windward snow drift on sloped surface

huangyhg

windward snow drift on sloped surface
i am designing a new parapet on an existing structure.  we are building the parapet up on an existing wall and providing a sloped roof behind it to support the top of the parapet since we can't develop fixity at the base.  the parapet is 4 feet high and i am sloping the roof behind at 1 to 1.  
since the surface is sloped and almost flush to the top of the parapet (4" reveal for flashing), do i still need to check for windward drift up again a roof projection?  should i apply the drift the sloped framing or start from the toe of the slop towards the interior?
i remember an "old timer" tell me to slope the framing at 4 to 1 to match the anticipated drift and then ignore it.  he said that is what they did when they retrofit structures when the concept of snow drift was fist developed, but i can't find that as an alternate in any codes.  now the slope of the drift is modified if the height of the drift is taller than the projection anyhow, so i am even less convinced about that method.

i worked on an existing strip mall once where a tall facade was added.  i did exactly what you are describing (sloped framing) to prevent snow from drifting.  i think if the slope of the framing is close to matching the anticipated slope of the drift, you can ignore drifting.
daveatkins
daveatkins, did you go up on the roof after a windy snow storm and see if that was true?
my intuition agrees with you that it eliminates the "corner pocket" of the drift, but i also wonder if some of that snow would still pile up on the diagonal - sort of a valley drift.  maybe not as much as the original parapet drift, but higher than nothing at all.
what do you think?
if you slope the roof as described, you may want to consider the effects of snow sliding from the new roof and piling up at the existing roof.   
i agree with stenger and jae.
last year i was working as a laborer on a new building in maine. after we had just finished a lower flat roof with decking, we had a large blizzard. there was substantial piling of snow on the lower roof. the snow had slid from the sloped roof and blown backwards onto the lower roof instead of falling straight to the ground. there was so much pile up that our foreman contacted the structural engineer out of concern that the lower roof was overloaded. turned out that we had to add another layer of decking and double the number of bar joists supporting the lower roof.

bones206,
what you are describing is a sloped roof some distance above a flat roof, and yes, the code covers drifting and sliding snow.
what we are discussing is a roof that slopes, then flattens out.  i think the argument can be made that this is a "valley" condition, but i choose to ignore snow buildup in this type of valley.  for example, when i design an apartment building with multiple ridges and valleys, i design for flat roof snow load across the entire roof.
daveatkins
i see the argument for the 'valley' theory.  but a 'valley' is from a change in the horizontal plane that could potentially block blowing snow.  i too ignore this condition.  the sloped roof example used to eliminate drifting creates a change in the vertical plane.  if the slope is too great, vertical, then drifting occurs.  a gradual slope to the top of parapet would not block blowing snow and not drift.  somewhere in the middle, i would argue that a sliding snow condition may occur, dependant on roof slope and roofing material.  i'm not sure the drift length or angle makes a difference but could be a good rule of thumb.   
dave,
yes, my anecdote was a different situation from the one being discussed. i just wanted to echo the points made by jae and stenger, that there may be unanticipated drifting or sliding. i don't necessarily think it is conservative to ignore these effects.