Is it reasonable to assume a CBM reservoir to be homogeneous and isotropic with reference to porosity and permeability - even during - the later stages of production?
No, even simplest form of analysis show correlation between performance and reservoir properties on all datasets that went through my desk. While some properties might be very relevant to CBM beyond poro / perm, poro / perm remain usually fiarly relevant.
No, reservoir properties vary significantly in time and space. So such generalization are strictly theoretical and are not applicable in practical situations.