Roof snow removal 16

[mijowe]

I have been ask by an owner for the number of inches of snow fallen on a roof that should trigger its removal. This is for a building that I designed and was recently completed.

The simple answer is to provide the snow load I designed for, in psf not inches. I feel like I have not exposed myself to more liability if I define what I designed for. Without getting compensated I do nto want to make a recommendation for their snow removal plan, and I also do nto want to provide a conservative number that would have them cleaning off the roof more that necessary (it is a fairly large roof).

I have never been asked for this and am curious how others address the question.

 

[KootK]

The correlation between snow depth and weight is certainly imperfect. That said, the structural design community has implicitly deemed the correlation adequate for our own purposes. That's how our code drift load provisions work, by relating drift depth to drift weight.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[DaveAtkins]

Aside from safety issues (staff slipping on roof, or worse, falling off roof), shouldn't we be concerned about damaging the roofing materials? I can see someone with a metal shovel slicing open an adhered EPDM roof membrane...

DaveAtkins

 

[CANPRO]

Drift calculations require an assumption on the density of snow as mentioned - It might be worth pointing out what the NBCC has to say about that (Commentary G, paragraph 7):

“To calculate loads due to snow on roofs, a measurement or good estimate of the unit weight is necessary, The unit weight of snow on roofs, gamma, obtained from measurements at a number of stations across Canada varied from about 1.0 to 4.5 kN/m3. An average value for use in design in lieu of better local data is gamma = 3.0 kN/m3. The unit weight of snow may be considerably greater than 3.0 kN/m3 in some locations such as regions where the maximum roof load is reach only after contributions from many snowstorms, coastal regions, and regions where winter rains are considerable; in such locations, a unit weight as height as 4.0 kN/m3 may be appropriate.”

Perhaps any recommendations for correlating snow depth to weight should include a similar precaution provided by NBCC Commentaries.

 

[BUGGAR]

The basic issue is probably to avoid roof collapse. I saw a sales guy in my local big box giving a demo of laser distance measuring gadgets. Just buy the owner one of these and have him "shoot" the distance from floor to midspan of several key beams (or?) when the roof is unloaded and then again during the snows. He can shovel at >l/400?

 

[Teguci]

2 things:
- The snow needs to be removed before it reaches the design load (not the next morning).
- The snow removal has to be staged so that it doesn't exceed the design load during removal (no 10 ft pile of snow waiting to be pushed off when the dump truck returns).

I don't think there is an option but for him to clean "off the roof more that necessary."

 

[MotorCity]

If the roof was designed per code requirements, there should be no need to remove snow from the roof unless some other load has been added to the roof since it was originally designed.

Code prescribed loads are meant to capture the maximum expected load for say 90% of the time, think bell shaped curve. The codes do not guarantee that those loads will not be exceeded, as such a claim would be ridiculous given the unpredictability of mother nature.

To design a roof with the expectation of the owner to shovel snow of the roof once it reaches a certain level is, in my opinion, below the standard of care expected of engineers.

 

[DETstru]

I've seen a number of roofs over the last two years that experienced severe deflection due to the heavier than usual snows and cold temperatures we've been getting (no failures but damage to nonstructural components). All of these cases were not due to snow alone, it was always a combination of a layer of ice with snow on top.

My opinion is that it would be irresponsible to give an "inches of snow" number due to the variable density of snow itself and the potential for the presence of an underlying ice layer. Additionally, you have areas that may be designed for drift, which would need different numbers.

My approach would be to tell the owner what design load you've used for the roof, how you came up with that load (code reasons, etc), and explain why you can't give a precise number for "inches of snow."

 

[msquared48]

wannabese:

I know that, but it appeared to me that the OP's client was asking for a far more strict definition of the limits than what we are mandated to provide by code.

If you think about it, a 3O psf snow loading could range anywhere from 6" of ice, to 2.5 feet of powder snow.

Mike McCann, PE, SE (WA)

 

[jtdpsu]

mijowe:

i think it's reasonable for a building owner to ask his engineering consultant--a presumed expert with respect to buildings--how much snow his roof can handle. they don't understand what "20 psf" means--he wants a height of snow. i think that's a reasonable request. i'd explain that converting a uniform snow load into an equivalent height of snow isn't a straightforward thing due to variable density of the snow--give him both ends of the spectrum. "worst case, really heavy snow, we recommend removal upon xx inch accumulation". done.

 

[RHTPE]

As I understand it, a powder snow here in New England equates to about 1" of water for 10" of snow. Obviously this varies depending on the kind of snow that has fallen, and is nowhere near being a standard. 1" of water weighs 5.2 PSF. The easiest way to be sure is to take a core using say a 6" PCV pipe, making sure that you've collected all of the snow down to the roofing. Melt the snow and determine the water volume - from there one can work the numbers to determine the weight of water over 1 square foot. This may not be an easily understood math problem for an owner, but it is an accurate way of determining the actual snow load. Using this method, one could easily develop a table or graph to translate volume of water into PSF.

I think that JAE has it right:

The only thing I'd suggest, for the owner here, is to tell them the design psf, then perhaps offer a range of approximate depths where the code safety factor is reached. Then educate the owner further about how these densities can vary and if there's any question about too much snow, a quick study should be done on the roof to see what is truly up there.

It is my strong belief that we owe our clients a bit more than simply stating the design snow load in PSF.


Ralph
Structures Consulting
Northeast USA

 

[KootK]

For a spell here, I'd like to set aside our collective litigaphobia and consider simply what we believe to be the right thing to do. It's an ethical discussion for another thread but I truly believe that liability is minimized, and positive outcomes maximized, when all parties simply do what is right. When all parties seek only to minimize their own risk exposure at any cost, I think that a "tragedy of the commons" kind problem develops. Things fall through the responsibility cracks and real problems crop up.

In this instance, we have an owner that has requested access to an engineering parameter so that he may use it to make future decisions. As engineers, this is our wheelhouse: providing people with technical information that they can use to make good decisions. Further, the owner wants to use the information to clear excess snow from his roof and reduce the likelihood of it collapsing. And if he's successful in that endeavor, that success itself will serve to reduce OP's liability in this situation more than any CYA measures that OP could take.

So, what's the right thing to do here? That seems pretty obvious to me. The right thing to do here is to supply the owner with the information that he needs: a simple metric, snow depth, which he can use to make his decisions. I believe that letting the owner convert PSF to depth on a snowstorm by snowstorm basis is fundamentally flawed because the likelihood that the owner will take on such DIY meteorologist duties is slim to none.

Will there be accuracy problems with a snow depth estimate? You bet. Disclaimer away!

Should you get paid for the associated risk? You bet. If you didn't build this into your original contract, and you don't mind annoying your client, nickel and dime away!

Will the roof somehow be less likely to collapse if you refuse to provide a snow depth and force the owner to guess/DIY? Not. A. Chance.

Alright then, back to our regularly scheduled program: lawyer avoidance.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[JAE]

No one is forcing the owner to guess. And I for one am not responding to this with liability in mind....only practicality for the owner.

The issue is that the parameter he is asking for is terribly flawed.

I would rather take the time to educate the owner and give them a rough range - but qualify your number with "it can vary significantly".

Keep in mind that our variability of snow load (in psf terms) doesn't vary to the degree that the snow depth does.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[msquared48]

JAE et al:

If you must be Mr. Nice Guy here, at least place it in writing.

That said, liability is involved. The courts are still very likely to hold you responsible if the client misunderstands or misapplies the instructions, written, verbal, whatever - the court is very likely to say that it was not clear enough and place the blame on you regardless since you are the professional.

That is my experience, and I for one will stay away from it.

Cheers.

Mike McCann, PE, SE (WA)

 

[Triangled]

I am curious to see a recommendation for descriptive verbiage to satisfy the client's interest, say for a 30psf roof snow load on a flat roof, to keep it simple

 

[Engineerataltitude]

In the Sierra Nevada, LA department of Water and Power does regular snow surveys to determine both snow depth and water content. They have been doing this since the 1920's. Based on that historical data, the weight of a foot of snow is right around 25 psf. I use that all the time to help people with older roofs in snow country to know when to shovel snow off their roofs. Been doing this for 25+ years with no problem.

Personally I think that if a good engineer has a thoughtful, reasoned basis for his/her recommendations they shouldn't be afraid of litigation. That's what we are trained for. Lay people don't know these things. They come to building engineers for this kind of information. I think it is a perfectly reasonable request that a good engineer should be able to answer.

Find some kind of credible resource in your area that you can use to make an educated guess about what the water content typically is in your snow and go with that.

Besides, most roof building materials have enough margin in them that even if you are wrong a bit, the first thing they will notice in the roof is excessive deflection before actual fracture failure or collapse.

 

[medeek]

I don't have anything really useful to add to this discussion other than to say it is an interesting topic with a lot of opinions. For the western states I would use the Rocky Mountain Conversion Density (RMCD) for calculating your snow weight/density.

A confused student is a good student.
Nathaniel P. Wilkerson, PE

http://www.medeek.com

 

[JAE]

Triangled - here is my attempt but I believe it wasn't kept too simple:

[blue]Dear Mr. Client:

You have requested information regarding snow loading of the roof of your building located at 123 Main Street, Anytown, USA. As you are aware, our firm designed your roof system for you recently based upon the applicable building code for the city. That code references snow load requirements which, for this area requires a minimum uniform roof snow loading of 36 lbs. per square foot.

Your request was to understand the types of situations where occasional, higher snow loads may occur that would necessitate snow removal to avoid damage or collapse of your roof structure.

In general, the design snow load of 36 psf, based upon the code, relates to an average depth of 1.8 feet (22 inches) of snow.

However, please allow us to point out several concerns with relying on this simple depth:

Density Issues
The density of snow that accumulates on the roof can vary significantly. Multiple snowfalls in several days or weeks can create snow densities much higher, and result in a lower depth of snow with the same design loading. Thus, a 22 inch snow depth may actually be higher or lower than the design load.

Drifting Issues
Snow may accumulate near roof irregularities causing drifting and wild variations in the snow depth. Knowing where and how to measure non-uniform snow depths and how they relate to our design load can be difficult. And excessive drifting could create possible structural concerns in smaller, localized areas. Our design included code required loading for the irregularities included in the design as the building is currently configured.

Water Ponding
Snow can be partially melted in sequential warm days causing significant amounts of water to pond on the roof, especially when snow and ice block or clog drains or scuppers. This can cause serious problems with the accumulation of water on the roof, creating large loads in low or depressed areas of the roof.

Changes in the roof structure
Over time, there may be variations in other types of loads on our roof that could affect how snow loading affects your roof. In the future, additional roof top units, heavier roofing, or equipment hung from the roof inside the building can alter the amount of snow that the roof can support.

So relying on a single snow depth that would trigger action on your part to begin removal operations may not be correct for any given snow or building condition and this might cause you significant unneeded expense if the density is really lighter than assumed, or create an unsafe condition if the density is heavier than assumed.

To at least offer you a general idea or range of values that is possible, consider that snow density can vary between 15 lbs/cu. foot and 30 lbs./cu. foot. This would create a range of depth values for your roof of:
Low depth: 14”
High depth: 29”

These values are based upon our safe maximum load which includes a safety factor. If we remove the safety factor from the depths above we would have the following range of depths that would approach, or equal the theoretical collapse load:
Low depth: 22”
High depth: 46”

From this you can see that due to variations of density the maximum critical depth is not easy to rely on, but hopefully this provides you with at least some understanding of what you have.

Our recommendation would be to not rely on a simple depth measurement. Instead we would recommend that you monitor any situations which would warrant special attention and prompt further investigation. These would include:
1. Significant snowfalls in the higher ranges listed above
2. Snowfalls on sequential period which follow one another prior to the early snowfalls melting away.
3. Rain-on-snow conditions.
4. Significant snow storms noted as historical or above average.
5. Changes in equipment, units, penthouses, screens, etc. that are newly placed on the roof.
6. Changes in equipment or elements newly hung from the roof inside.
7. Changes in roof framing such as sagging, water ponding, or leakage.
8. Any evidence of drains blocked.

If those conditions should occur, we would be happy to assist you in further looking at your roof and how those conditions might affect its structural performance.

Sincerely,
ABC Structural Engineering, Inc.


John Doe, P.E.
Engineer At Large
[/blue]


Check out Eng-Tips Forum's Policies here:
faq731-376

 

[msquared48]

JAE:

Iceing too...

Mike McCann, PE, SE (WA)

 

[LittleInch]

A bit shorter.

Dear Mr Client.

Your roof has been designed to cope with a certain weight of snow [insert weight per sq ft here].

Unfortunately snow and ice comes in a variety of densities - heavy wet snow and ice at worst or light fluffy powder. Therefore the depth of snow the roof can handle could be as low as [insert low value here based on "heavy snow] to [insert high value here].

We would recommend that you clear the roof when the depth becomes greater than [ insert value not much more than the low value here] or install a device like this

http://www.hydrologicalusa.com/product/meteorology/snow-measurement/ssr

to measure the actual load.

REMEMBER: When clearing snow do NOT load snow from one part of the roof onto another part before clearing it as this could overload, take precautions against falling off the roof and try not to use implements which may damage the roof covering.

I'm sorry I can't be more precise, but "snow" is a highly variable thing, a bit like people so e.g. the weight of 10 cheerleaders will be very different from the weight of 10 line backers.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[JAE]

LittleInch - doesn't your phrase: We would recommend that you clear the roof when the depth becomes greater than [ insert value not much more than the low value here] sort of negate everything else? You've essentially said, "blah blah blah blah - your max depth is x - blah blah blah." That is pretty much what the client will ultimate see.

Your link was great - that would actually be a great thing to steer the client to.

Check out Eng-Tips Forum's Policies here:
faq731-376