Estimating the backscatter spectral dependence and relative concentration for multiple aerosol materials from lidar data

摘要

Detection and estimation of materials in the atmosphere by lidar has heretofore required that the spectral dependence of the relevant cross section coefficients - backscatter in the case of aerosols and absorptivity for vapors - be known in advance. While this typically is a reasonable assumption in the case of vapor, the aerosol backscatter coefficients are complicated functions of particle size, shape, and refractive index, and are therefore usually not well characterized a priori. Using incorrect parameters will give biased concentration estimates and impair discrimination ability. This paper describes an approach for estimating both the spectral dependence of the aerosol backscatter and relative concentration range-dependence of a set of materials using multi-wavelength lidar. The approach is based on state-space filtering that applies a Kalman filter in range for concentration, and updates the backscatter spectral estimates through a sequential least-squares algorithm at each time step. The method is illustrated on aerosol-release data of the bio-simulant ovalbumin collected by ECBC during field tests in 2002, as well as synthetic data sets.