parapet height leeward step

huangyhg

parapet height & leeward step
i have a one story building abutting a two story building.  the upwind distance on the two story building is 250'.  my surcharge on the low roof from the leeward drift is quite high (about 78 psf).  however, my parapet height is 7' above the high roof elevation and i feel like there is some logic to reducing the lower roof drift surcharge, i just don't see the justification in asce7-98.  any thoughts?
i don't know the parameters of your specific roof, but carefully read 7.7.1, 7.7.2 and 7.8 (asce 7-98)
thanks. i am thoroughly familiar with those sections, and have been through the commentary many times over, as well.  this is a point that does not seem to be covered.
                    !
                    ! parapet =7'
    lu-250'        !
--------------------
                   !
                   ! roof elev difference =20'
                   !
                   !----------------
my windward drift on the upper parapet is about 3.2 feet (based on 250' upwind length). base snow is 1.05', for a total of 4.25'.  my parapet is 2.75' higher than my snow drift and, in my mind acts as at least a partial impediment to blowing onto the low roof.  commentary section 7.7 implies that the upper roof snow is "blown clear" as shown in fig 7-8 (2nd to last paragraph in section).  i wonder if that is really the case for a really tall parapet.  how tall, i do not know.  hence the post for opinions.  in the end, i put all the leeward drift on the low roof, as if the parapet was not there. better safe than sorry.
i do not believe that you can use a parapet to "block" snow from drifting onto a lower roof.  the actual dynamic behavior of the snow is not linear (i.e. the snow does not translate perfectly horizontally but rather is blown upward and deposited below due to the wind lines around the building shape.
i have wondered this before too
until i added forensic engineering/failure analysis to my workload i also questioned what seemed to be overly conservative code requirements.  i now feel that the snow drift articles of asce/boca are far from conservative and actually may need to be made more demanding, particularly the density formula and the length of the drift.  
i have investigated the failure of far too many buildings and roofs, some of which were designed in accordance with these codes and still failed under what was nothing more than a "once a decade" snowfall.
the vast majority of my investigations unfortunately revealed that the roof failed as a result of engineers either ignoring or misunderstanding snow drifting at changes in rooflines, parapets, etc.  
a code article is a minimum requirement and that is what the state board will hold you to upon their investigation into any roof collapse.  if the situation is as black and white as you describe it, you better do it by the book or pray for warm winters.
the lower roof should be designed for full snow accumulation.  the effect of the parapet will reduce the lower roof load a bit, but it's not likely that this can be determined.
the wind velocity drop will happen irregardless of the parapet and snow will still be 'precipitated' as a result of the reduction in wind velocity.  in addition, there will be small 'vortex' that will form due to the 'suction' developed...
the code provides that, for drifting snow wind can be from any direction. wind from one direction (behind the parapet) may reduce drift in that direction due to aerodydamic shading, however, from the other direction the parapet acts as a collector. the worst case must be considered.