Mitigation of Increased Design Snow Loads on a Sloped Roof

I was recently asked to provide advice regarding the suitability of the roof of an existing structure with regard to the `tons' of snow we have been getting the last couple winters. Though the owner noticed no `distress' to the structure, she was worried enough about the snow accumulations as winter progressed that she shoveled snow off the roof so as to not let the snow build up. Her motivation for calling on me was her wish to `not have to shovel', or more specifically, `not have to worry about having to shovel'; and if she doesn't (shovel), will the roof withstand the snow(s) we have been getting.

In terms of background, the design snow load(s) for the area in question were increased not long ago. The local building authorities felt that the design snow loads at the time were insufficient, and thus increased them. (Yes, there were some structural failures.) Presently prescriptive Basic Roof Snow Load values are provided for various parts of our county (by `Map'), as well as incorporation of a study, Ground and Roof Snow Loads for Idaho, University of Idaho, 1986. In other words, one can pull a roof snow load value off the Map, or have it site-specifically calculated.

And in our city (Moscow), some ten miles distant from the site in question, the design roof snow load was raised from 30 pounds per square foot (psf) to 40.

So, we could say that a lot of in-place roofs suddenly became `under-designed'. (Note that I did not say `unsafe'; and note further that I am not saying `under-designed' and `unsafe' are, or are not, the same. That (conversation) will be left for another time.)

Design Roof Snow Load

The structure in question lies right on the boundary of 40 psf and 60 psf Basic Roof Snow Load on the Map. (The Map divides the County up into four rather large zones of 30, 40, 60, and 80 psf.) As the structure is on the top of a hill, and as the owner described her situation being `much more snow' than Moscow, and since Moscow is 40, leaning toward 60 psf would seem more prudent. But since I am familiar with the more technical report referenced in the County Code, I used it to determine the site-specific load for the structure.

The report (Ground and Roof Snow Loads ...) provides a map of the entire State showing `normalized' ground snow loads, normalized meaning psf of snow load per foot of elevation. For the site in question the value is about 0.028 psf / ft. From a topographic map of the site I obtained an elevation of 3090 ft. So the ground snow load, p_g, for the site is ...

... 0.028 psf /ft x 3090 ft = 86 psf.

The snow that settles on roofs is generally taken to be less than that on the ground. The amount on the roof can be calculated using ASCE 7 Minimum Design Loads for Buildings and Other Structures (American Society of Civil Engineers), Chapter 7. The formula is,

p_f = flat roof snow load = 0.7 C_e C_t I p_g,

where,

C_e is an exposure factor,

C_t is the thermal factor, and

I is the importance factor (for the structure).

For this particular structure I chose (using the tables in ASCE 7)

C_e = 1.0 (neither too exposed nor too sheltered),

C_t = 1.0 (warm roof), and

I = 1.0 for a typical residence.

So,

p_f = 0.7 (1.0)(1.0)(1.0) 87 psf = 60 psf.

(NICE! ... same as the `heavy' side of the boundary on the Map.)

But the roof is sloped. Does (doesn't) that make a difference?

In many cases, `yes'.

ASCE 7 also provides adjustments for slopes, by way of a C_s factor from their Figure 7-2. The figure has three sub-figures for warm, cold, and `really' (my word, not theirs) cold roofs, and each of these shows two graphs for `unobstructed, slipper surfaces' and `all other surfaces'. The roof in question is `asphalt shingle', which falls under the `all other ...', and has a slope of approximately 5/12. From the Figure, C_s = 1.0, so the sloped roof snow load, p_s = C_s pf = 1.0 p_f = 1.0 (60 psf ) = 60 psf, no change.

For this particular roof (surface and slope) no benefit may be taken for the slope.

Sloped Roof Joists

The owner was particularly concerned with one lower and westward facing roof surface with 2 x 8 DF No. 2 joists @ 16 in. o.c. spanning 12 ft (in plan) and extending beyond the exterior bearing wall to provide 2 ft eaves. Engineering calculations (not included here, but an example of some pitched rafter calcs is in another article, here) indicated that the joists as is are over-stressed in the design load condition by between 10 and 20%. Hence, by present design standards, we engineers might say something like `not good'. (That doesn't mean the house must be immediately evacuated; after all, it's currently summer. Nor does it mean the house will necessarily collapse this winter, or next. What it does mean is that the in-place roof structure does not meet `current code' which incorporates certain factors of safety and the best prediction of future loads. Over the life of the structure there is a very real probability that the design snow load (or even greater) will be realized, and when that happens, we are beginning to crowd accepted safety factors (but in this case not crowd too much).

Metal Roof

So, what to do? For the particular roof in question, the owner is considering replacing the asphalt shingles with a (slippery) metal roof. Returning to Figure 7-2 of ASCE 7, for a warm roof of slope 5/12, a value of C_s = 0.7-ish is obtained. The design roof snow load becomes,

... p_s = 0.7 p_f = 0.7 (60 psf) = 42 psf.

Re-running the calculations for the joists using 42 psf instead of 60 psf puts stresses for in-place joists down into the `allowable' range based on current building codes. We engineers often say `good'. And I often say `REALLY good', adding the REALLY, as I sleep better at night when things are `good'.

And so, the owner's idea of going to a metal roof is (also) good.

Closing Remarks

Interestingly, since the design snow load around here increased by a ratio of about 4/3, and since the reduction for unobstructed slippery surface factor for this roof is about 3/4 (the inverse of 4/3), and a lot of roofs around here are that steep, or steeper, the increased anticipated snow loads can generally be accommodated by going from non-slippery to unobstructed slippery surface roofs. (Yes, I know; Dead load (weight of roof itself) must also be included ... which messes up the ratio thing a bit ... but it still worked out.)

The word `unobstructed' with `slippery surface(s)' is an important one. Obstructions such as valleys hang the snow up. The sliding snow must not be prevented from sliding; it must have a place to go; and the surface must indeed be slippery, to be able to `take the reduction'. The effects of obstructions and elements that prevent slipping are real. The snow will hang up on the roof, even if metal is used.

References

Latah County Building Code Ordinance, Latah County Building and Planning Department, P.O. Box 8068, Moscow, Idaho.

Residential Construction Guide, Community Development, City of Moscow, 221 East 2^nd Street, Moscow, Idaho 83843.

Climate and Geographic Design Criteria, Design Standards, Community Development, City of Moscow, 221 East 2^nd Street, Moscow, Idaho 83843.

Minimum Design Loads for Buildings and Other Structures, ASCE Standard ASCE/SEI 7, American Society of Civil Engineers, www.asce.org.

Bending Check of a Pitched Roof Rafter, Jeff Filler, Associated Content.

See also:

Snow Loads on Roofs and Decks, Jeff Filler, Associated Content.