davis uv sensor

[jerryg]

I got a used uv sensor and have been comparing it to the two nearby metars for the last few days and when they are both reading 13 mine shows 11.7. My uv reading starts about an hour later than when the metars start reading in the am and mine shows 0 an hour before the metars. Is this normal? How close should they be to the metar readings? There is not alot of info about the uv sensor on the forums. I saw the specs from Davis but what is the real world accuracy of the sensor? I have never been really into the uv reading with it available from the metars and on clear days should be the same everywhere in the area but now that i have one my interest is peaked lol. If it is off i can send it in to Davis for refurb alot cheaper than i could buy a new one but need the info to decide if it needs to go in.

[johnd]

You get UV in your Metars? What resolution are the Metars reporting to? Integer UV Index or 0.1? If it's integer then I'd say it's possibly not a big enough difference for concern, because a UVI of 13 could be eg 12.6 and not 13.0. And then what accuracy might the Metars be reporting to - might just be 5% or even 10% depending on the type of sensor etc. And then the difference between say 11.7 and 12.6 would just be around 7% - potentially well within the combined specification of the two (ie Davis vs Metar) sensor readings.

But if it was consistently 11.7 vs 13.0 then that's obviously more of a difference, though perhaps not much more than borderline in significance.

Presumably the altitude of the all the sensors is not significantly different - higher altitude is obviously going to give higher UVI potentially.

[jerryg]

One of the metars is 8 miles nw of me and the other is 12 miles se of me, they both read the same on clear days so i feel comfortable in thinking i should read the same being as i am near halfway between the two. They are read out in whole numbers up to a max of 16. The altitude is not an issue here on the coast of Texas nearly sea level. I know Davis gives a 5% spec and i was not sure how close others that have a sensor have noticed how they compare to the metar near them. I am right between the two and never really saw the need for uv sensor but now that i have one it gives me something else to play with lol.

[johnd]

Ok, so in integer values, you're reading 12 and they're reading 13 (in your particular example) - a difference but not a major one. As you say, the nominal accuracy of the 6490 sensor is 5% and then there's a 2% potential annual drift in the specs. Did you say yours is a used one? So if it was eg 4 or 5 years old then it wouldn't be surprising for it to be off by say 10% compared to a brand new one side by side.

So I think the answer is that yours might be reading a little low, but not by much and probably well within expected margins of error given all the uncertainties in the various readings, locations, age of your sensor etc involved.

[jerryg]

That is kind of what i thought and was just more or less wondering how close a new one would be by comparison.

[dalecoy]

Did you clean the diffuser?  Is it possible that some time in the past, the diffuser was improperly cleaned, and is now more/less diffusing?

Maintaining the Sensor
For the most accurate readings, clean the diffuser after mounting, and then periodically.
Use ethyl alcohol (not rubbing alcohol).

Also of note:

Due to the sensitivity of ultraviolet and solar radiation sensors, it is common practice
for manufacturers to recommend recalibration after a period of time. Here at Davis
Instruments, we have seen approximately 2% drift per year on the readings from these
sensors. For applications demanding higher accuracy, the sensors should be calibrated
once every year

[jerryg]

I have no way of knowing the history of the sensor it came with a used station i bought. I did clean the lens before i put it outside and i think it is probably an age issue. It does not seem to be as sensitive as i think it should be because it is over an hour before i get a reading compared to the metar and quits reading an hour or so before the metar in the pm. Since i never had one i thought i would seek some input as to how they function when they are newish. Hard to tell anything if you do not have a frame of reference to go by. If i get more enthused about the uv reading i will send it in to Davis for a refurb. I am just not too sure how much use the reading is.

[graculus]

I didn't know that metars could include UV, do you have a link to them?

[jerryg]

I guess most metars report uv. I get my readings off of weather underground  site for my area. Wu shows the nearest metars to me.

[johnd]

I guess most metars report uv.

Not so far as I know (which was why I was rather surprised at the suggestion), but maybe there's a newer or US-only Metar specification? Though, equally, Metars are for primarily aviation use and it's not obvious why UV readings would be any use to flight planning. Plus, all the well-known Metar specifications - eg http://en.wikipedia.org/wiki/METAR - don't mention UV AFAICS.

So I suspect that your WU readings are picking up UV data from elsewhere than a genuine Metar. But that does perhaps imply that they're less 'official' than a Metar would be and so potentially less accurate (though I've also seen plenty of inaccurate Metars too).

[dalecoy]

I presume you don't have a solar sensor.

OK, how long after sunrise do you start getting UV data?

Here's my info for comparison:  http://www.wunderground.com/weatherstation/WXDailyHistory.asp?ID=KMOLEESS9

Sunrise today 5:55 AM
First solar data (.09) at 6:00 AM
First UV data (0.5) at 7:50 AM

Of course, there's a "latitude" effect.

[jerryg]

Firstly, i thought it came from the metar locations. My wu shows four sites in my area and they all show uv, they read differently depending on clouds etc. The two i am between read the same with clear skies and read different with cloud cover over their site. I do have solar sensor and i got my first solar reading at 6:31 and my uv at 8:30. So my uv lags solar by 2 hrs in the am. I can see the airport from my location and when i see its cloudy over the airport the uv reading goes down and back up when the clouds move out so i guess the sensor is at the sites.

[dalecoy]

I do have solar sensor and i got my first solar reading at 6:31 and my uv at 8:30. So my uv lags solar by 2 hrs in the am.

Well, I presume you noticed that my UV lags my solar by about 1:50 in the AM (today).  That's close enough to 2 hours, and differences in latitude, humidity, etc. can easily account for small differences.  So, I'd say that based on a single sample, your Davis Solar-and-UV setup is working exactly like mine.

What hardware and software is in use at those other WU stations?  ("About This Station" section)?

Could you post the station IDs you are looking at, as well as yours?

[jerryg]

Ok that answer some of my questions about the uv sensor. The uv readings i am refering to are on the wu site for Victoria, Texas and is real time on updates not a forecast type. Here is the one i look at most http://classic.wunderground.com/cgi-bin/findweather/getForecast?query=zmw:77901.1.99999  notice the uv reading, how it gets there in real time i just assumed it came from the site.

[dalecoy]

... and is real time on updates not a forecast type.

I'm not sure what you mean by that.

Here is the one i look at most http://classic.wunderground.com/cgi-bin/findweather/getForecast?query=zmw:77901.1.99999  notice the uv reading, how it gets there in real time i just assumed it came from the site.

When I go to that link, the current conditions box is labeled by default "Colony Creek Area, Victoria, Texas (PWS)".  When I look at Colony Creek Area, Victoria, Texas (PWS) which is at KTXVICTO11, http://classic.wunderground.com/weatherstation/WXDailyHistory.asp?ID=KTXVICTO11 -- I see that it's not reporting solar or UV, and Hardware: Davis Vantage Vue.

The only listed nearby WU reporting station that is reporting UV is KTXVICTO6, and I assume that's your station?

And I have no idea where WU gets the UV data for "Current Conditions" in Victoria, TX.  Perhaps you'll have to ask WU?

Or in the Weather Underground forum on WXforum.net at http://www.wxforum.net/index.php?board=90.0

In either case, when you find out, please come back to this topic and tell us the answer.

[jerryg]

Yes o6 is my station. the airport is kvct. What i mean by real time is with each update of the airport info the uv can change based upon cloud cover or time of day. It is not a generic forecast for the day. All four of the airports that are shown in my area all show uv on the conditions part. I just thought the uv was reported with the equiptment at the airports, i do not see how else you could get real time uv readings without a sensor at the sites. The readings at the different sites are different so that rules out an area type reading. Guess i could query wu about it.

[jerryg]

I just emailed wu and ask them where the uv readings come from on the airport current conditions. Will let you know what i get from them.

[C5250]

That could be interesting... AWOS stations do measure UV, but it is not a part of a normal METAR.

[archae86]

My uv reading starts about an hour later than when the metars start reading in the am and mine shows 0 an hour before the metars.

An oddity of the Davis UV sensors is that as reported through the ISS they have a distinct threshold at 0.5  Below that they report zero.  So the dynamic range on a day that peaks around 12.0 here in Albuquerque has a reportable UV dynamic range of just 24:1.  By contrast, the Davis solar sensor will report a minimum value of 5 or perhaps less on a day with the peak over 1000.  Thus it has a practical dynamic range here at the moment of at least 200:1, perhaps more.  So, although at my site at this time of year the beginning and end of both UV and solar are badly compromised by obstructions (Sandia Peak in the morning, and a row of Arizona Cypress trees in the afternoon), it is routine for the non-zero UV reporting to begin and end over an hour inside of the solar reporting.

I should mention that aggravating the artificial thresholding difference, the real scattering of UV by low path near sunset and sunrise is more than that of aggregate visible radiation, so one should expect the UV to have a different daily shape in a way which accentuates the thresholding artifact.

I've had the UV and solar sensors since January 2012.  On May 27 I had my first try at cleaning them, taking care to do so near solar noon on a cloudless day to see what difference it made.  I used 190 proof Everclear, and found no observable change in solar reading, but about a 5% increase in UV reading.  We had not had much rain in a while, and had at least a little smoke from a massive fire in SW New Mexico--so my sensors may have been dirtier than average for that amount of exposure.  

[archae86]

One of the metars is 8 miles nw of me and the other is 12 miles se of me, they both read the same on clear days so i feel comfortable in thinking i should read the same being as i am near halfway between the two.

You may be able to find some people with Davis UV sensors reporting to weather underground not so very far from you to get an idea of typical station-to-station variation and light curve shape.

While it won't find people with add-on sensors, one method is to check the personal reporting station list at:

http://www.wunderground.com/weatherstation/index.asp

setting the scroll box to Texas, then searching the resulting list for "plus", as most of the Davis Vantage Pro 2 Plus stations report UV curves to wunderground.  I don't know the map of Texas well enough to guess which of the many dozen Texas stations probably reporting UV detected by this method might be near you based on the town name, but I imagine you do.  Clicking on the station name at left will give you the data they are reporting today, and clicking on the previous day link a time or two or three will probably find a nearly cloudless day which can make comparing the light curve more direct.

[kaymann]

My Solar to UV Senors here in AZ (DW4959 KAZMARAN7):

UV consistantly one hour later than Solar and UV does not kick in until .5 threshold is met..

[jerryg]

The .5 threshold makes sense in that the change over point for noaa is .5, so 1.5 shows to be 2 and 2.5 shows to be 3 etc. Then 1.4 is 1 so anything below .5 would be 0 and that is that .lol.

[dalecoy]

My Solar to UV Senors here in AZ (DW4959 KAZMARAN7):

UV consistantly one hour later than Solar and UV does not kick in until .5 threshold is met..

If you get a chance, please check the Browse window data and see if the lag is not closer to 2 minutes.

In any case, we can clearly see that jerryg's Davis UV sensor is performing just like other Davis UV sensors (relative to Davis solar sensors).  So there's apparently nothing "wrong" with it.  It's just how Davis UV sensors work.

Where WU gets UV data, and what kind of sensors, etc. is still an open question, of course.  But we resolved jerryg's real (initial) question.

[kaymann]

Actually looking closer it is actually .1 on first reading and within the 10 minute arc int it then jumps to .5 index on the next reading.  I used to have a 15 minute interval and under that arc interval it was always .5 (probably the max by the end of that 15 minute reading) Given that it rises about .5 for first couple of intervals that would make sense.

BTW it it is still more than 40 minutes after first solar energy reading given going through so much atmosphere probably filters out alot of UV.

[graculus]

That could be interesting... AWOS stations do measure UV, but it is not a part of a normal METAR.

I asked WU for the source of UV number they are showing for my local airport. Turns out that it's not a measured value:

"We compute UV Index data based on the time of day, elevation, latitude and cloud cover.  UV will update on our site as frequently as the local airport updates conditions which should always be at lease once an hour."