I generally only look at longitudinal and transverse conditions on the roof.  Oblique directions have more effect on the primary structure than they do on the roof surface.

Ron, are you saying when you run oblique wind cases you put no load on the roof?

dozer:

I think what Ron says is that Oblique directions will not be the controling factor for roof design, but longitudinal and transverse conditions will control the roof design. that does not means you don't need to put load on roof when you run oblique wind cases.

dozer...Chris is correct.  Can't neglect the roof loads and their affect on the structure, but for the roof system and structure, the longitudinal and transverse will generally define the structure of the roof system.

Yeah, I get that. My question is how do you apply wind loads to the roof in an oblique case? I could see somebody having a load case in a structural analysis program for wind on the roof parallel to the ridge and another for wind perpendicular to the ridge and without giving it much thought multiplying each of these by 0.75 and adding them together.  That wouldn't be right.

I believe you when you say it doesn't control the roof design but you still have to put something on the roof. I'm just surprised that ASCE 7 doesn't say anything about it.  

ASCE 7 addresses the general setup conditions, but doesn't dictate how you get there.  Many engineers still do hand calcs, so the technique for doing that type of analysis is going to be different than if you are doing a full FEA.

As an example, it is sometimes easier to do a few hand calcs and develop the roof structure reactions, then apply those to an "open box" FEA model.  That saves the time and effort to build each truss/joist/dummy beam for the roof into the model.  It makes the interpretation of the results easier and also makes it easier to troubleshoot your model if you are working with a notoriously "buggy" FEA program.  Depending on the complexity of your model, they can "blow up" on you pretty easily sometimes.

So, IMHO, ASCE 7 sets general parameters from which we are to use engineering judgment in the application of those parameters.  They cannot set parameters with such site specific parameters as building geometry, building design parameters and other items that would be specific to your building, but they give you the tools to set those yourself.

In your case, you can do it either way you'd like.  You might even want to try it a couple of ways so that you'll know the differences and can look for special cases.  If you have built a full FEA model, it's pretty easy to change the load conditions.  If you have a partial model, same thing.  If you're doing it by hand...simplify!

The link below is from the India Code. Check out page 35 of 106 for the negative roof pressure values. They are very high in the corners at all angles.

http://www.iitk.ac.in/nicee/IITK-GSDMA/W05.pdf

I wouldn't think the quarterly wind direction governs roof truss design. The greatest wind load occurs immediatley adjacent to a windward edge. Considering the quartely wind loads will only reduce this action.

kikflip...I agree.  That's why I use the longitudinal and transverse conditions for the roof, then transfer the reactions to "the box".

Dozer,
I only have ASCE 7-05 but it sounds similar to the newer 7-10. It is my understanding that you take the loadings shown in Fig 6-10 (7-05) which include both wall and roof loads and apply them as shown in Case 1 - 4. So in case #3 you would apply 75% of the loadings from a transverse direction and 75% of the longitudinal direction.