Where you do not have snow blowing off the upper roof onto the lower roof, as in your case, you can look at the adjacent or common wall as being similar to the parapet condition above it.  In this case 7-93 allows you to consider 0.5hd as the drift load height  (para 7-8).

Ron, at what point would you guess that snow might begin to blow up and over the top of the parapet and add to the lower roof?(ex. What if there were only 1 foot between the highest point of upper roof snow drift and top of parapet.) This could cause a much greater lower roof accumulation if the upper roof length is much larger compared to the lower roof length.

I would have to agree with haynewp.  We ALWAYS use full drift on the lower roof, regardless of whether there is a parapet above.  

The "aerodynampic shade" developed from a drop in roof elevation still occurs even with a parapet, as the parapet typically doesn't significantly affect the overall effect of the shielding caused by the high roof.  Yes, some small drifting occurs on the parapet side, but the amount of snow "caught" by the parapet normally is small compared to the total amount of snow initially deposited on the high roof, and then dropped onto the low.  ASCE 7 doesn't really address the leeward step drift with a parapet so we play it safe and go with full drifting.  

The other reason we do not take advantage of the parapet above is due to the possibility of windward drifting.  This is where low roof drifts occur against a high roof wall due to snow being blown into it.  Ron's suggestion to use 1/2 x hd is based on this kind of drifting.  There is a good commentary on this in ASCE 7-98 (C7.7) which illustrates this.  Some recent research has forced a revision in ASCE 7-98 (from 7-93) to require raising the 1/2 x hd to 3/4 x hd for this condition.  We just simplify and go with full, typical drifting.

Finally (but less likely), someday, somehow, the owner may wish to remove portions of the parapet.  

Two reasons for my answer as stated....

(1) The reference was for ASCE 7-93, which is different than 7-98 in the way it handles the windward drift

(2) The parapet height was given as 8 feet, which will trap windward snow for the top roof, but presents a leeward condition on the lower roof.  For wind from the opposite direction, little accumulation would occur leeward of the parapet, as is borne out by Haynewp's statement of much less than 8 feet of drift expected behind the parapet.

I agree that we should be conservative with snow drift, as it does not always follow prediction.  The "problem" as stated can be interpreted within the provisions of the Standard.

Haynewp...with regard to your second question, you have changed the parameters a bit.  If the parapet drift gets to the point that wind is blowing the snow directly off the upper roof onto the lower roof, then the issues JAE raised are certainly applicable.

Oops...I guess I didn't see the 8 foot height.  At that height, there would certainly be a significant amount of snow trapped by the parapet for the leeward step condition.  But even so...I would still, at least, use the 3/4 hd indicated for the windward condition.