Pitfalls in atmospheric correction of ocean color imagery: How should aerosol optical properties be computed?: Reply to comment

Gordon [Appl. Opt. 42, 542 (2003)] argues that use of external rather than internal mixing when aerosol optical properties are computed will not seriously affect atmospheric correction of ocean color imagery, in spite of the fact that top of the atmosphere reflectances computed with the two approaches differ significantly as shown by Yan et al. [Appl. Opt. 41, 412 (2002)]. We apply an algorithm for simultaneous retrieval of aerosol optical properties and chlorophyll concentrations to demonstrate that use of the internal-mixing approach leads to atmospheric corrections that differ significantly from those obtained with the more realistic external-mixing approach. For relative humidities of 90% or more, the differences in retrieved aerosol optical properties and chlorophyll concentrations, incurred by application of the internal-mixing approach, become unacceptably large.