Pretty sure I am in the same boat with low UV readings. Station 1648 hit 5.5 and Station 3231 hit 7.7 today. I just don’t have any other nearby station to compare against.
1 Like
I tracked it today because of the warnings. We’re getting high temperatures and with a clear sky and thus high UV.
So I spent today with AWOS and readings close by that I know are accurate. If I had professional equipment I’d spend a week measuring and recording.
1 Like
Thanks for all the work compiling that list, @WLPD! I’ve added 18 new stations to the CL beta group for UV and you should see their readings get better soon. We’re a couple weeks away from pushing an update to the CL system to handle UV, so relatively soon all stations will benefit from the improvements being made there.
Note, one of those SKY units in your list (station 2170) has an extremely low UV reading. It may be un-calibrate-able, but it’ll be a good test case for the CL system. That system will ultimately not only auto-calibrate the UV and other sensors - it’ll also alert owners to problems (ideally before they are aware of them). The one SKY reporting NULL (“—”) UV is also reporting a failed UV sensor, so we’ll reach out to that owner directly.
3 Likes
Hi, you can add me to the way too high UV index Czech Republic
station ID 2986
1 Like
Will do. Since it looks like your station just came online, it may take a day or two for changes to be visible.
2 Likes
Below is UV index from Weatherflow stations across NZ at noon yesterday. For clear skies all but one Weatherflow station overestimates UVI.
I pulled data from NIWA to compare (a Govt funded Climate Institute). They measure UVI in a couple of key locations and predict for other towns and cities.
Its winter here, so UVI is at the bottom of the index. I hope that this data is useful for Weatherflow to add to the CL system and is a nice contrast to the current northern hemisphere summer data.
I had a look on Weather Underground, quite a few stations (non-professional) are also over-reporting UVI. Maybe I am expecting too much from a low-cost sensor. Si1132 data sheet shows it has reasonable correlation under clear skies but poor with cloudy conditions.
Solar radiation is much better. My comparison is rough, as I could not find direct data to compare, its midday MJ converted to watts.
Let us hope the Weatherflow AI will bring the UVI calibration closer.
| Location (approx) | Cloud cover | Sky UV | NIWA UVI | Sky Radiation | NIWA Measured Radiation | Station |
|---|---|---|---|---|---|---|
| Timaru | Clear | 3.6 | 1.1 | 348 | 325 | 2856 |
| Christchurch | Clear | 3.7 | 1.1 | 328 | 336 | 3772 |
| Christchurch | Clear | 1.1 | 1.1 | 343 | 336 | 2618 |
| Christchurch | Clear | 2.8 | 1.1 | 286 | 336 | 2595 |
| Christchurch | Clear | 4 | 1.1 | 385 | 336 | 2582 |
| Christchurch | Clear | 2.7 | 1.1 | 378 | 336 | 2627 |
| Christchurch | Clear | 4.7 | 1.1 | 407 | 336 | 2933 |
| Wellington | Clear | 5.8 | 1.4 | 377 | 2568 | |
| Wellington | Cloudy | 6 | 1.4 | 505 | 2729 | |
| Masterton | Cloudy | 5 | 1.5 | 478 | 2613 | |
| Levin | Cloudy | 4.1 | 1.5 | 261 | 2609 | |
| Palm. North | Clear | 4 | 1.6 | 456 | 417 | 2809,Device ID9662 |
| Palm. North | Clear | 3 | 1.6 | 458 | 417 | 2809,Device ID9849 |
| Napier | Clear | 2 | 1.7 | 328 | 439 | 2590 |
| Napier | Clear | 2.7 | 1.7 | 372 | 439 | 2592 |
| Whakatane | Cloudy | 1.7 | 1.9 | 396 | 2594 | |
| Hamilton | Cloudy | 1.2 | 1.8 | 158 | 2620 | |
| Hamilton | Cloudy | 5.9 | 1.8 | 486 | 2654 | |
| Auckland | Cloudy | 2.7 | 2 | 250 | 2593 | |
| Auckland | Cloudy | 3.7 | 2 | 216 | 1675 | |
| Auckland | Cloudy | 2.8 | 2 | 118 | 2626 | |
| Auckland | Cloudy | 2.4 | 2 | 240 | 2765 |
Excellent analysis, Peter - thanks! The production system is currently not auto-calibrating the light sensor, but I’ve added these stations to the light sensor CL beta group. Full roll-out is still a few weeks away.
I don’t think so - since your station can learn over time, it’s not limited to the one-time factory calibration that other stations are. Let’s see how the data changes now that they’re in the beta group.
2 Likes
Thanks David, Looking good, NZ stations in beta group are reporting UV index’s much closer to national predicted values.
2 Likes
yeah.mine is now showing a more reasonable 2.5 today (sunny day)
2 Likes
The UV at my station seems to be doing a little better. Today it was tracking real close.
2 Likes
Another day of good UV readings.
2 Likes
Morning. Two readings about 10-15 miles apart in my city (Sudbury, Ontario). His UV levels are at least twice as high as mine. For my northern latitude, his readings are wayyyyy too high (11 range).
Just sending these example for study. Looking forward to updates that improve other readings also. My lower readings first pic, his house higher readings 2nd pic.
Either that station is at a very high altitude, the location is incorrect or the sensor is faulty. That’s as high UV readings as we get here in Denver.
Is that Station 2211?
1 Like
This is definitely a calibration issue. We’re getting closer to adding a UV component to the network-wide continuous learning system, but it’s still only being applied to a limited beta group. I’ll add your station, along with the two others in Sudbury, to that group. Watch for those to improve soon.
1 Like
Thanks much. I know you’re all working hard there. Rome was not built in one day :).
Daniel
Sudbury.
1 Like
Yes his is station 2211. Mine is 3778. No high mountains in my area. Wish I had Denver mountains. :).
Max UV forecast for today was supposed to be 8, and that is not common. his sensor was better today or closer to mine anyway.
1 Like
Hi there,
I installed my new Sky last weekend. Unfortunately it’s not ideally sited, and doesn’t get direct sun until a couple of hours after sunrise. But it’s nicely exposed for most of the day.
I’m just wondering how long the calibration process should take.
In my area, the UV doesn’t get above 3 at this time of the year (late winter), yet my Sky is telling me it’s upwards of 12 in full sunlight middle of the day. (It rarely gets above 12 in summer.)
I’d have thought it might have sorted itself out by now.
Any ideas?
Thanks.
It can take a few cycles. The UV auto-calibration sub-system needs a set of data with clear skies for about an hour each side of solar noon. Looks like your system just updated recently.
1 Like
Hi David,
A follow up for the NZ stations that are in the beta group. Its looking pretty good! Maybe under predicting a little. We need to wait for higher UVI’s later in the year to really see.
A couple of stations don’t seem to be auto calibrating. 2594, 2620, 2593. They are excluded from the data. Maybe they were missed from the Beta group. Two new NZ stations 2589 and 4074 have popped up since last time too.
Compared to before calibration it’s amazing…
Take the data with a grain of salt, its using NIWA predicted data and I have no way of knowing the cloud cover condition at each station. It’s from different days over the last 10.
I’ll check again in a couple of months, the UVI should be up to 10 by then.
I was wondering if the app should have a UV warning (like it does for rain commencement)? I know you can do it via IFTTT. Determining a world wide warning level would be difficult as standards vary, AU/NZ its set at 3. US is 6. Maybe a user setting.
2 Likes