The *_FS tables are the "flow split" tables (for drainage line and catchment feature classes). They contain the HydroID of a downstream feature if there is more than one downstream feature at a given feature (HydroID of the first downstream feature will be stored in the field NextDownID of the feature itself, while the _FS table will have the additional values).
In the _FS table, "FEATUREID" field contains the HydroID of the feature that has a flow split, while "NEXTDOWNID" field contains the HydroID of the downstream feature.
The only time there will be an entry in those tables is if you have braided streams in your system (there are flow splits). If you are using dendritic terrains and you do not provide streams with braids as inputs into the process (as in your case where streams are unknown), these tables will always be empty.