A note on splitting errors for advection-reaction equations