Snow on solar panels is worth a review in our climate because of our long winter, cold days and nights, lots of snow fall, and often cloudy days. I've collected the power data daily with the Fronius datalogger on my system. I really like the datalogger but if someone asked me "should I get one", I'd say it depends. Once the panels are on your roof you are not going to change anything quite likely. I get it simply as a way to better understand the nature of PV generation in our climate and for my curiosity factor. I also wanted to establish data on a system for sharing.
The impact of snow on solar panels is very significant on total production. That is production by the individual panels as well as the "string". [I have 20 panels that are 245 watts each. My inverter model 5100 can take up to 3 series of panels, called strings. There are 2 strings going into the inverter, each with 10 panels]. Here is is a little experiment I observed about solar heat and panels in the winter:
- snow covers the panels as the snow either falls directly on the panels or as snow in the wind drifts over the house and lands on the roof and panels
- snow and ice will melt during direct sunshine when the temperature is below 0 degrees C
- snow and ice won't melt during cloud cover and overcast days when the temperature is only a few degrees below 0C - just not enough heat coming through the clouds
- very cloudy days and snow covered panels produces virtually no energy [meaning, 0.05 kWhr per day]
So, just how much do we like the bright, direct sunshine during the winter when there is snow on the panels? LOTS! I observed my indoor/outdoor min/max thermometer and grabbed my infrared thermometer. [Of course, it is not a thermometer, it would be a thermal radiometer to be exact]. I happened to have a convenient solar thermal radiation absorber - a glass mirror with the dark backing facing outside that was propped up on our window sill. The sun shone fully on this dark backing and it served as an example of how much heat is in the sunshine. I had the indoor and outdoor temperatures in the thermometer. And I used the IR thermometer to measure the solar panel temperatures [approximate].
I photographed the temperatures at 10am on 9/01/2011 - a bright winter day. Our panels and window face south east so the sun was nearly straight on the panels. You can see in the photo below the outside temperature was -6C and indoor temperature was 20C.
 |
| Photo 1. Indoor and outdoor temperatures on 9/01/2011. |
I tried to measure the temperature of the solar panels. This requires an adventurous climb out a window and onto a lower roof below the panels. I used the infrared thermometer to measure the panel surface temperature. I am not sure what the actual temperature is - the glass could be reflecting heat, the cold sky temperature could be reflecting off the panels and into the thermometer, and the emissivity of the panels is not know. [The IR thermometer is set at assumed 95% emissivity of the object]. Anyways, the thermometer was between -3 and -4C. About a 2-3C temperature increase above ambient. I should do the same measurement on a cloudy day. Note that the roof is snow covered. The snow has melted and slid off the most of the panel area and to a lesser extent off the shingled roof yet.
 |
| Photo 2. Panel surface temperature estimated to be -3C, ambient is -6C. |
The second panel photo [photo 3.] shows several things. First there is still snow on the panels, but it is along the top and bottom edges. The snow slides and stops at the bottom of the panel on the aluminum edge. An ice dam forms and bunches up at the junction of the top and bottom panel edges. Second thing to note is the icicles forming at the lower edge of the panels. Clearly snow and ice has melted and dripped down the panel where it freezes in the colder ambient air. The panels must be above 0C and the outside air below 0C. Third thing to notice is the water drops and what appears to be melting snow. Even though the IR thermometer is measuring -4C, the panel surface must be a little warmer for the water drops and melting snow to exist.
 |
| Photo 3. IR thermometer measurement solar panel reading -4C while outdoor temperature is -6C. Note icicles, water droplets, melting snow, and snow remaining at upper and lower edges of panels. |
Finally, the thermal IR thermometer measured 51C on the reverse side of a mirror that was facing the sun in the south facing room. The surface is darker and has a low reflectivity and high absorption of solar visible and thermal radiation. This surface had no wind and cold outdoor ambient air to keep it cool. This shows the tremendous heat coming from the sun that can potentially be warming the panels. However, if this mirror surface was covered in highly reflective/low transmission thin material its surface temperature would be close to room temperature. The same holds for the solar panels that are covered with snow. The snow reflects most of the light and heat away from the panels. The snow starts to melt slowly from the bottom of the panels or where the snow very thin. Once a section of the panels are exposed to direct sunshine, that area starts to warm up. The melting front advances upwards and more panel is exposed.
As the panel warms and warm air starts to flow up under the panels snow can start to slide down the panels. I will post some pictures of snow that has accumulated on the ground show that snow clearly slides off in fairly large sheets.
 |
| Photo 4. IR thermometer measurement of back side of mirror that was facing outside on indoor window ledge. The surface temperature reads 51C while room temperature is 20C. |
|
There are some interesting observations about snow and panels. I have not tested these observations so I could be misinterpreting things:
- be aware of snow sliding off panels on sunny days - people and property may be hit with falling sheets of snow and ice
- the panels certainly warm up in the direct sunshine by at least 3C and I expect even higher - wind speed can reduce their heating rate
- ice and snow can slide down the panels and refreeze, often at the bottom edge of the panel at the aluminum trim
- diffuse sunlight does not significantly increase the solar panel surface and snow can remain on the panels for many days
- the low winter sun does not provide a long period for the panel and air to warm up during a sunny day - several sunny days may be required to clear the panels particularly if the air is cold and it is windy
I contacted several solar companies to ask about their systems and to get quotes. Nearly all of them were promoting the Enphase micro inverter. I have a string inverter and I don't have a similar system as mine to do any comparisons. However, the micro inverter was encouraged since they generate power from the panels individually. They are not subject to one shaded panel in a string bringing down the power generation. However, from my observations of snow on panels to date [1/2 winter] I normally have all the panels with at least some snow on them. Seldom, or for only a short duration, is one or more panels fully clear and others partly covered. This suggests to me that individual panels would also be experiencing shading and therefore reduced generation. Further study of snow on panels in Ontario, and in different installation configurations, is required before the argument can be made that micro inverters will have significant benefits over string inverters because of snow shading. My observations to days would suggest that solar developers should not over promote micro inverters solely due to increased generation during the winter - this may not have much merit and may only be suggesting the solar developers' lack of knowledge and experience.
Finally, a very rough approximation is that when the panels are half covered with snow, power generation is approximately one quarter of clear panel generation.
