Content area
Abstract
Atmospheric aerosols play an important role in earth climate by scattering and absorbing solar and terrestrial radiation, and indirectly through altering the cloud formation, life- time, and radiative properties. However, accurate quantification of these effects is in no small part hindered by our limited knowledge about the particle size distribution (PSD) and refractive index, the aerosol microphysical properties essentially pertain to aerosol optical and cloud-forming properties. The research goal of this thesis is to obtain the aerosol microphysical properties of both fine and coarse modes from the polarimetric solar radiation measured by the SunPhotometer of the Aerosol Robotic Network (AERONET). We achieve so by (1) developing an inversion algorithm that integrates rigorous radiative transfer model with a statistical optimization approach, (2) conducting a sensitivity study and error budgeting exercise to examine the potential value of adding polarization to the current radiance-only inversion, and (3) performing retrievals using available AERONET polarimetric measurements.
The results from theoretical information and error analysis indicate a remarkable increase in information by adding additional polarization into the inversion: an overall increase of 2–5 of degree of freedom for signal comparing with radiance-only measurements. Correspond- ingly, retrieval uncertainty can be reduced by 79% (57%), 76% (49%), 69% (52%), 66% (46%), and 49% (20%) for the fine-mode (coarse-mode) aerosol volume concentration, the effective radius, the effective variance, the real part of refractive index, and single scattering albedo (SSA), respectively, resulting in their retrieval errors of 2.3% (2.9%), 1.3% (3.5%), 7.2% (12%), 0.005 (0.035), and 0.019 (0.068).
In real cases, we demonstrate that our retrievals are overall consistent with current AERONET operational inversions, but can offer mode-resolved refractive index and SSA with sufficient accuracy for the aerosol composed by spherical particles. Along with the polarimetric retrieval, we also performed radiance-only retrieval to reveal the improvements by adding polarization in the inversion. The comparison analysis indicates that with polar- ization, retrieval error can be reduced by over 50% in PSD parameters, by 10–30% in the refractive index, and by 10–40% in SSA, which is consistent with the theoretical results.
