Estimating the atmospheric turbulence strength (Cn2) is crucial for optical wave propagation. Site-specific prediction is difficult because classic empirical models are built from measurements across disparate climates. We train a quadratic Fourier neural network (QFNN) on winter-season field data and macro-meteorological variables from a rural site; the model delivers reliable estimates with R2 ≈ 0.92 for measured Cn2, outperforming well-known baselines.
Çelik, U., & Yaşar, H. A. (2024). Estimation of Ground-Based Atmospheric Turbulence Strength (Cn2) by Novel Neural Network Architecture. Applied Optics. https://doi.org/10.1364/AO.532723
@article{CelikYasar2024AO,
title = {Estimation of Ground-Based Atmospheric Turbulence Strength (Cn2) by Novel Neural Network Architecture},
author = {Çelik, Uğurcan and Yaşar, Hüseyin Avni},
journal = {Applied Optics},
year = {2024},
doi = {10.1364/AO.532723}
}