Method to retrieve single scattering properties of aerosols using multi-wavelength scattering and absorption coefficient data measured by integrating nephelometer and absorption photometer

Abstract

Scattering and absorption coefficients, measured with an integrating nephelometer and absorption photometer, respectively, are often used to characterize aerosols. This study developed a method for retrieving the single scattering properties of an aerosol from multiwavelength scattering and absorption data and evaluated its performance and accuracy using simulation data based on Optical Properties of Aerosols and Clouds (OPAC) models. This statistical retrieval method retrieves the volume-size distribution and the complex refractive index simultaneously, which reconstruct the measured values taking into account the angular truncation and non-ideality of the light source. These values are then used to more accurately estimate single scattering properties (scattering, absorption, and extinction coefficients: single scattering albedo (SSA); and the asymmetry factor). With no systematic (bias) error, the root mean square error (RMSE) of the measured scattering coefficient was 4.42 × 10–5 to 6.61 × 10–5 m–1, whereas the RMSE of the retrieved scattering coefficient was 9.48 × 10–6 to 1.09 × 10–5 m–1. The RMSE was thus reduced to about 20 %. The RMSE of the SSA calculated directly from the measured values was 0.014–0.021, and that of the SSA calculated from retrieved values was 0.002, corresponding to a relative error of 0.2 %. In each case, the error in the SSA calculated from the retrieved scattering and extinction coefficients was less than the SSA calculated from the measured values. Sensitivity tests of the systematic (bias) error of absorption and scattering coefficients in SSA retrieval demonstrated that, with a 10 % systematic error, the maximum difference between the true value and the retrieved SSA exceeded 0.02 for a small SSA, but with a systematic error of 3 % or 5 %, the maximum difference was small. Therefore, a systematic error of less than 5 % is desirable. The retrieved volume size distribution and complex refractive index were qualitatively similar to the original values.