Aero, the table suggests that this is a computer print out opposed to the FCOM (although the manufacturers FCOM was rewritten some time ago). Data prints could be supplied by any service provider.
Also, the 15kt tailwind clearance was not universal for all operators; Australia was one of the first clearances and the distances were almost certainly a cautious extrapolation of existing data.
Tabulated computer data was based on the AFM charts, and had to demonstrate sufficient agreement such that there were no significant differences. This may have involved data points every 20m erring on the safe side of the line. Thus the landing distances were within a cricket pitch length on the safe side of a pencil line over the AFM.
AFM performance was based on a more accurate manufacturers computer model of aircraft performance. This model would be use a set of actual landings, in a limited range of conditions, as agreed by the regulator, averaged across at least five pilots with five landings each, etc, etc, and adjusted with additional time margins and factoring required by regulation. There would be few actual actual approach and landings as the data was an assembly of each phase; i.e. flare and touch down, then separately deployment of landing devices and commencement of braking, and then brakes on to a full stop. Actual approach and landing distances may only have been a spot check, as with wet data, which is normally based on extrapolation - hence the caution required in very wet conditions.
The 146-100 and -300 data were flown over a range different tests and locations, several years apart and possibly biased towards the market for each type, e.g. the -300 at higher weights. Some Australian aircraft had the updated 507-1h engine but I would not expect this to change landing performance.
Thus the differences in the data might only reflect the range of variables in collecting, preparing, and publishing performance, but always erring on the safe side w.r.t. the AFM; but this still requires human judgement of actual conditions and similar bias towards a safe margin.
P.S. Carbon or steel brakes? I cannot recall any differences; but there were for high / low pressure tyres.
IIRC, distances were measured in ft, but published as required in meters.