Aerosol models from the AERONET database: application to surface reflectance validation

• Published in: Atmospheric Measurement Techniques
• Main Authors: Roger, Jean-Claude, Vermote, Eric, Skakun, Sergii, Murphy, Emilie, Dubovik, Oleg, Kalecinski, Natacha, Korgo, Bruno, Holben, Brent
• Format: Text
• Language: English
• Published: 2022
• Subjects: Aerosol Robotic Network
• Online Access: https://doi.org/10.5194/amt-15-1123-2022; https://amt.copernicus.org/articles/15/1123/2022/

Aerosols play a critical role in radiative transfer within the atmosphere, and they have a significant impact on climate change. In this paper, we propose and implement a framework for developing an aerosol model using their microphysical properties. Such microphysical properties as the size distribution, the complex refractive index, and the percentage of sphericity are derived from the global AERosol RObotic NETwork (AERONET). These measurements, however, are typically retrieved when almucantar measurement procedures are performed (i.e., early mornings and late afternoons with clear sky) and might not have a temporal correspondence to a satellite overpass time, so a valid validation of satellite-derived products cannot be carried out. To address this problem of temporal inconsistency of satellite and ground-based measurements, we developed an approach to retrieve these microphysical properties (and the corresponding aerosol model) using the optical thickness at 440 nm, τ 440 , and the Ångström coefficient between 440 and 870 nm, α 440–870 . Such aerosol models were developed for 851 AERONET sites within the last 28 years. Obtained results suggest that empirically microphysical properties can be retrieved with uncertainties of up to 23 %. An exception is the imaginary part of the refractive index ni, for which the derived uncertainties reach up to 38 %. These specific parametric models of aerosol can be used for the studies when retrieval of microphysical properties is required as well as validation of satellite-derived products over land. Specifically, we demonstrate the usefulness of the aerosol models to validate surface reflectance records over land derived from optical remote sensing sensors. We then quantify the propagation of uncertainties in the surface reflectance due to uncertainties with the aerosol model retrieval that is used as a reference from radiative transfer simulations. Results indicate that individual aerosol microphysical properties can impact uncertainties in surface reflectance retrievals between 3.5 × 10 −5 to 1 × 10 −3 (in reflectance units). The overall impact of microphysical properties combined yields an overall uncertainty in surface reflectance < 0.004 (in reflectance units). That corresponds, for example, to 1 to 3 % of the retrieved surface reflectance in the red spectral band (620–670 nm) by the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument. These uncertainty values are well below the specification (0.005 + 0.05 ρ ρ is the retrieved surface reflectance) used for the MODIS atmospheric correction.