Eliminating Photon Noise Biases in the Computation of Second-Order Statistics of Lidar Temperature, Wind, and Species Measurements

The precision of lidar measurements is limited by noise associated with the optical detection process. Photon noise also introduces biases in the second-order statistics of the data, such as the variances and fluxes of the measured temperature, wind, and species variations, and establishes noise floors in the computed fluctuation spectra. When the signal-to-noise ratio is low, these biases and noise floors can completely obscure the atmospheric processes being observed. We describe a novel data processing technique for eliminating the biases and noise floors. The technique involves acquiring two statistically independent datasets, covering the same altitude range and time period, from which the various second-order statistics are computed. The efficacy of the technique is demonstrated using Na Doppler lidar observations of temperature in the upper mesosphere and lower thermosphere acquired recently at McMurdo Station, Antarctica. The results show that this new technique enables observations of key atmospheric parameters in regions where the signal-to-noise ratio is far too low to apply conventional processing approaches.