Quasi-stable thermal gradients driving atmospheric optical anisotropy

We propose a paradigm shift in the study of optical anisotropy, moving from ad hoc empirical models to a physics-based framework with predictive capability. Long-standing observations, previously described by introducing fitting parameters not linked to atmospheric characteristics into otherwise isotropic models, are explained here by the combined effect of isotropic turbulence and atmospheric temperature gradients. Since temperature gradients can be measured or obtained from fluid-dynamic simulations, this approach enables quantitative predictions of anisotropy strength and orientation. Based on previously published near-surface profiles, we demonstrate for the first time with numerical simulations (to our knowledge) that commonly observed gradients can strongly modify long-exposure beam properties, including beam shape, scintillation index, and covariance ellipse rotation. Weak gradients above grass surfaces can alter the beam diameter by up to a factor of two over ~2 km, while stronger gradients near asphalt yield comparable effects at hundreds of meters. Transitions between shadowed and sunlit profiles observed at the same location produce covariance ellipse rotation, and evolving gradients explain the observed switch between vertical focusing and defocusing on hour-long timescales.