Cumulus Support

I have had the B-L sunrecorder since June 2011 and it feeds my Cumulus 1

Until May of this year I had the VP2 without solar/UV, and then replaced it with a VP2 Plus ISS.
My Cumulus 1 still uses the B-L data www.komokaweather.com/weather/trends.htm
I have been testing CumulusMX on another console and it is using the VP2 Plus solar/UV and the Cumulus sunshine calculation www.komokaweather.com/cumulusmx/trends.htm. I have not changed the default settings yet,

I just happened to look at the MX trends pages and it was interesting on the difference. Attached screen shots show the Cumulus trends compared to the B-L sunshine hours. You will see, as expected the Cumulus 1 trends matches the B-L, and the CumulusMX is quite different as with my current CumulusMX default settings it is showing much more sunshine, with the B-L being the most accurate.

I know very little about solar/UV, lux, etc. but will try to do a little reading and see how I should configure MX to be more in line with B-L.

Enjoy,
Paul

The best place to start with your tuning is the solar graph in MX - on a nice cloudless day compare the theoretical and actual curves. You then need to change the Transmission Factor (if using the RS calculation) or Turbidity value (if using the Bras calculation) to get the values around midday to match the best you can.

Decreasing the RS Transmission factor will decrease the RS calculated values.
Increasing the Turbidity value will decrease the Bras calculated values.

Then you can compare the curves near Sun rise/set and see if the minimum cutoff value can be tweaked.

An extra complication that I see in using classic Cumulus is that here in Albuquerque New Mexico there is systematic annual variation. So I get the best average match between cloudless day reported solar and the "theoretical" curve year-round by tweaking the transmission factor to get a good match near the equinoxes. That way I'm off one direction in December and the other direction in June.

I find the same in the UK, *generally* in Winter the skies are clearer (when its not cloudy! ) and I up the transmission factor, in summer the air is more murky and transmission factor comes down.

Really it should be set on a day-by-day basis depending on atmospheric conditions, but we aren't going to do that.

Maybe adding a seasonal variation might be a good idea, and MX simply interpolates between the two values. We run this risk of making this thing over complex though!

A silly question but what is the B-L’s definition of sunshine - is it inline with the WMO definition?

The reason I ask is on the 9th of December you have >6 hours of sunshine. I don’t know Canada at all but going on location did you actually get more than 6 hours where the theoretical solar max was >120w/m2? I know here in London (UK) it is nothing like 6 hours.

Don't confuse the Cumulus MX theoretical value which is a global irradiation value on a horizontal plane, with that WMO definition which is 120 W/m2 of *direct* irradiation - ie on a plane directly facing the Sun.

A rough estimate+ puts 120 direct as roughly equivalent to minimum global values of 20 (Bras) or 5 (R-S) and a Sun elevation of around 5 degrees.

+ Using the solrad spreadsheet.

jrhilton wrote: ↑Mon 31 Dec 2018 10:21 am did you actually get more than 6 hours where the theoretical solar max was >120w/m2?

For standard cumulus, the threshold value above which detected solar radiation counts as sunshine time has two configurable values. One is the fraction of "theoretical", and the other is a simple watts per square meter. Here in Albuquerque New Mexico I configure mine for 75% and 20 W per square meter. I am unaware of what standards I may be violating, and just set the configuration parameters to give what seemed to me a result reasonably descriptive of the actual day on a year round basis.

archae86 wrote: ↑Mon 31 Dec 2018 1:35 pm I am unaware of what standards I may be violating, and just set the configuration parameters to give what seemed to me a result reasonably descriptive of the actual day on a year round basis.

You are not violating any, as above, the standard refers to direct radiation, our weather stations measure diffuse global radiation.

A silly question but what is the B-L’s definition of sunshine - is it inline with the WMO definition?

The reason I ask is on the 9th of December you have >6 hours of sunshine. I don’t know Canada at all but going on location did you actually get more than 6 hours where the theoretical solar max was >120w/m2? I know here in London (UK) it is nothing like 6 hours.

Interesting that you would ask and pick that date, as that had a period of about 1.5 hours were the B-L was offline and we did get that much bright sunshine that day. Also my B-L sensor is 20 feet above the top of roof mounted on unused chimney (about 40+ feet high) but there is still some tree obstruction when the sun is low in morning and evening

I recently have been starting my WL 6.0.2 to do some testing. I now have a VP2 Plus (UV/solar) and while in WL I noticed sunshine hours are recorded in WL reports. I have not entered any calibrations for solar and uv. In a quick comparison of the Davis/WL vs the B-L there is a huge difference in that Davis/WL is usually much higher than B-L In a 15 day period Dec 6 to 19:
Davis in Hrs/ B-L in hh:mm:ss
0 vs 0
1.8 vs 0
4.7 vs 00:05:26
4.0 vs 06:11:38 (Dec 9)
3.2 vs 0
0 vs 0
3.7 vs 00:20:04
5.2 vs 02:40:51
2.0 vs 00:21:12
1.7 vs 0
4.3 vs 04:58:01
0 vs 0
3.7 vs 04:42:16
4.7 vs 03:19:00

Attached are a couple of charts for Dec 9

I need to work on comparing the sunshine between the B-L and CumulusMX.

Enjoy,
p.s. I need to contact Ole as see that the B-L website is not available so my charts are not displaying.

I would ignore the Weatherlink figures completely - the algorithm is much too simplistic in just assuming that a suitable threshold for calculating BS hours from direct irradiance can also be used for total irradiance. The WL figures are probably better than nothing, but not by much!