Earth and Atmospheric Sciences, Department of


Date of this Version



Xu, X., and J. Wang (2015), Retrieval of aerosol microphysical properties from AERONET photopolarimetric measurements: 1. Information content analysis, J. Geophys. Res. Atmos., 120, 7059–7078, doi:10.1002/2015JD023108.


Copyright © 2015 American Geophysical Union. All Rights Reserved. Used by permission.


This paper is the first part of a two-part study that aims to retrieve aerosol particle size distribution (PSD) and refractive index from the multispectral and multiangular polarimetric measurements taken by the new-generation Sun photometer as part of the Aerosol Robotic Network (AERONET). It provides theoretical analysis and guidance to the companion study in which we have developed an inversion algorithm for retrieving 22 aerosol microphysical parameters associated with a bimodal PSD function from real AERONET measurements. Our theoretical analysis starts with generating the synthetic measurements at four spectral bands (440, 675, 870, and 1020 nm) with a Unified Linearized Vector Radiative Transfer Model for various types of spherical aerosol particles. Subsequently, the quantitative information content for retrieving aerosol parameters is investigated in four observation scenarios, i.e., I1, I2, P1, and P2.Measurements in the scenario (I1) comprise the solar direct radiances and almucantar radiances that are used in the current AERONET operational inversion algorithm. The other three scenarios include different additional measurements: (I2) the solar principal plane radiances, (P1) the solar principal plane radiances and polarization, and (P2) the solar almucantar polarization. Results indicate that adding polarization measurements can increase the degree of freedom for signal by 2–5 in the scenario P1, while not as much of an increase is found in the scenarios I2 and P2. Correspondingly, smallest retrieval errors are found in the scenario P1: 2.3% (2.9%) for the fine-mode (coarse-mode) aerosol volume concentration, 1.3% (3.5%) for the effective radius, 7.2% (12%) for the effective variance, 0.005 (0.035) for the real-part refractive index, and 0.019 (0.068) for the single-scattering albedo. These errors represent a reduction from their counterparts in scenario I1 of 79% (57%), 76% (49%), 69% (52%), 66% (46%), and 49% (20%), respectively. We further investigated those retrieval errors over a variety of aerosol loading and fine-/coarse-mode prevalence, which indicates that observations in scenario P1 can yield the retrieval of refractive index and single-scattering albedo for both fine and coarse aerosol modes, when aerosol optical depth at 440 nm is larger than 0.2 and 870/1020 nm Ångström exponent ranges between 0.7 and 1.6.