An Inverse Model for Estimating the Optical Absorption and Backscattering Coefficients of Seawater From Remote‐Sensing Reflectance Over a Broad Range of Oceanic and Coastal Marine Environments

Abstract

We present an inverse model (referred to as LS2) for estimating the inherent optical properties (IOPs) of seawater, specifically the spectral absorption, a(λ), and backscattering, bb(λ), coefficients within the ocean surface layer, from measurements of ocean remote‐sensing reflectance, Rrs(λ). The nonwater absorption, anw(λ), and particulate backscattering, bbp(λ), coefficients can be derived after subtracting pure seawater contributions. The LS2 requires no spectral assumptions about IOPs and provides solutions at arbitrary light wavelengths in the visible spectrum independently of one another. As the LS2 can operate with the inputs of Rrs(λ) and solar zenith angle it is applicable to satellite ocean color remote sensing. The model can also operate with additional input of the diffuse attenuation coefficient of downward irradiance, which provides somewhat improved model performance for applications using in situ radiometric measurements as inputs. The evaluation of LS2 with a synthetic data set that is free of measurement errors indicates good performance for IOPs in the visible spectrum, except for anw(λ) in the long‐wavelength portion of the spectrum where anw(λ) contributes only a few percent to a(λ) under typical open ocean conditions. The good performance is characterized by a median absolute percent difference between the model‐derived and true values of IOPs, which is generally <20%, and the median ratio of model‐derived to true values <10%. The satisfactory model performance is also demonstrated through validation analysis based on extensive data sets comprising coincident in situ measurements of Rrs(λ) and IOPs as well as a match‐up data set comprising satellite‐derived Rrs(λ) and in situ IOP measurements.

Optical characterization of marine phytoplankton assemblages within surface waters of the western Arctic Ocean

“…There is, however, an increasing capability to obtain measurements of the spectral absorption coefficient from in situ instrumentation (Slade et al 2010;Wollschläger et al 2013) or to estimate it through the use of inverse models applied to observations obtained from air or spaceborne ocean color sensors (Werdell et al 2018). Recent work suggests that these latter inversion models can perform with reasonable accuracy in the Arctic (Zheng et al 2014;Loisel et al 2018). In parallel, there has been recent progress in the development of mathematical approaches to partitioning data of the absorption coefficient into the separate contributions of various seawater constituents, including phytoplankton (Lee et al 2002;Ciotti and Bricaud 2006;Zheng and Stramski 2013).…”

Semianalytical Derivation of Phytoplankton, CDOM, and Detritus Absorption Coefficients From the Landsat 8/OLI Reflectance in Coastal Waters

“…The performance of our approach in estimating a(412) is comparable to other SA algorithms, which used the R rs (412) measurements for the inversion (e.g., Loisel et al, 2018;Werdell et al, 2013). The MAPD values varying between 30% and 39% (see Table 1) are not a small uncertainty for a ph (443) and a dg (443) estimation.…”

Nature of optical products inverted semianalytically from remote sensing reflectance of stratified waters

“…Remote sensing algorithms (Loisel and Stramski ; Loisel et al ) were also developed to invert a RS ( λ ) and from a combination of R rs and K d (termed as RKA hereon for such R rs ‐ K d algorithms). This is based on that both R rs and K d are functions of a and b b (Gordon et al ; Gordon ; Lee et al ); therefore, it is a system of two equations for two unknowns.…”

“…This is based on that both R rs and K d are functions of a and b b (Gordon et al ; Gordon ; Lee et al ); therefore, it is a system of two equations for two unknowns. When a RS ( λ ) and are inverted from R rs ( λ ) via RKA, however, K d ( λ ) is first estimated from the R rs ( λ ) spectrum through empirical regressions (Loisel et al ) or neural networks (Loisel et al ), that is, the K d ( λ ) in RKA is not an independent measurement. Because generally a dominates K d at a given wavelength for most natural waters (Gordon ; Lee et al ), a RS ( λ ) inverted via RKA is mainly dependent on the empirically estimated , with mainly determined from R rs ( λ ) and the inverted a RS ( λ ).…”