, “Yes, now you’ve got it!”, only to be pulled away from the correct values within a short time.
And humidity, don’t get me started.
Perhaps someone can help me understand the CL process a bit better. The Air unit’s barometric pressure graph is shown below and this unit was placed in service on 1/19/19. So it’s been in service about 2 weeks.
The behavior I notice is that often the Air’s barometric pressure is in line with my Davis VP2, but then for no apparent reason (see graph) will overshoot the mark and then gradually come back to a reasonably close value. This morning was a classic case. In the early morning my Davis was reading 30.42" and my Air was within .01" of that value. Then, within a 45 minute time period, the Air shoots up to 30.66", an increase of .25" (a very steep increase in meteorological terms, for that short a time period). Keep in mind that virtually all stations within my area (including my Davis) were either steady or saw an increase of .01-02" during that same time period.
Also keep in mind that no station, NWS or otherwise, had any pressure reading even close to 30.66. So I can’t see that the CL process was using other stations in my area as benchmarks.
So I can only assume that this behavior was somehow the CL process manifesting itself. But why after 2 weeks? If the correct pressure was being displayed on the Air for hours on end, why would the CL process increase the pressure .25" over the correct reading? It has since come back as I write this to 30.52, still .07" high and not the .01" it was prior to this ‘adjustment’. To me it looks like the CL process should have left well enough alone. At the very least it looks like there’s too much tinkering going on.
It actually makes me wonder if there’s some way to pull the ‘CL plug’ once you’ve got an accurate value for a parameter…assuming this odd behavior continues.
My suspicion is that this is not CL, because it were CL it would have had to run twice, once at around 8.00 and once around 8.30. I don’t think CL does run that often. It would be useless if it did as one really wants to see local changes, and not get adjusted to some other value every half hour. What is causing this, I have no idea.
Just out of curiosity…What timezone are you in? I was looking at your other parameters for a similar hump and looked at your brightness. you station only catches some light starting 1 pm?
Thank you very much!!! This is just what I needed.
you can do that with weewx if you run that, but be careful with you fiddling switches+dials (so to speak) and WF’s automated continuous learning stuff also tweaking tuning. Who knows what your effective result would be then…
Sunny, thanks for the reply and I agree there really seems to be no reason for the CL process to occur that often. It would seem at some point we need to have more ‘faith’ in our local readings.
I’m in the eastern time zone of the U.S. (N.Y.)
Brad, what was it that you needed? You’ve got me intrigued now.
Vince, I’ll pass on that approach.
Nothing special, its just WF Staff posted the CL information for me.
For the past couple of weeks my AIR - Mobile (5080) has been running about 0.5 inHg higher than my home AIR - Front Porch (5075). I first noticed it when I was camping at Deception Pass State Park, which is on the Puget Sound, so not much above sea level. When I got home I compared the station pressures and 5080 is significantly higher than 5075. Both are at the same height and location when I’m home. @dsj is this a calibration issue or something else?
Here are some screen shots of the two AIRs:
Hi @gizmoev, yes this was a calibration issue related to the station changing locations. We cleared out the cal data from the time device #14489 was in a different location, so it should look better going forward. Let us know if you see any other issues. Thanks!
Thank you. Would it have eventually recalibrated? If so, about how long would it take? When I move it, should I disconnect it from the internet during the move so if I’m driving over mountains and such it doesn’t get thrown off?
A few months back, I noticed a slight discrepancy between the air pressure figures at my station and those in the surrounding area. The last days above southern Europe is a field of extremely high air pressure, which is near 1050hPa. The WF only measured about 1040hPa. Specifically for my city, the data from my WF station was about 5hPa lower. I wrote yesterday during the day mail WF Support, but they still have not answered me.
it might be that your height above sea level and you height above the ground aren’t setup correctly. But for 5hPa that difference is about 50m. In your case the stations height should be around 270m and the height above ground should be whatever the height you placed the Air (probably 1-2m). If that is all set to the right values, support is needed. But, hey, give them some time
Hi @bocestar thanks for reaching out. The sensor appears okay so either reference sources for the Continuous Learning calibration are off for your location or an elevation value is off. Your height above ground looks okay.
It’s also possible there is some uncertainty with the SLP reduction with such a high pressure value.