Methodology for error analysis and simulation of lidar aerosol measurements

Abstract

We present a methodology for objective and automated determination of the uncertainty in aerosol measurements made by lidar. The methodology is based on standard error-propagation procedures, a large data base on atmospheric behavior, and considerable experience in processing lidar data. It yields algebraic expressions for probable error as a function of the atmospheric, background lighting, and lidar parameters. This error includes contributions from (1) lidar signal; (2) molecular density; (3) atmospheric transmission; and (4) lidar calibration. The validity of the algebraic error expressions is tested by performing simulated measurements and analyses, in which random errors of appropriate size are injected at appropriate steps. As an example, the methodology is applied to a new airborne lidar system used for measurements of the stratospheric aerosol. It is shown that for stratospheric measurements below about 25 km, molecular density uncertainties are the dominant source of error for wavelengths shorter than about 1.1 μm during nonvolcanic conditions. Because the influence of molecular scattering (relative to particulate scattering) decreases with increasing wavelength, stratospheric measurements with a Nd:YAG lidar can thus be more accurate than those made with a ruby lidar, provided that a suitable detector is used.