These various techniques/recommended practices have, as their intent, assessment of three problems -
(a) the steady wind is expected to reduce as the aircraft gets nearer to the ground. The typical presumed profile is the one-seventh power law flight test/certification boundary layer equation
speed(1)/speed(2) = (height(1)/height(2))^(1/7)
or some other empirically derived relationship. Typically, the atmospheric boundary layer is considered to be appropriate below, say, 2000ft and provides a useful means of estimating speeds at heights different to the measuring station
(b) gusts are random and may bite you at any stage during the approach and landing
(c) come over the fence increasingly too fast for the runway and conditions and you progressively increase the risk (probability) of an overrun
What's the answer ? Good question ..
The recommended practices are intended to provide some guidance on how to manage the three considerations. None is precise, none is intrinsically more generally "correct" than another .. but all are usefully helpful in the real world.