Hybrid optical turbulence models using machine-learning and local measurements

Accurate prediction of atmospheric optical turbulence in localized environments is essential for estimating the performance of free-space optical systems. Macro-meteorological models developed to predict turbulent effects in one environment may fail when applied in new environments. However, existing macro-meteorological models are expected to offer some predictive power. Building a new model from locally measured macro-meteorology and scintillometer readings can require significant time and resources, as well as a large number of observations. These challenges motivate the development of a machine-learning informed hybrid model framework. By combining a baseline macro-meteorological model with local observations, hybrid models were trained to improve upon the predictive power of each baseline model. Comparisons between the performance of the hybrid models, selected baseline macro-meteorological models, and machine-learning models trained only on local observations, highlight potential use cases for the hybrid model framework when local data are expensive to collect. Both the hybrid and data-only models were trained using the gradient boosted decision tree architecture with a variable number of in situ meteorological observations. The hybrid and data-only models were found to outperform three baseline macro-meteorological models, even for low numbers of observations, in some cases as little as one day. For the first baseline macro-meteorological model investigated, the hybrid model achieves an estimated 29% reduction in the mean absolute error using only one day-equivalent of observation, growing to 41% after only two days, and 68% after 180 days-equivalent training data. The data-only model generally showed similar, but slightly lower performance, as compared to the hybrid model. Notably, the hybrid model's performance advantage over the data-only model dropped below 2% near the 24 days-equivalent observation mark and trended towards 0% thereafter. The number of days-equivalent training data required by both the hybrid model and the data-only model is potentially indicative of the seasonal variation in the local microclimate and its propagation environment.