Improvement in the technique to extract gravity wave parameters from lidar data

Abstract

From the analysis of 11 years of INPE lidar profiles and some data from the WIPM lidar, we find that the technique proposed by Gardner and Voelz (1987) for gravity wave parameter extraction does not work well at these sites. At São José dos Campos (23° S) and Wuhan (31°N), Nas occurrence is very high, and the background layer is often far from a single symmetrical Gaussian, with the result that most spatial power spectra (∼80%) will be distorted as the given examples show. The power spectrum is used to extract wave parameters in the Gardner and Voelz (1987) technique, but this only works well for an undistorted spectrum. To solve this problem, a new method is developed. First, a “connection layer” is constructed, allowing the determination of approximate wave parameters and an improved background layer. The background layer parameters and the wave parameters are then used as the initial input parameters for a Levenberg‐Marquardt fit, leading to a much better fitted background layer and better wave parameters. This new method achieves good agreement between simulated sodium variations and observations, as well as artificial numerically simulated “observations,” as shown in the examples presented in this paper.