When I went to school (Pilot school) a long time ago.............
If as I do one remebers the lift formula, and who doesn't, then one knows that if all the variable are constant except, say TAS (Wind in the static case) and one knows that at lift off L=W then a prety accurate graph can be constructed from that point back to zero.
Since the aerodynamic force depends on the square of the velocity, doubling the velocity will quadruple the lift.
In the light aircraft case if the lift off speed was say 60 knots then:
1. There would have to be a head wind limit on weighing that aircraft of 60 knots. In practice probably more like 40 Knots operationally to allow for x/wind taxying.
2. If the head wind was say 15 knots the lift produced would be 33.3% of that required at lift off.
3. If the weight in still air is known by accurate measurement then a correction factor in this case of -33.3% would produce an accurate model weight.
Prototype trials in still air and in various wind conditions , over a 6-12 month time frame would generate an accurate and reliable weight correction model.
Ipso facto, no need for wind tunnel test data or even a copy of the test pilots notes.