Sorry - I'd have to sit down with pencil and paper to really think it through, and don't have the inclination right now ;-)

What I CAN offer is that there has been some recent work (I believe if not in a paper already, soon to be in one from Fekete) that basically says a decline exponent of 2.0 (where 'normal' decline parameters should not go over 1.0) is a close approximation to transient flow. In fact they are using this to help characterize tight gas declines - from assume drainage radius and perm/poro they calculate time to reach boundary against which they use exponent of 2.0 until boundary 'seen' then something less than 1 (ie 0.6) after this

Hope this helps.