I have been trying to find a chart that alows us to take into consideration snow that sits on the modules. Nrel dose have a spot for this but im not sure witch percentage to put into this spot. We have always taken off January and Febuarys production amounts and half of March since those are our snowest months and we have had months when it never slides in some elevations and locations.
They say Eskimos have hundreds of words for snow. There are so many different types of snow specific to different areas, that you have to know your area and take an educated guess or look at data from similar systems in your region.
When I was working in the winter in Toronto, it was cold and the snow didn't stick much. It would just drift.
I live in the Sierra Nevada foothills of California and it does stick, but is almost always gone in a day.
I spent some time in New Jersey and would see it also be gone as soon as the sun came up.
When I was in Mongolia and it was -40 degrees C = -40 degrees F, the snow would not stick and would just blow off.
If you are in a place with deep snow banks, I would expect a problem during those times of the year (which are usually not so filled with sunshine in the northern latitudes).
For the most part:
If it is sunny, the sunlight will melt the snow and heat up the PV. There also can be an effect of the shaded cells heating up a little.
If it is not sunny, the snow might slip off or don't worry too much if there is not any sun, because there will not be much production.
I measured the array at my house with an inch of snow covering the entire array and no sun to be seen. The 4kW system was producing 300W.
If you can get data from another PV system in the area, that is probably your best bet. I think that people usually overestimate the amount of losses due to snow, because when there is snow, production is not that good.
Usually the people most concerned about snow are the people that are off-grid and they will sweep it off.
My colleague had an O&M contract with a 2MW First Solar solar farm in NJ. They did a study about cleaning the snow off and determined that there would be module breakage and it would not be cost effective to clean the snow.
Thanks for the good question,