POSTER Session 3

PERFORMANCE AND UNCERTAINTY ASSESSMENT OF A NOVEL MULTI-STEP SEMI-ANALYTICAL ALGORITHM FOR ESTIMATING SEAWATER INHERENT OPTICAL PROPERTIES FROM OCEAN REFLECTANCE

A current challenge of inversion models aiming to derive inherent optical properties (IOPs) of seawater constituents from radiometric measurements is that they often provide a limited number of IOPs or rely on highly restrictive assumptions about the spectral shapes of output variables. To alleviate these limitations, a novel multi-step (4-step) Semi-Analytical Algorithm (4SAA) has been developed that implements four independent component models to retrieve the hyperspectral diffuse attenuation coefficient for downwelling irradiance, Kd, and a suite of total and constituent IOPs from input remote-sensing reflectance. This structure is advantageous as it enables independent optimization of each component model, in-depth characterization of uncertainties associated with each component model, and uncertainty propagation through the multi-step sequence of the algorithm. Here, we describe one approach to validate the first two steps of 4SAA utilizing a recently published synthetic optical database. The first two component models, referred to as KdNN (step 1) and LS2 (step 2), provide estimates of Kd and the total and non-water absorption and backscattering coefficients. To conduct our uncertainty analysis, we utilized an approach to perturb the input parameters of both KdNN and LS2. The performance and uncertainty assessment of both models are determined from 350 to 700 nm. Initial results suggest a satisfactory performance of the two models versus the synthetic database and provide insight into the propagation of uncertainty through the first two steps of 4SAA. We anticipate that this validation analysis will support the estimation of IOPs from ocean color measurements from the NASA PACE satellite mission.

*Matthew Kehrli, Scripps Institution of Oceanography, University of California San Diego, USA, [email protected], https://orcid.org/0000-0003-1268-676X

Dariusz Stramski, Scripps Institution of Oceanography, University of California San Diego, USA, [email protected], https://orcid.org/0000-0002-8518-072X

Rick Reynolds, Scripps Institution of Oceanography, University of California San Diego, USA, [email protected], https://orcid.org/0000-0002-1579-3600

Daniel Jorge, Laboratoire d’Océanologie et de Géosciences (LOG), France, [email protected], https://orcid.org/0000-0002-4812-9239

Cédric Jamet, Laboratoire d’Océanologie et de Géosciences (LOG), France, [email protected], https://orcid.org/0000-0001-6988-6506

Hubert Loisel, Laboratoire d’Océanologie et de Géosciences (LOG), France, [email protected], https://orcid.org/0000-0003-2905-6465

Code of Conduct
Privacy Policy