Re the Moscow market roof collapse. I’m curious how much load a given depth of snow would put on a roof per sq ft.
Any ideas?
In our Structures class at school, (part of the architectural curriculum) we consider snow to be a live load on a roof and measure it at 50 pounds per square foot. However, it should be noted that this number is the maximum that is generally accounted for in an educational environment.
The actual calculation of snow weight is slightly more involved and is explained in greater detail here.
50 lbs per sq ft? I’ve trying to picture a cubic foot wide square of snow. What’s that like 6 feet of snow height or something?
Well, snow’s just frozen water, so we can start from there.
I can always remember that 1 liter of water = 1 kg (well, at 4 degrees celsius, but good enough to estimate with).
Using google to convert to units I can understand, water weighs 62.5 pounds per cubic foot.
The weight of snow will be less than that, because ice is slightly less dense than water. Plus snow isn’t solid ice. There’s a lot of air in loose snow, although the lower portions of a pile of snow will compact due to the weight of snow above it. 50 pounds per cubic foot sounds reasonable. Water’s heavy stuff.
Typical snow doesn’t come in any where near 50 pounds per cubic foot. It is more like 10 - 12 although it depends on the density of the snow and how many freeze thaw cycles it has been through to make dense snow or ice.
Mostly ice would push it into the 50 lb range.
blace81 didn’t say 50 lbs per cubic foot, blace81 said square foot. Presumably the 50 lb figure is the engineering limit for the maximum expected volume of snow. The link provided by blace81 sez:
In addition, your link says:
still talking about square feet.
More direct to the OP, the first link also says:
or 62.4 lbs for a depth of 1 foot. So 12" of light fluffy snow would be a load of 1.87 pounds per square foot. An inch of ice would be a load of (slightly less than) 5.2 pounds. A load of 50 pounds per square foot would be (slightly more than) 9.6" of solid ice, or 4 feet of compacted snow.
It varies greatly. A very wet snow can have a density approaching that of water. A very cold, dry snow can have a density a tenth as much, or less even. So a roof with a foot of snow on it has a load somewhere in between if it had a foot of water and if it had an inch of water. I don’t know what sort of snow Moscow typically gets, or what they had in this particular case.
In a very heavy, wet snowfall, each 6" of snow will contain about 1" of water. The lightest, fluffiest snow I’ve personally encountered was almost a 20:1 ratio. Using your numbers:
1 cubic foot of “wet” snow =~ 10.5 pounds
1 cubic foot of “dry” snow =~ 3 pounds
Obviously, as you pointed out, the snow on the bottom will be compacted, but if you’re looking at how much snow has to fall out of the sky and stick to your roof before it exceeds engineering limits, then designing for 50 pounds per square foot would accomodate almost a five foot “wet” snowfall (equivalent to just under 10 inches of rain).
That’s a heavy storm, but it’s certainly not unheard of.
The other factor is that if you have a reasonable pitch to your roof (and if you live in an area with heavy snowfall, you had better!), then the snow will slide off the roof long before you accumulate five feet of it.
Most current building codes in the U.S. are based on ASCE 7-98, “Minimum Design Loads for Buildings and Other Structures” from the American Society of Civil Engineers. Chapter 7 has to do with snow loads. A roof is required to support a certain snow load in pounds per square foot of roof area. On pages 74-75 there is a map of the U.S. which shows snow loads for different areas of the nation. Here in Florida we have zero snow load, while the northernmost states require the roof to support 50 to 70 pounds per square foot of snow load.
Thanks for answers!
But you do actually receive snow infrequently. Do you have a lot of roof collapses when you do get a snowfall, if it is not engineered to take any snow load? I know this occurs in Vancouver, BC once in a while - they get a snowfall and it is disastrous for a lot of roofs. I guess I’m wondering how you engineer for something that doesn’t happen often, but when it does, it can be extremely expensive (and dangerous).
You can’t engineer a roof for no load at all, or the first time you climbed up there to fix the TV antenna or replace a broken shingle, the roof would collapse. Personally, I’m a big fan of overengineering. I don’t like building anything to minimum specifications, whether it’s a house or a piece of electronics.
Overengineering means never having to say “oops”!
Well, the snow load is only one of several live loads that are added up. The minimum design live load for a roof is 20 psf, which covers miscellaneous conditions, and there’s also rain and wind loads that total much more than any conceivable snow load.
Another problem is if the the water is not drained properly it will re-freeze and turn into a solid slab of ice. Much heavier of course. Where I live, in the Colorado rockies, we engineer roofs for 100 lbs per square foot. Imagine 6 inches of ice and perhaps 6’ of compacted snow.
I just put an addition on. I had the trusses engineered for 100 lbs per square foot for placement at 24 inches apart. To be safe, I ordered an additonal truss so that I could install them at 16" on center.