I suspect that the extra deep snow, coupled with drifting and the recent heavy rains, loaded the roofs much more than their design snow loads. It would be interesting to know what the actual loads and design loads were in each case. But, I'm sure we will never find out.

One hopes that the code "officials" out there will be updating their ground snow load maps.  But even so, every now and then the statistical probability of any loading map will be exceeded.

Just ask the members of the El Paso, TX surf club.  They are waiting for the western half of the U.S. to drop into the Pacific from the "big one".

Well, as far as the snow maps go... I suppose they're adequate for design.

BUT, in "my" part of the country (I'm in New York City, grew up in Montgomery County, MD) we'd get hit with "a big snow storm" once every couple of years. In Washington, the Washinton's Birthday storm where we got 4 ft of snow in 1978 (I think that's when it hit) is still talked about. In NYC, every couple of years we'd get nailed with big snow in February, sometimes in December, too. "Big snow" means in the 36" within a couple of days. It made the news this time because we've experienced a six-year lull in the big snow so people got all bent out of shape, etc.

DaveViking & PEinc are both right.  I am in the middle of a project where 2 wood beams (3 ply, SPF center, with 1" plywood on each side) have failed, but didn't collapse.  Luckily the restaurant manager had the presence of mind to shut the place down until it was safe.  

The beam was a 24' simple span and the failure location was 6' from the nearest support.  In addition to the recent snow, the beams also had HVAC equipment on the roof.

When I inspected the structure I was shocked to see that each failure was the result of the center SPF ply not being continuous throughout the span.  The connection was only a butt joint, which was only toe nailed.  The 2 outer plys were all that was keeping this beam up.  The thought that any engineer would design something like this, or a building inspector would let this go absolutely horrifies me!

I was wondering the same kinds of things myself when I heard about these failures.  My suspicion is that these collapses were probably due to a combination of extreme loading conditions especially in some localized spots, poor or incomplete structural detailing, maybe a construction error or two, and what I might call 'gnat's butt' designing, by which I mean when the designer, in a justifiable attempt to keep costs to an absolute provable minimum, meets just exactly what the building and design codes require and not one thing more, a practice very typical in our profession.  Of the four, only the last can be planned for or controlled, as the others are unforeseeable.  From engineering school onwards (and though I've been away from there for a longer period of time than I might like to admit, I'd bet it's still basically the same), we as engineers always look to the most efficient solution or lightest structural section when doing a structural design.  While there's nothing wrong with this, and there's no doubt we are doing our jobs when we do this, there is also something to say for what one might call 'judicious overdesign' when it comes to a basic structure.  I'll give you an example.  I work for an institution which has a design manual for new construction.  It calls for using a 30 psf live load for roofs or the prevailing snow load whichever is higher.  I live in a region where the BOCA roof load would generally be around 25 psf or so when you factor in everything.  So is designing a roof for a live load of 30 psf 'overdesign'?  Not in my view.  It allows for a margin of safety (I didn't have to worry too much about any roof collapses here) plus it adds a little safety against minor construction or detailing flaws.  It also affords flexibility, which often proves to be its most beneficial aspect.  We do a lot of renovations to existing buildings here, and when the mechanical guy says he wants to put a new HVAC unit up on the roof, I can often accommodate this without expensive structural renovations which could often make the whole project infeasible.   The first cost for adding a little extra beef in a building's basic structure is pretty cheap when you compare it to many other first cost items. After all, the entire structural cost on a new building is probably around 10-15% of the final cost.  If you added 10% to the structural costs, which is probably more than necessary, you'd still wind up adding 1.5% to the overall building costs at most.  10 years down the road, or even just during construction, that may be the best 1.5% you could spend.  I know most engineers don't get to make these kinds of decisions, unless you happen to be in a facilities job like I have, but it's worth thinking about.  Anyway, just my opinion on the subject.   

We need to remember that minium code requirements are only that: MINIMUM.  As jcoffey suggests, I try to provide just a bit more than code requirements.  I do this more for my own protection than anyone else's.  I like knowing that I went beyond what the code would require.  I just don't tell anyone, especially the owner.  Oops, did I say that out loud?

The combination of heavy snow, followed by rain, making it heavier wet snow, may have exceeded some of the minimums.  While some margin is advisible, maybe you can't design for everything.  Drainage may have been stopped up with leaves or debris (lack of maintenance) and contributed to the loading.

Backsmith

We will forever grapple with the economics of our designs, that is certain.
I've been amazed in my short career how varied the interpretations and opinions can be of highly experienced engineers on particular subjects.  Just look at some of these threads.  It pushes me further along watermelon's line of reasoning, but then again, I've never owned a business and had to make payroll either.    

I think jcoffey nailed it!

(I would also have a little "fluff" in there for re-roofing as well.)

Looking at the overall life of the building should be in the back of our minds. Yes we look at what the current client wants but we usually get dragged into the fray when the lawyers come and play. Regardless of who now owns the building now.

That's my 2 cents...

I also think the code acts as relative maximum as well as an absolute minimum.  Taking jcoffey's example of using 30 psf when 25 is required, I would argue that even if the code was increased to 30, engineers would begin using 33 or 35 just to be safe.  Heck, I'd design for 200 psf and never worry about it.  That is our nature.  If it becomes widespread that engineers are using 30 instead of 25, again only an example, then the code needs to be revisited.  Although we are the ones who end up in the courtroom and it is our conscience that will haunt us with "why didn't I use 30 even though the code required 25" if there is a failure, we must have some faith in the integrity of the code and those who develop it.



  

I sat on a code review committee for my city last year.  We were reviewing the IBC 2000 and noticed that the required snow load was significantly reduced in the new code.

Old timers around my city typically used a roof snow load of 30 psf.  Some years later, the UBC changed to a ground snow of 30 which gave us a roof snow of 21 psf.  We'd add 5 psf for rain on snow and get 26 psf for our design.  Now, the IBC requires 25 psf for ground, making the roof snow 17 psf.  We on the committee said no way and set the minimum roof live load at 25 psf.

Beyond loads, the codes and standards out there are constantly researching the "exact" capacity of any system or material.  So we get more and more complex codes and design methodologies that take lots more time.  Just look at what you really have to do to properly design a concrete column these days.

Most of the failures due to snow that I saw back in 1984 in San Antonio, Texas (13" of snow in Texas!  Do you believe it!) were due to stability issues and not capacity issues.  Many times there were beams running over the tops of columns and no web stiffener plates were provided.  The beams, with the significant snow load, simply rotated off the top of the column and dropped the roof....with the columns still standing vertical.

Maybe we all have to go back to using a slide rule instead of the calculator. Then you can always read the larger number and add your safety factor.  It is strange that you don't see a mandatory safety factor of 2-3 in bldg. design but to design sewer rehab.material thicknesses we have a minumum safety factor of 2.

jcoffey made some excellent points that we should all re-consider.

A few points to ponder....
The code requirements are statistically based minimums.  These are based on PROBABILITIES not POSSIBILITIES.  These recent failures could easily have fallen into the "possibilities" not probabilities.  This simply means that given the appropriate information at the time and designing, constructing, and maintaining the structure in a reasonable or "non-negligent" manner might not have prevented the failures.

There is a need for local review and application as JAE pointed out....it just doesn't usually get done in a formalized fashion as a code supplement.  Our local engineering societies should be driving committees such as the one JAE described.  Remember....one size fits all usually doesn't fit anyone well!

Our Standard of Care does not require perfection.  It requires a lack of negligence.  There's a big difference between those two points on the practice map and let's hope they don't get too close together.  

Details, details, details.

If you use a generally good detailing practice, i.e. web stiffeners on overhanging beams at the column, proper lateral bracing, etc, even if the actual code forces showed that they weren't required, etc., then you can probably tolerate 30 vs 26 psf of snow.  Most catastrophic failures are the ones that were SUDDEN, (Hiatt Regency mezzanine)and not ductile.  

Designing to Code is not enough, you must focus on proper detailing, and CHECK THE SHOP DRAWINGS...

GA

trainguy,

Details, Details, Details..."Hyatt" not "Hiatt".

Sorry, I've been checking shop drawings all day in Kansas City, Mo. I couldn't resist the jab.

Your point is a valid one!
Thanks for the post.

An aside - CT uses BOCA, and supplements BOCA with using a Pf of 30psf minimum.  I cant recall a roof collapse in CT due to snow in recent memory.

Here's one for you all.  As an engineer for the Air Force in Colorado Springs, I have a unique situation to deal with.  The Department of Defense is not constrained by model building codes (except for fire) so we have a slew of regulations of our own that discuss engineering design.  The Air Force Civil Engineering Support Agency, which is the technical experts for Air Force engineering, have always said that the building codes (UBC, BOCA, SBC) are not enough for the federal government and must be supplanted with DoD regulations.  The Corps of Engineers has a regulations (TI 809-1) which lists snow and wind loads for military locations.  My location is listed as 15 psf snow load.  The 97 UBC snow load chart shows the entire State of Colorado to be determined by the code official.  ASCE 7 shows a chart where Colorado Springs falls very close to the border line between 15 psf and CS (case study).  The regional building department here in Colorado Springs requires a Pg of 30 (below 7,000 feet) which may be factored, usually resulting in a Pf of 21 psf.  Most engineers in this town however use the full unfactored 30 psf for design.  So here's my problem.  When we have private engineering firms design buildings for us, they want to use 30 psf.  When the Corps of Engineers designs buidlings for us, they use 15 psf.  Who's right?  The regional building department can't substantiate their 30 psf requirement.  The Corps of Engineers says their 15 psf value is based on statistical analysis.  As a compromise, I have told private firms that I hire to use the 30 psf with factor adjustments such as exposure and sloped roofing.

Hartford Arena, 1978.  Snow loading initiated, but poor design/construction allowed collapse.

vmirat,

I had the fortune of visting the VMI campus last July. Really fine engineering university, I must say.

Anyway, USACE is leaning heavily towards IBC 2000 as "the" code. However, the Uniform Facilities Criteria folks in TI 809-02  "Structural Design Criteria for Buildings" dated 01 September 1999 (http://www.hnd.usace.army.mil/techinfo/ti.htm) references EVERYTHING, including ASCE-7, which it should, as pertinent to structural design. For snow and wind, MIL-HNDBK 1002/2A "Loads" has all sorts of geographical info, though my opinion is that every project is case-by-case and I use these loads as a starting point.

I've heard the "well, if it's really windy, all the snow will blow off the roof, so wind governs" argument. OK, but, gosh, my next-door neighbor's got a lot of FROZEN SNOW stuck to his roof and it hasn't blown off in these gail-force winds. Strange...

I go with whichever code or reg is the most conservative; if the local code says 30 psf and the USACE/NAVFAC/AFM says 15 psf, I would use 30 psf and make a note in the calcs why I chose 30 psf. Federal LAW may trump state and local law, but engineering judgement answers to a higher power, no?

I am somwhat bugged by the recent collapses because it makes structural engineers look bad - even though it *might* not be the fault of the engineer.

Ron - Good point.  I stand somewhat corrected!

I thought this might be of interest.

The collapse of the Borders Book Store in Newark, Delaware, during a severe winter storm in January of 1996.

http://www.sbcmag.info/past/2001/01nov/borders.htm

The caused - The open web bar joists involved in the collapse had been poorly manufactured and the defective welds involved simply failed to sustain the loads transmitted thereon.

I have a buddy who's an insurance agent.  He told me about a claim where a tenant in a building used as a garage cut out the bottom chords of the joists to fit in some equipment.  The owner of the building wanted the repairs paid for.  (My buddy didn't pay the claim.)
My point is that before we are too hasty in blaming the building design, let's wait for the investigation.  While it could be a bad design, there are a several other reasons like bad construction, materials (see post above)and owner misuse which might of caused or contributed to the failure.

Right on Uncle Jed.  All men are innocent until proven guilty.

12 inch deep water is 62.4 lbs/SF.
I think my rafter tables were based on span/something max deflection with max stress in parentheses.  The few buildings I have seen collapse were either damaged from rot or not built to code.

Wrightguy,

You're absolutely wright! My mistake...