The data would seem to at least falsify the warfare hypothesis: if they were mostly prisoners-of-war, you'd expect a few localities & demographics to be overrepresented (as they are from the usual enemy city-states, or a burst of young male sacrifices from a victorious battle with many prisoners).
And a gradual dispersion of victims falling off steadily with distance yet with a long tail of very remote origins sounds to me like what you'd get from a mature long-distance network of slave traders.