Snow Drift on at Grade Deck w/o Canopy 2

[KootK]

I'm working on a project where the ground snow load is 175 PSF. That's about six times what I'm used to and has prompted me to consider snow drift on decks in a little more detail than I've done in the past. And it turns out that I have questions. How do we feel about the interpretations shown below? Consider the bits highlighted in yellow to be questions.

 

[IceNine]

Not sure if this helps, but in ASCE 7, the term 'roof' is used in every instance regarding snow drift requirements.
Perhaps because roof implies occupied space below, and a deck is unlikely to be occupied above or below during an extreme snow event, there is less concern.
That's not to say I wouldn't design for some amount of drift.

 

[KootK]

That's a data point, I'll take it. My gut feel is to design for the leeward case only but I don't know how to justify ignoring the windward case in an open field kind of situation. If you took my windward interpretation literally, you'd wind up at more less the same place even with 30 psf ground snow and a deck 4' above grade. And nobody's doing that to my knowledge.

 

[phamENG]

As you know, I'm not in a heavy snow region, so take this as you will:

Fetch is only infinite if it is infinite. Are there nearby houses , wooded areas, or other major obstructions? If so, I'd shorten it a bit. Even if it's in a wide open space, there has to be some limit to the amount of snow that can be reasonably carried on the wind. While the Great Blizzard of 1888 did leave some whoppers in New York, I'm not sure I'd be worried about that happening to a deck behind somebody's house in the boonies. And taking that to infinity would mean that even here, where I routinely ignore snow drifts because even with an abnormal parapet on a long roof it's still smaller than the roof live load, I'd have snow up to the eaves of a house.

I'd probably stick with 1.2 for Ct and use the full density. If anything, wouldn't the drift snow in the leeward case be heavier? If it's coming from a heated roof and had a chance to melt a little, then refreeze on the ground...seems like you could get some additional packing from that.

 

[KootK]

Are there nearby houses , wooded areas, or other major obstructions?

No nearby houses or obstructions. I'm sure there are woods around but I'd be relying on Google earth for the most part and who knows what the future holds. That said, I do feel that there should be some upper limit ground based fetch. What that is though... I don't know. It's weird when you start thinking about this magnitude of snowfall. The ground snowfall may be 4'-6' high without drift. With that in mind, is it really unreasonable to add a little drift and have it then up at the same 9' height as the adjacent roof?

I'd probably stick with 1.2 for Ct and use the full density.

In reference to the snow drifting off of the upper roof (leeward) or from the ground (windward)? In either case, your answer surprises me as it means that you're basing your thermal factor on where the snow lands, not where it came from, correct?

If anything, wouldn't the drift snow in the leeward case be heavier? If it's coming from a heated roof and had a chance to melt a little, then refreeze on the ground...seems like you could get some additional packing from that.

I honestly have no idea. A counter argument to that might be that I'd mostly expect it to be the light and fluffy stuff on the high roof that would blow off and drift to the low roof.

 

[KootK]

I've actually got two versions of this on the go right now if anybody would want to chime in on the other one: Link.

 

[Rabbit12]

How close is the depth of 175 psf of snow to the height of the roof eave?

 

[phamENG]

Fair point. In this case, it can certainly reach the roof line. And then...there's these sorts of pictures...

I've always used the thermal factor for where it comes to rest. Using ASCE 7, C[sub]t[/sub] is used not only to determine the flat roof snow load but also to categorize the roof as warm or cold. Based on that determination, different values of the sloped roof snow load are determined based on Ct and whether or not it's classified as slippery. So Ct is mostly about the ability of a structure to shed snow either through sliding or melting and draining. So if you have a pocket of ambient air under your 'ground drift', then you'd be unheated since the geothermal won't melt it as you pointed out.

Your counter argument makes sense, and I think it also supports my argument here. If you're only getting the light and fluffy stuff from the roof above, it never got within 3 feet of the roof surface. Why does it care what the Ct of the roof where it came from was?

Glancing through ASCE 7, I'm wondering if you are dealing with a different set of equations, though...where'd the 0.7 come from?

Side note - it feels strange reading up on snow as my AC cranks all day to combat the 90+ degree temperatures...

 

[KootK]

Throwing some numbers at it, it only takes about 90' of fetch to get maxed out to a 10' roof height.

 

[KootK]

Glancing through ASCE 7, I'm wondering if you are dealing with a different set of equations, though...where'd the 0.7 come from?

Just this from ASCE7-10

Your counter argument makes sense, and I think it also supports my argument here. If you're only getting the light and fluffy stuff from the roof above, it never got within 3 feet of the roof surface. Why does it care what the Ct of the roof where it came from was?

That's a persuasive argument, thanks.

 

[Celt83]

on mountain or off mountain?

This one is from mid march in Alta, Utah combination of drift and "normal" snow fall with the snow dunes reaching the roof levels pretty easily.

My Personal Open Source Structural Applications:

https://github.com/buddyd16/Structural-Engineering

Open Source Structural GitHub Group:

https://github.com/open-struct-engineer

 

[dold]

I'd design it for drift up to the roof line. I've definitely seen snow pile up to that extent on decks and otherwise up against houses. But...like you said, it's only 18" off the ground. Other than a potential code violation, what's the worst that can happen as far as the foundation is concerned - some settlement?

I took this pic late Feb of last year. Also at Alta, UT. Ground snow at Alta is 292psf!

I took it because I saw the massive pile and thought "hmm, what sort of lateral pressure might a massive drift put on to a wall?" The entire ground level of the house is nearly buried. Granted, there is a bit of a slope but the pile must be at least 10ft tall.

Any thoughts on that one, re: lateral pressures? I suspect the load would be minimal, if any at all. Snow seems like it would be fairly stable once it starts to compact itself...?

Side note: some of the guys in our office have designed buildings up at Alta for avalanche loads. They are rather...robust.

 

[phamENG]

Ah...that 0.7...of course. I usually ignore that equation because, again, snow load is so small here I can just apply ground snow load and still ignore it 95% of the time.

Taking a quick look at the commentary, it appears that the 0.7 is sort of a baseline exposure factor. So it sounds like you should drop the 0.7 for your deck if you're that low to the ground.

The normal, combined exposure reduction in this standard is 0.70 as compared with a normal value of 0.80 for the ground-to-roof conversion factor in the 1990 National Building Code of Canada. The decrease from 0.80 to 0.70 does not represent decreased safety, but arises due to increased choices of exposure and thermal classification of roofs.

dold - I can't point to a reference, but I looked into it once when I had a project with some snow drifts on the roof. I wanted to check the metal stud connections for lateral load. My research at the time suggested snow does not apply a lateral load. The lattice of ice crystals is stable enough - look at snow plows going through a large drift. You can get vertical faces of snow on each side several feet high. If you can cut away that material and have it not collapse, that suggests it's not applying a load laterally there.

 

[Enable]

phameENG: I'm not so sure about no lateral loads due to snow accumulation.

3.5.1 Ground snow depth
Windows located near grade are susceptible to lateral pressures from the depth of the ground snow. Drifting snow can cause a significant snow depth increase against the side of a building over and above a uniform ground snow depth. O’Rourke6 states that ASCE 7-05 remains silent on the issue of snow lateral pressure. He also indicates that a somewhat conservative but general approximation of the lateral load can be calculated by utilizing Rankine’s theory for cohesionless soil and a frictionless wall. O’Rourke6 provides an example of this calculation.

Rice and Dutton22 indicate that snow load lateral force was considered in the design of the glass facade for the La Cite de Sciences et de La Villette, Paris, France. Their lateral force rationale was not disclosed. A window in deep snow that is subject to a lateral snow force is shown in Fig. 3.17.

I don't have the full text, however, so I cannot verify anything beyond that quote.

Also, in this eng-tips thread civilperson tells us that vertical piles of snow next to metal sheet buildings have been known to crimple them. Same thread an J19 says ASCE actually has a design example with lateral loads, which I think might be on O'Rourke example referenced in the above quote (i'm far too cheap to buy it just for that though lol)!

I've never had cause to consider the matter in any detail. But this cursory look does provide some indication that there might be something going on even if it is only P-delta within the pile that causes a slope failure leading to a block of snow gaining momentum and hitting the buildings or something.

 

[phamENG]

Well then. I retract my previous statement.

Thank you for the correction.

 

[jayrod12]

Blizzard in Gillam last year, drifts 10 or 11 feet high for unprotected buildings. I'll try and find some pictures of it. Ground snow up there isn't even as high as yours.

 

[bones206]

Since the full height drift scenario seems plausible, using that as your design case seems reasonable to me. I would set aside all the various C-factors and roof load equations and just calculate the snow density using the code equation: γ = 0.13pg + 14 ≤ 30 pcf

Then just multiply that density by the eave height and design to that load.

If the design ends up being impractical, maybe consult with the AHJ to see what their take is. They would probably have a good sense of what’s typically done in the area.

 

[Rabbit12]

I agree with Bones. With the 175 psf load I'm guessing you aren't too far away from the eave anyway. With high snow loads like that it's crazy how deep drifts can get. I have pictures of Old Faithful visitor's center in Yellowstone that has snow piled up to the eaves.

 

[jayrod12]

I stand corrected, it was 2017.

This was the door to the arena. For what's it's worth, the design typical roof snow is 50 PSF up there.

 

[KootK]

@Everybody: thank you kindly for your contributions thus far. Fun pics to share with the architect.

on mountain or off mountain?

In a valley near the base of a ski mountain renowned for the amount of snow that it receives.

With the 175 psf load I'm guessing you aren't too far away from the eave anyway.

That's right. Per the MathCAD that I posted earlier, the base ground snow get us 6' up a 10' wall to begin with.

Any thoughts on that one, re: lateral pressures?

My thought is that I may barf:

1) The glass thing and what my liability exposure is on that.

2) These custom homes are all glass and no shear wall. Barely there lateral. I'm almost starting to wonder if drift lateral pressures should be considered for the lateral design of the building globally. As phamENG pointed out at the beginning, most locations have a dominant wind direction. As such, an imbalance of drift seems altogether plausible.

I was really hoping for a little more "you're nuts KootK, just design if for 60 PSF".