The variation in center location as a result of a change in wind across the face of the target isn't all that hard to rough estimate.
Thinking in 2 dimensions:
In a windless environment,a sound wave will propagate in a circle with a radius related to the speed of sound. The time to reach a sensor on the very corner of a 6 ft square target is close to .03771236 seconds from the center (using 1125 as the speed of sound).
If we insert a crosswind across the face, our nice, pretty circle becomes and egg shape. The sound is moving through the air at the same speed as before, but the air mass is also carrying the sound downwind with it.
If we calculate out the elliptical shapes (half upwind, taller than wide and half downwind, wider than tall), we can calculate the effective speed of arrival of the sound wave. In the case of a 5 mph wind, I guesstimate the change in geometric location would be roughly Correction - 0.235 inches at the water line. That number shrinks as the distance between sensors decreases. For a 3 ft square target, it is more like Correction - 0.12 inches.
The problem is that the effects of wind across the target face (even if my guesstimates are completely off) should be the same regardless of distance the bullet travels getting there. It should represent more noticeable effects as the range shortens because the error represents a larger percentage of a ring width.
I know that Adam has the math on this down to a science. Perhaps he will comment.
4/22/19, Corrected a math error.