Kevin, this is something that I seem to encounter very often. The significance of whether it is one tree affects its (their) categorisation as well as RPA(s).
If it's one tree, I'd say it's quality is diminsihed by what is probably maturing epicormics. Furthermore, removal of the subsidiary stems is going to open up the whole tree to infection, and Ash wouldn't stand a chance fighting it off.If that's right, then there is no management recommendation that would restore the tree to the quality of a sinlge stemmed A or B.
If it's more than one tree, the subsidiary stems could be removed, and it could be a sound management recommendation that they be removed. When gradign them they could be seen as of "very low quality ... suppressing adjacent trees of better quality", and as such they would be Cat U.
Regarding RPAs, there is I recall some published support for the view that where two trees are within each others' RPA the rooting areas are shared, without needing to increase the size of either. This is irrational except and unless the trees are diminished in development because of their competition for resources. Even then it's a questionable argument, but it is overpowered anyway by the arbitrary nature of the 12x multiplier.
To apply 5837 to this situatuon, you have to know or at least settle on an assumption about whther it is one or two or more trees. Then one or multiple RPAs can be calculated. But if anyone then draws multiple RPAs as circles, they're kidding themselves on. I mean, how would the roots of the smaller stems get under the large stem?
Also Duncan Slater published an article a coupel of years ago making a good case that trees are positively hostile to each other, especially if the same species.
That's why I investigate and show it's one or multiple failing which state a reasonable assumption of whether it's one or many, and if it's many I skew both RPAS away from each other.
And back to my first point, the real significance is in the Categorisation. One tree with multistemmed cna cause downgradde with no managment solution. Multiple treess can result in the subservient ones being downgraded to U, and a valid management solution being their removal.
For what it's worth, the body language of your situation is saying to me multiple trees.