Davis VP2+ solar energy readings too high...?

[Tony_J]

I have just upgraded from an Oregon station to a Davis Vantage Pro 2+ which has the UV and solar energy sensors fitted. Today was the first day since I installed the station that there has been any sunshine - the solar energy sensor is clearly working, but the reading during sunny spells is a factor of 2 (or more) greater than the theoretical maximum value indicated by the Cumulus software for this lat/long. Does anyone have any light to shed on this problem? (pun intended)

[archae86]

The Cumulus number is more of a reference than a "theoretical maximum". It depends on your location, time of day, day of year, and the setting that you have for "trans factor".

I don't remember what the default was, but I just looked today and I currently have my "trans factor" set to 0.8. Here in Albuquerque New Mexico that value means that in mid-summer with clear blue sky the Cumulus provided maximum is somewhat higher than what I actually get, whereas in midwinter the Cumulus provided maximum is appreciably lower than I actually get.

But the really big error in this department comes with partial clouds. When the sun is striking your light detector directly unobstructed by cloud, reflections from nearby clouds commonly push the total light detected considerably higher than the blue sky value for that moment. On my station, and many others I have observed, this takes the form of a very busy graph ducking both well above and well below the "maximum" curve, while fully blue-sky days have very smooth curves indeed, with only a gradually changing difference from the maximum curve.

Enough of generalities, if you post a link to your station site I may be able to make comments more specific to your situation.

[Tony_J]

Thanks - that makes a lot of sense; today has been generally cloudy with occasional breaks in the cloud, so what you suggest about reflections would certainly apply. My web page is here:

http://www.jeffree.co.uk/suidhe/weather/trends.htm

Location is in the Isle of Mull, west coast of Scotland.

[archae86]

Your graph for today certainly shows strong cloud reflection effects. Given your location, it may be a while before you have a sufficiently clear sky day to judge the difference between the Cumulus "maximum" and your actual detection. I hazard a guess that it will be within 20% or so, unless there really is a problem.

Eventually you may wish to adjust the "trans factor" to give a better year-round fit between your displayed "maximum" and actual clear-sky values. As in my case I see a clear systematic offset between winter and summer solstice, I think you might for a first try attempt this match somewhere near spring equinox. Aside from the curve generated on one graph, the practical effect of this maximum is that it is used in the calculation which decides whether the most recent increment of time is added to the hours of sunshine for the day or not.

[Tony_J]

Thanks - yes, we don't see many clear days here at this time of year, so it may be a while. Much appreciate the help!

[archae86]

Some of the many systematic effects in solar observations as sensed by a Davis VP2 solar sensor and logged by Cumulus may be reviewed in these images from the Cumulus history for my station in Albuquerque New Mexico.

Clear day near Winter solstice

Clear day near Summer solstice

Almost clear day near spring equinox

broken clouds in spring


Some of the effects and artifacts visible in these graphs include:
1. I have a mountain nearby to the East, so sunrise here does not include direct sunlight for up to an hour after nominal sunrise. This issue affects all the graphs.
2. There is a row of Arizona Cypress trees to the West of my detection station, so depending on time of year, there is a sudden drop in detected light even on very clear days. This issue is prominent in the summer graph, not a factor in the winter graph, and intermediate in the spring graphs.
3. I have a neighbor's house to the Southeast, which blocks morning light even after the mountain no longer blocks it, for a little while, depending on time of year. This issue is prominent in the winter graph.
4. I have guy wires to stabilize the weather station mounting pole. Even though there are very thin (80 pound test braided fishline), they make a visible dent on each side of noon. A guy working on my roof last summer broke one, which I have neglected to repair, so only one dent after mid-summer 2015.
5. The Cumulus "maximum" calculation overestimates the summer to winter difference at my location.
6. The Cumulus "maximum" calculation has very slightly the wrong shape for diurnal variation here.
7. Broken cloud conditions give very spiky graphs, which many early users think signal some sort of station problem. It is entirely possible for short periods that reflected light from broken clouds added to direct path sun shining through a hole in the clouds add to more light than would be present on a perfectly clear day.

[Tony_J]

Those are very instructive graphs - thanks again. I can see that I should not worry about station errors and simply give it some time to see what kind of behaviour I should expect at this location. The winter to summer differences are quite striking.