Atmospheric correction of ocean color imagery: use of the Junge power-law aerosol size distribution with variable refractive index to handle aerosol absorption

Abstract

When strongly absorbing aerosols are present in the atmosphere, the usual two-step procedure of processing ocean color data-(1) atmospheric correction to provide the water-leaving reflectance (rho(w)), followed by (2) relating rho(w) to the water constituents-fails and simultaneous estimation of the ocean and aerosol optical properties is necessary. We explore the efficacy of using a simple model of the aerosol-a Junge power-law size distribution consisting of homogeneous spheres with arbitrary refractive index-in a nonlinear optimization procedure for estimating the relevant oceanic and atmospheric parameters for case 1 waters. Using simulated test data generated from more realistic aerosol size distributions (sums of log-normally distributed components with different compositions), we show that the ocean's pigment concentration (C) can be retrieved with good accuracy in the presence of weakly or strongly absorbing aerosols. However, because of significant differences in the scattering phase functions for the test and power-law distributions, large error is possible in the estimate of the aerosol optical thickness. The positive result for C suggests that the detailed shape of the aerosol-scattering phase function is not relevant to the atmospheric correction of ocean color sensors. The relevant parameters are the aerosol single-scattering albedo and the spectral variation of the aerosol optical depth. We argue that the assumption of aerosol sphericity should not restrict the validity of the algorithm and suggest an avenue for including colored aerosols, e.g., wind-blown dust, in the procedure. A significant advantage of the new approach is that realistic multicomponent aerosol models are not required for the retrieval of C.

Sensor Capability and Atmospheric Correction in Ocean Colour Remote Sensing

“…A number of atmospheric correction methods, such as the spectral optimisation algorithm (SOA) [164,165] and the NeuroVaria (neuro-variational) method [166,167], have made use of a simple model that uses a Junge power-law distribution to characterise the size of aerosol particles, together with a complex refractive index that is wavelength independent. For the spectral optimisation algorithm, the diameter of the particles is D, and dN is the number of particles per unit volume in the size interval D ± dD/2 [164]:…”

“…For the spectral optimisation algorithm, the diameter of the particles is D, and dN is the number of particles per unit volume in the size interval D ± dD/2 [164]:…”

“…The six values for size parameter v were chosen at intervals of 0.5 between 2.0 and 4.5 [164]. As models based on the Junge power-law distribution do not require a comprehensive set of in situ measurements, they have the potential to be applied universally.…”

“…The spectral optimisation algorithm (SOA) [164,165] employs a simple aerosol model which determines the size of aerosol particles with a Junge power-law distribution. This distribution is a simplification, which enables SOA to achieve optimisation when there are a limited number of spectral bands [223].…”

Simultaneous determination of aerosol optical thickness and water leaving radiance from multispectral measurements in coastal waters

“…CC BY 4.0 License. observation data (Gordon, 1998;Fougnie et al, 1999;Wang and Franz, 2000;Murakami et al, 2005;Yoshida et al, 2005;Franz et al, 2007).…”