UV seems too high or too low (solved with auto-calibration CL system)

,

Thank You @dsj
Appreciate it verry much. :hugs:

@dsj

Today a bright and warm autumn day

The UV seems way better than the last weeks.
Did you calibrate the Sky?

This is exactly what we normally have around this time of year.
Verry happy with it.
No more extreme spikes.

Thumbs up for Weatherflow.
:ok_hand::+1::+1:

1 Like

Here are my UV readings over a couple of days.
Last 3 days the readings where way better.
This is how it’s supposed to be around this time of year.

smartweather

2 Likes

Hello @poider1982 Glad you noticed. We quietly launched the Continuous Learning system for UV in early October. See CL system for UV explained.

1 Like

Hope it handles when the Sky is briefly in the shade when the sun ducks behind a utility pole every day :slight_smile:

1 Like

Yup. And have that pole removed! Ha.

2 Likes

I assume that the system uses the actual station location and not the public location. Right now 5075 and 5080 should be reading the same or very nearly the same since they are right next to each other as shown in this post: Calculate wind vector-direction. How long should I expect it to take for the two stations to start to agree?

As mentioned in this post, roughly once a week,
Also if you still have values way out of possible values, contact support and they’ll have a closer look at your data

support via the website : http://got.wf/contact

support via mail : support@weatherflow.com

3 Likes

Good luck with that! :rofl: This is the final result from the last time that I tried to get the power company to move a pole:

pole-driveway2

8 Likes

Yes, the CL process uses your actual location. I took a look this morning and your “E Kelso” SKY didn’t have enough “clear days” to apply a UV calibration, which explains why they were not matching. I forced an override so they should fall in line today (both with reality and with each other!).

4 Likes

How many “clear” days are needed? Yesterday marked two weeks of clear days like this afternoon photo with my chimney blocking the sun. That being said, it looks like the upper E Kelso (5075) station was reading high as the calibration adjustment is clearly visible whereas Mobile (5080) shows only a slight drop in peak UV value. It is still lower but maybe after a few more sunny days they will come in line with each other.

I see that there are several spider webs floating between and possibly through the wind channel in Mobile. Can this likely mess with wind readings? SKY Mobile had some low wind readings when the upper E Kelso did not, usually it is the other way around.

Thanks for the follow-up. The CL process needs at least four clear days, although several other criteria must also be met. I just took a look and both of your SKY units received another calibration update today. Please report back how things seem over the next couple days.

1 Like

Must have been some of the other criteria wasn’t met. When I saw how clear it was today I wondered if another calibration was going to be done. Sure enough! I just checked and the UV index is different by only 0.1-0.2 now. :smiley: Looking forward to the other CL processes to get implemented. Keep up the good work!

1 Like

@dsj my uv seems to be still reading low. Today’s official recording was 6, I recorded a max of 4. Station 5008
Gary

1 Like

UV is showing mega high levels here for the time of year. Sky was showing 6 earlier, it shouldn’t be any higher than 2.

1 Like

Neither of your SKY units have seen enough clear days yet for the CL process to suggest an update, but a couple humans looked at the data and have applied what should be a pretty good one. You should both see improvements starting today.

Hi David,

The UVI is rising in the southern hemisphere. Daily clear sky maximums today are between 6 and 8 across NZ. I had a quick look at NZ stations you added to the beta group in an earlier post. It seems they are under predicting UVI (about 30-50%).

I’ve have temporary added third sky to my collection so I have 3 Sky within 5m. First two where those that where added to the beta test group although under predicting are have stayed nicely aligned together. Number 3 has been in operation for about 1 month but the CL system has not corrected it yet, is about 3x the other two (all are on 2809). Blue-sky days are rear, but its sunny today (-: and the predicted max is about 7. I cannot find any official/trusted public source for my city. The nearest is 90km away and was 6.4 measure at 1pm today. Thanks.

1 Like

Much better today! A max of 1.8 which looks to be perfect. Thank you! :ok_hand:

2 Likes

Hi Peter. we just had a look at the last couple weeks of data and while the automated CL system did not have enough data to confidently suggest a change, we have a applied a tweak to all three that should bring them better in-line with each other and reality. The CL process is running daily now and it will apply an additional correctly if/when it collects enough valid data.

Glad to hear it!

Thank you David. Great support as usual.

All three are close aligned now. Values look close to what is being predicted. Will need wait for a blue sky to really see but rain forecast for the next week (rain gauge testing time (-: )

2 Likes

@dsj my UV seems to be out by quite a lot still. Reading today was 5.2 and should have been around 7. We have had plenty of fine clear days so that shouldn’t be the issue.
Gary

1 Like