Thoughts on Mitigation of Snow on Roofs

So, there is a bunch of snow on your roof. And it is more than normal, maybe way more. And you are worried that your house, or garage, shop, or cabin, or whatever, is going to cave in. First let me ask you the question: is removing the snow going to be more dangerous than the danger the snow itself poses presently? Roof replacement is always better than people replacement. Ask a lawyer; people are expensive. Before you climb up there to remove the snow, or send someone else up there, consider a few things to see if it is really that necessary.

1 . First, see if you have the design and construction documents for your structure. These `docs' may contain the `Snow Load' for which your structure is designed. Contact the local building department. If the snow is more than normal hopefully someone is keeping track as to just how much is out there. And, they will also know how much buildings in their jurisdiction should be designed for.

2. Consider drifting and other accumulations. Generally the snow on roofs is some amount less than the amount on the ground. However, this is not necessarily the case if there has been wind and accompanying drifting of snow. And it is not the case where snow has slid from one surface to another. So, while 18 inches of snow on the ground may be no alarm, if it has drifted to depths of 48 inches in some areas, that may indeed well be.

3. Also consider accumulations from maintenance work. Some of the most serious (and dangerous) accumulations I have witnessed were those from well-meaning maintenance workers clearing snow from around roof-top air-handling equipment, and, as the winter progressed, ended up making very deep and heavy piles of snow nearby.

4. Evaluate the present effect of the snow on your structure. Any roof built under modern building codes should have been designed/constructed to be stiff enough to resist their design loads while not sagging (deflecting) an amount more than, say, 1/200^th to 1/300^th of the span. (What does that mean?! Well, I'll tell you.) Let's say you have a wood-construction roof. Roof members are supposed to be designed to carry their design loads while not deflecting (sagging) more than 1/180^th their span, or 1/240 if crackable stuff like gypsum (`sheet-rock') is attached to them. So, in other words, if you have a modern wood construction roof, with sheet-rock ceiling, and your roof members span 16 ft, then the total sag should not be more than 16 x 12 inches of span divided by 240 gives 0.8 in. ... a bit over ¾ of an inch (under design load).

So, measure the distance from floor to ceiling at one end, and then the other. If the roof is straight the mid-span of the roof/beam/whatever should be the average of the two distances. If the distance from the floor to the mid-span point of is less than the average, the difference is the sag (deflection). If, for a 16 ft span, it is less than 0.8 inches, you can expect the load is less than the design load. If it is right at 0.8, then, hmmmmm, the in-place load might indeed be right at the design value. If the sag is greater than 0.8 inches, the load may be in excess of the design load.

Does that mean danger is imminent? ... not necessarily. There are factors of safety built into design. These factors of safety are generally on the order of 2 for a thing like a roof. The factors of safety take into account uncertainty in loads and materials (and workmanship). Generally the factor of safety regarding (future) loads is about 1.6. The factor of safety taking into account materials and workmanship makes up the difference. So, if the sag is right at 0.8, you know the load is `right at design' ... and that there is some factor of safety yet remaining in the system.

If, however, you measure a sag of 1.6 inches, for this same roof ... you can suspect that you are getting close to the edge of the actual capacity of the system. There is no more factor of safety to count on; the situation is not safe. (But see Note below.)

5. Indirectly measure the Snow load. Around here, this year, snow has fallen with a density of about 9 inches of snow equaling 1 inch of standing water equivalent (up to about 2 feet or so deep). Maybe where you live it is different. I'll address that. But what that means is that 1 inch of snow `weighs' ...

Well, let's do the math ... 1/9^th inch of water, or 1/9^th divided by 12 = 0.0093 ft of water ...

Water weighs 62.4 lb / cubic foot, so ...

1 inch of snow weighs ... 0.0093 ft x 62.4 lb / cu ft = about 0.6 lb per sq ft (psf).

At this density, then, 27 inches of snow weighs ...

27 in. x 0.6 psf per in. of snow depth ... gives ... 16 psf.

All the roofs I know of should be able to carry that.

So, a foot or two of fresh snow (around here, Inland Northwest) ... not to worry.

But, if it has been melting, or raining on the snow, the density can be much greater, ... two, three, or more times greater. That same 27 inches could mean 30 psf, or 45 psf. Presently, roofs around here are required to be built to carry 40 psf of Snow load.

But please understand me ... around here it has been falling at 9 to 1 ... but I can certainly imagine other regions, perhaps warmer, wetter, more `coastal' regions ... where it would fall heavier, perhaps MUCH heavier, than that.

And, there is information that, on average, once past a couple feet of depth, the snow is in general ... on average denser (University of Idaho, 1982).

6. As snow melts and takes on rain, the density increases. You might be able to gage the snow on roofs by what is on the ground, or you might not. I propose that snow falls and settles on a roof somewhat similar to that on the ground, but as things melt, the sun shines, the wind blows, and the rains fall, as buildings heat (and cool), the similarity loosens.

7. At some point you may need to get on the roof, and measure. And I have an article on the subject (here).

8. Decisions. So, let's say you determine that the load on the roof is near the design roof Snow load. Or maybe even more. If there has been no collapse ... fabulous! In my opinion one should only try and get snow off the roof if there are conditions forecast or reasonably possible that would increase the load. Such conditions might include more snow in the forecast, or rain. Initially, rain is absorbed into the snow and only adds to its weight (even though the snow depth will likely decrease). Another possibility would be wind, which may move snow from one spot and deposit it (even more deeply) on another.

Otherwise, keep people off the roof, turn up the heat in the building, and help it melt.

Note: for longer and longer beams and other roof members, typically they end up being extra strong, just to be stiff enough. So, even if the deflection is, say, double the `allowed' amount, it may still be safe. You'll need to hire a structural engineer to sort it out.

And, also note, that for shorter spans the roof may be at capacity even before there are significant deflections, so again, you might need a structural engineer to sort that out.

Again, these are some things to think about before you rush right out and send someone up there (to start shoveling it off).

Okay, maybe it really is that bad!

Then what? ... See Part 2 (here).

References

"Climatic and Geographic Design Criteria," City of Moscow Community Development Building Division, 221 E. 2nd Street, Moscow, Idaho 83843.

Ground and Roof Snow Loads for Idaho, R.L. Sack, A. Sheikh-Taheri, 1986, University of Idaho, Department of Civil Engineering, Moscow, Idaho 83844.

Measurement of Snow Loads on Roofs, Jeff Filler, Associated Content.

Mitigation of Snow on Roofs Part 2, Jeff Filler, Associated Content.