How rich the environment is in $\ce{Sn}$ vs $\ce{S}$ affects the chemical potential $\mu_\ce{Sn}$ as described in Section IIIA of the linked paper:
... $\mu_\ce{Sn}$ for the Sn-rich environment is equal to $E_\ce{Sn}$, where $E_\ce{Sn}$ is the total energy of Sn per atom in the diamond crystal structure.
... $\mu_\ce{Sn}$ can be calculated for the S-rich
environment as $\mu_\ce{Sn}= E_\ce{SnS}-E_\ce{S}$.
(Note $E_\ce{SnS}$ and $E_\ce{S}$ in the quote above are the energy per formula unit and atom respectively).
As your image shows, the formation energy depends on $\mu_\ce{Sn}$, so it also depends on how rich the environment is in $\ce{Sn}$ vs $\ce{S}$.
Richness here is just referring to how abundant a given element is in the surroundings. This won't change the simulations of the monolayer, but it does affect the formation energy. This is because in an Sn-rich environment, you are not just producing a lone Sn atom when you remove it from SnS. Rather, you are implicitly incorporating it into bulk metallic Sn, which has different energetics. For an S-rich environment, they assume the Sn is instead producing a site of SnS at the expense of a site from bulk S. So depending on the environment, the overall formation process is different.
If you aren't concerned about the environment effects, you could use the same definition for the chemical potential as was used for the doping transition metal (e.g. the energy of one isolated atom). They actually do this earlier in the linked paper, where they compute vacancy formation energies for SnS either using the environment affected chemical potentials or just the energy of a lone atom.