I've updated the tri-modal plot highlighting in yellow the objects with parents (i.e. deblended objects). They only lie one one of the three branches. I'm not sure yet what the driver is between the other two branches (will likely need to compare the images of several cases to get clues).

Fred Moolekamp I have created DM-9537 to investigate the cause of the third branch. Did you want to create a few other issues in regards to using blended objects as PSF model stars and the behavior of the footprint growing before we can mark this as done?

Created DM-9546 and DM-9547 to investigate the use of blended sources to model the PSF and the behavior of the footprint growth algorithm respectively.

The cause of the bimodal distribution has been identified by Lauren MacArthur as a difference in the footprint sizes after they are allowed to grow between blended and isolated sources. DM-9537 has been created to investigate a third distribution that appears in some fields, DM-9546 has been created to investigate why blended sources are being selected to model the PSF, and DM-9547 has been created to understand and document why growing the footprints is working differently for isolated and blended sources.

Great, thanks Fred Moolekamp!