scott, there may be a way to improve your "complex impedance" modelling.
inductance is kind of like the inertia of a moving mass but in the electrical domain--you push the electrons, but they push back and don't want to move. after reviewing the location of the impedance peak in a winisd model for the current fi q in 4 cubic feet sealed (~50hz) and what ricci actually measured (~30hz), i figured that i'd experiment to see how much added physical mass would be required to shift the location. it turns out that ~2x mms brings the impendance down to the measured result.
that is a pretty significant amount of "electron weight".
adding that kind of mass to a driver significantly lowers the efficiency (-~6db) in the upper bass region, while increasing it in the bottom end (+~3db).
yellow is sub with 2x mms added first, then 1khz inductance increased to match actual impedance measurement at 45ohms, which is 7mh. red is just increasing the inductance. across the subwoofer region, the added weight method ends up matching the measured result quite well.
maybe there is a way to make this into a predictive model, by first modelling the sub in the modelling program and adding weight until the location of the impedance peak is properly located, then adding 1khz inductance until the impedance at that point is about the same.
this method however requires an impedance plot, which is something that not all companies provide.