Yes and no. The empirical study could identify correlations and the strength of those correlations, but does nothing to say why they are correlated (maybe just spurious). But a CFD analysis may give you insight into what may be happening to cause the issue, which can then lead to a hypothesis that can then be tested further. All models are wrong, but some are useful.
Only if the depth is the only variable that’s changed. You can’t just use two entirely different pools at different times and locations and conclude that the depth is the reason for different performance.