Comparison of Raman LIDAR signal estimation and smoothing methods and correlation between the Pierre Auger side scattering method for determining aerosol content in the troposphere, A

Abstract

Raman backscatter LIDAR is the standard method in atmospheric physics for measuring atmospheric aerosol optical depth profiles. Cosmic ray observatories, including HiRes and Pierre Auger, measure the aerosol optical depth using an elastic side scattering technique. A first ever comparison between the two methods was carried out in southeastern Colorado at the Pierre Auger R\&D site. Between September 2010 and June 2011, over 300 hours of data was collected by the side scattering and Raman LIDAR system in parallel and over 900 hours of data was collected by the LIDAR alone. LIDAR backscattering signals become increasingly dominated by noise as height increases due to an ever decreasing photon return. Smoothing of the signals is required to obtain a usable aerosol optical depth profile. Free-degree density estimation and a customized kernel density estimation smoothing technique were applied to the Raman LIDAR data. It was found that both the free-degree density estimation and the kernel density estimation smoothing techniques work well for LIDAR signals. A strong linear correlation coefficient above 0.9 was calculated between the two techniques. These smoothing techniques were compared with the Savitzky-Golay smoothing technique currently used by a Raman LIDAR group in L'Aquila, Italy. Although the correlations between the density estimation techniques and Savitzky-Golay technique were still strong (above 0.8), there is a systematic difference in the aerosol optical depths observed of around 0.02. Since the two density smoothing techniques smooth the LIDAR signals well, this shift might be explained by differences in the two analyses. A similar systematic offset is seen when comparing the density smoothing methods to the side scattering data.