@ 2736 cM they are full siblings, 100%.
To help, this is the DNA share results with my half Sister, 26% shared DNA: 1,817 cM across 33 segments
I have seen similar sibling and half sibling results when using Pro Tools and looking at Shared Matches. The lack of trees and apparent activity suggest that they may not even be aware of each other, especially if all they were interested in is their ethnicity.
Conversely if one of them is aware of Shared DNA amounts that then may have answered their search for a missing sibling and that could result in them not including them in their online info.
DNA shares in the 200-400 range do have a lot of relationship possibilities and can be difficult to map into a feasible tree and the clue here is the unweighted figures which do indicate closer MRCA. If you have an idea of any generational distances between these matches then that too may help narrow down the possibilities.
Hopefully the 2736 cM siblings have enough information to incorporate them into a tree together which in turn could show family breakdown or indeed potentially a baby being passed on for adoption and if this is the case a paper trail will show totally different families. In these situations we trust the DNA rather than records.