Intercomparison of aerosol volume size distributions derived from AERONET ground-based remote sensing and LARGE in situ aircraft profiles during the 2011–2014 DRAGON and DISCOVER-AQ experiments

Joel S Schafer,Patricia Sawamura,Kenneth L Thornhill,Tom F Eck,Edward L Winstead,Richard H Moore,Alexander Smirnov,Ilya Slutsker,Bruce E Anderson,Andreas J Beyersdorf,Brent N Holben,Alexander Sinyuk,Luke D Ziemba,David M Giles

doi:10.5194/amt-12-5289-2019

Abstract

Abstract. Aerosol volume size distribution (VSD) retrievals from the Aerosol Robotic Network (AERONET) aerosol monitoring network were obtained during multiple DRAGON (Distributed Regional Aerosol Gridded Observational Network) campaigns conducted in Maryland, California, Texas and Colorado from 2011 to 2014. These VSD retrievals from the field campaigns were used to make comparisons with near-simultaneous in situ samples from aircraft profiles carried out by the NASA Langley Aerosol Group Experiment (LARGE) team as part of four campaigns comprising the DISCOVER-AQ (Deriving Information on Surface conditions from Column and Vertically Resolved Observations Relevant to Air Quality) experiments. For coincident (±1 h) measurements there were a total of 91 profile-averaged fine-mode size distributions acquired with the LARGE ultra-high sensitivity aerosol spectrometer (UHSAS) instrument matched to 153 AERONET size distributions retrieved from almucantars at 22 different ground sites. These volume size distributions were characterized by two fine-mode parameters, the radius of peak concentration (rpeak_conc) and the VSD fine-mode width (widthpeak_conc). The AERONET retrievals of these VSD fine-mode parameters, derived from ground-based almucantar sun photometer data, represent ambient humidity values while the LARGE aircraft spiral profile retrievals provide dried aerosol (relative humidity; RH <20 %) values. For the combined multiple campaign dataset, the average difference in rpeak_conc was 0.033±0.035 µm (ambient AERONET values were 15.8 % larger than dried LARGE values), and the average difference in widthpeak_conc was 0.042±0.039 µm (AERONET values were 25.7 % larger). For a subset of aircraft data, the LARGE data were adjusted to account for ambient humidification. For these cases, the AERONET–LARGE average differences were smaller, with rpeak_conc differing by 0.011±0.019 µm (AERONET values were 5.2 % larger) and widthpeak_conc average differences equal to 0.030±0.037 µm (AERONET values were 15.8 % larger).

Highlights

Atmospheric aerosol volume size distribution information is relevant to the modeling of radiative transfer, weather processes and human health due to air quality concerns (Peng et al, 2018; Li et al, 2015; Sheng et al, 2019; Eilenberg et al, 2018; Gong et al, 2003)
The Aerosol Robotic Network (AERONET) global monitoring program provides an opportunity to capture seasonal and diurnal trends in extinctionweighted column-integrated aerosol volume size distributions and concentrations for ambient atmospheric conditions derived from frequent sky radiance measurements and spectral aerosol optical depth
This paper presents a large number of comparisons of multiple fine-mode volume size distribution datasets from four US regions for in situ measurements from repeated aircraft profiles during a series of month-long intensive DISCOVER-AQ (Deriving Information on Surface conditions from Column and Vertically Resolved Observations Relevant to Air Quality) campaigns conducted between 2011 and 2014

Summary

Introduction

Atmospheric aerosol volume size distribution information is relevant to the modeling of radiative transfer, weather processes and human health due to air quality concerns (Peng et al, 2018; Li et al, 2015; Sheng et al, 2019; Eilenberg et al, 2018; Gong et al, 2003). During the Indian Ocean Experiment (INDOEX) Clarke et al (2002) computed lognormal fits of volume size distributions for in situ measurements for fine-mode pollution acquired by ship and aircraft in the Arabian Sea under high aerosol loading and found that the average accumulation mode volume peak radius ranged from 0.17 to 0.18 μm with computed geometric standard deviations for ship data equal to 1.51 and aircraft data equal to 1.43. These values are similar to the AERONET-retrieved averages from 1998 to 2000 in the Maldives (Kaashidhoo) in the same region. Given the typical complexity of aircraft campaigns and the fact that the validation of AERONET retrieval products is rarely a central campaign goal, this well-coordinated effort resulted in a dataset unique for the quantity and nearsimultaneous nature of the comparisons

Instrumentation

Method

Aerosol volume size distribution comparisons

Comparisons with humidification adjustment of LARGE volume size distributions

Findings

Conclusions