I have no doubt that earlier this model had 6 subdomains from which the extra sets of elements must have come. I realize that this does not necessarily indicate a bug; stale data that is not referenced would not be a problem but it just seems suspicious.
Please let me know if there is anything additional that I should add.
Re: DC equation coefficients inconsistent or stale?
Yes, if you make a model with several subdomains and later change so that it now has fewer, it might be that the equation and/or boundary coefficients keep the excess entries, but they will not be used or accessed so there should technically be no harm (and if later the model is extended again the "hidden" coefficients will be used as defaults from before).