Impact of particle size, refractive index, and shape on the determination of the particle scattering coefficient – an optical closure study evaluating different nephelometer angular truncation and illumination corrections

Abstract

Abstract. Aerosol particles in the atmosphere interact with solar radiation through scattering and absorption. Accurate aerosol optical properties are needed to reduce the uncertainties of climate predictions. The aerosol optical properties can be obtained via optical modeling based on the measured particle size distribution. This approach requires knowledge or assumptions on the particle refractive index and shape. Meanwhile, integrating nephelometry provides information on the aerosol scattering properties directly. However, their measurements are affected by angular non-idealities, and their data need to be corrected for angular truncation and illumination to provide the particle scattering coefficient. We performed an extensive closure study, including a laboratory and a simulated experiment, aiming to compare different nephelometer angular truncation and illumination corrections (further referred to as “angular corrections”). We focused on coarse-mode irregularly shaped aerosols, such as mineral dust, a worldwide abundant aerosol component. The angular correction of irregular particles is found to be only ∼2 % higher than the angular correction of volume-equivalent spheres. If the angular correction is calculated with Mie theory, the particle size distribution is needed. Our calculations show that if the particle size distribution is retrieved from optical particle spectrometer measurements and the irregular shape effect is not considered, the angular correction can be overestimated by about 5 % and up to 22 %. For mineral dust, the traditional angular correction based on the wavelength dependency of the scattering coefficient seems more accurate. We propose a guideline to establish the most appropriate angular correction depending on the aerosol type and the investigated size range.