Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

Re: ground snow v. roof snow - which is it?

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Title: Re: ground snow v. roof snow - which is it?

Here's a little more background on the snow loads issue:

1.  Forensic investigations of snow load failures in the early 90's indicated that drifted snow loads were underpredicted by the unbalanced snow loads in early versions of ASCE 7 (pre-1998).

2.  As a result, the unbalanced snow loads were increased, first in ASCE 7-98 and then in ASCE 7-02, to account for the drifted snow.  HOWEVER, the model for unbalanced snow loads was the same as used in early version of ASCE 7... place a uniform surcharge + balanced snow load on the leeward side of the roof, rather than to try to place the actual triangular drift load on the roof.

3.  RPI studied the issue in more detail.  They realized that the use of the uniform surcharge model significantly OVERPREDICTS the total unbalanced snow load on the roof (controls the design of rafters, beams, headers) and UNDERPREDICTS the localized loads that can occur directly under the drift (controls sheathing, purlin, bracing, and localized bending in truss chords).

4.  The ASCE 7 Snow Task Committee debated how to address the problem.  Some wanted to define the drift load using a triangular load that mimiced the drift shape.  Some wanted to keep a uniform surcharge.  They finally developed a partial uniform load that is located on the leeward side of the ridge.  It induces the same moment in the rafters/beams as the triangular drift.  It also induces much higher localized loads near the ridge, though still not as high as a triangular load over a very small area... all in all, a reasonable compromise given some of the other discussions.

5.  For small roof dimensions, the partial uniform load nearly covers the whole roof.  It was demonstrated that the effective moment, induced on rafters/beams with the new loading scheme, approximates a uniform load on the leeward side equal to the ground snow load.  The ASCE 7 Snow Task Committee agreed and created a special loading provision for roofs with ridge-eave distances <=20' that allows the unbalanced snow loads to be approximated by applying the ground snow load on the leeward side.  This provision reverts the snow loads for small buildings back to the loads historically used for decades in the rafter tables.

6.  Due to concerns about bracing and localized bending in truss chords (and suspected failures as a result), the simplification in item 5 was not carried over to trusses; however, the total load on trusses should go down as a result of the new loadings.  It does, however, require some patience and proficiency in applying the provisions!

Please note the distinction in these comments between small buildings and larger commercial structures.


John "Buddy" Showalter, P.E.
Director, Technical Media
AF&PA/American Wood Council
1111 19th Street, NW, Suite 800
Washington, DC 20036
P: 202-463-2769
F: 202-463-2791

The American Wood Council (AWC) is the wood products division of the American Forest & Paper Association (AF&PA). AWC develops internationally recognized standards for wood design and construction. Its efforts with building codes and standards, engineering and research, and technology transfer ensure proper application for engineered and traditional wood products.

The guidance provided herein is not a formal interpretation of any AF&PA standard.  Interpretations of AF&PA standards are only available through a formal process outlined in AF&PA's standards development procedures.


From: Jim Wilson <wilsonengineers(--nospam--at)>
Subject: Re: ground snow v. roof snow - which is it?
To: seaint(--nospam--at)
It seems as if the code writers have eliminated the
option of reducing snow load when using prescriptive
design. Subsequently, it is up to the local enforcers
to decide if an engineered design is permitted to
include the reduction factors. Does that sound
--- Drew Morris <dmorris(--nospam--at)> wrote:
> Jim Wilson wrote:
> >I'm under the impression that typical roof span
> charts
> >for wood rafters, beams, etc. are based on ground
> snow
> >load. At least that's what the tables say. That
> >seems misleading since using the ASCE formula, one
> >could reduce that by 0.7 or more for roof loading.
> >I'm guessing that the prescriptive code takes the
> >conservative route by making no adjustment for roof
> >snow factors.
> >
> >Similarly, if "engineering" a wood roof, should the
> >same adjustment factors be ignored to maintain a conservative
> >consistency? Maybe there is a
> commentary
> >on this some where.
> >
> >Just curious...
> >
> >Jim Wilson, PE
> >Stroudsburg, PA
> >
> >
> >
> In Alaska, Anchorage and Juneau specify a ground
> snow load and a
> minmimum roof snow load. The roof snow load is
> calculated using ASCE 7,
> Section 7.
> In Anchorage, the ground snow load (Pg) is 50 psf
> and the minimum design
> roof snow load (Pf) is 40 psf.