Moving all species-level observations to complex level since they are inseparable: https://bugguide.net/node/view/15012 The species taxa would be recreated after the merge.
Unintended disagreements occur when a parent (B) is
thinned by swapping a child (E) to another part of the
taxonomic tree, resulting in existing IDs of the parent being interpreted
as disagreements with existing IDs of the swapped child.
Identification
ID 2 of taxon E will be an unintended disagreement with ID 1 of taxon B after the taxon swap
If thinning a parent results in more than 10 unintended disagreements, you
should split the parent after swapping the child to replace existing IDs
of the parent (B) with IDs that don't disagree.
See discussion here: https://www.inaturalist.org/flags/139926