Thanks for the invite PA, but I don't think I have much to say that isn't already well described and illustrated in Aerodynamics for Naval Aviators (fig. 3.9 about p. 220)
Only thing I might add after re-reading the OP is that the trim change might have been largely due to the reduction in downwash over the tail following loss of wing lift. This would reduce the negative tail lift and give a nose down moment. Shock induced separation on the upper surface, by itself, would just reduce the lift over the rear part of the wing and give a nose up tendency.
This might explain why the modelling of the wing flow was inconclusive, although approaching CL=0 it would be a negative AOA so not too surprising that the lower surface behaved a bit like the upper surface would at positive AOA