Measurement of atmospheric neutral wind and temperature from Fabry–Perot interferometer data using piloted deconvolution

Abstract

Nonlinear regression techniques, when applied to sky exposures obtained using a Fabry-Perot interferometer (FPI), are able to recover atmospheric neutral wind and temperature through inversion of the resulting fringe pattern. Current inversion methods often account for temporal fluctuation of the etalon's optical path length (caused by temperature variation in the instrument housing, for example) by characterizing the system function using isolated exposures of a frequency-stabilized laser. Because these path length changes correspond directly to shifts in the fringe pattern, they can significantly increase the total wind velocity uncertainty between laser exposures. We propose an extension to current regression techniques allowing for characterization of the optical path length and measurement of neutral wind and temperature simultaneously, thus reducing the need for frequent isolated laser exposures. This is achieved by using the laser as a pilot signal that enters the aperture of the instrument during sky exposures. We show that the extension can lead to a lower variance estimator for velocity when the optical path length has a significant time-varying component. Additionally, several pragmatic physical configurations that would allow for construction of a piloted signal in a real system are tested and compared using an FPI installation at the Urbana Atmospheric Observatory.