I'm fairly certain this is mostly a transient fueling issue. I've seen it before in mild form on dozens of vehicles but none so extreme as the LSA blowers, particularly with aftermarket injectors. I've mentioned it in detail here and elsewhere in the past, but essentially what you're up against is an injector that doesn't spray directly at the intake valve. To my knowledge only the actual LSA/LS9 injector has the proper axial offset to prevent this problem. I did have a lot of success with the ID1050x injectors I ran in the LSA though - gotta hand it to them for the quality. Other injectors were absolutely undrivable. I might try adjusting the transient fuel model to clean up those rich/lean excursions.
That much would explain the movement in the fueling during shifts. But it looks like you have something else going on as well. Can't be too sure with the data so far but it seems like the dynamic air zone transition at 2500rpm is clearing out the VE correction factor from entering steady state airflow (maybe coincidence?). Ignoring the VE correction in steady state is normal operation for the software, but the filter for the VE correction factor is too sensitive to tolerate poor steady state detection from not calibrating the steady state detection algorithm...but most of those variables, the filters and steady state definitions, aren't available to edit. They aren't consistently available in HP Tuners, even among the same controller in different applications.
Anyway, figured I'd share. Definitely look at that transient fueling though.