I am only a lowly Area controller so forgive my ignorance.
Are they really trying to say, in an overly complex and fancy way: We will use time based wake turbulence separations and then convert those into distances for a given wind, in the cases where the time separation (as a distance) would be less than the current published distance requirement?
E.g. H v H (Requirements: 4nm or 2 minutes)
No wind: GS ~160kts : 2mins = 5.33nm so distance based (4nm) is more efficient;
40kt headwind: GS ~120kts : 2mins = 4nm so either or;
60kt headwind: GS ~100kts : 2mins = 3.33nm so time based more efficient.