you should probably apply some methods for species delimitation (e.g. GMYC, PTP, SPNA etc.) to see where species limits lies. Then you can make better decision whether the encircled clade can be considered separate species. Applying morphological (or other, reproduction at best) traits is also an option. I would recomend to combine both and apply as much lines of evidence (molecular, morphological, biogeographical etc.) as possible before decision. However, as Lucas mentioned, Alba and A0 definitely can't be considered two separate species, because it would violate the rule of monophyly accepted in current systematics. While A0 might be separate species, Alba would be paraphyletic because A0 evolved within it and can't be excluded (according to phylogram you presented).
I'm a bit suspicious to that tree, to me it looks odd to have A0 being separated from Aalba with such a relatively long branch and still not have Aalba monophyletic. What methods is the tree estimated with and is it correctly rooted? I would guess you may get a different result if you reestimated the phylogeny with a different method and if needed additional outgroups. But as already said Aalba is paraphyletic and A0 is monophyletic in your tree
It would have been nice if you would have provided details such as type of data and the model used to derive of the tree. As can be seen, the branches are labelled with some numbers, I guess, the bootstrap values. Please clarify number of times data was simulated. In any case, refrain from interpretation of tree topologies especially when bootstrap support is low. PFA a book chapter on molecular phylogeny which we had written a few years ago. Hope this helps.
As Svante Martinsson mentioned; you should re analyse your data with different method like Bayesian Inference. I would recommend to re look at your sequence alignment and delete any noisy flanks. WhaT I would like to say is to refine the alignment and carry on toward another analysis. In most cases the long branch results from bad alignment.