POSTER Session 1

Retrieval of hyperspectral Kd from ocean color remote sensing measurements in support of PACE satellite mission

Many algorithms have been developed over the past decades to retrieve the diffuse attenuation coefficient for downwelling plane irradiance (Kd) using optical remote sensing data. Past approaches include empirical and analytical models, multi-band and band-ratio ocean reflectance approaches, or the use of other ocean color products such as chlorophyll-a concentration and the particulate backscattering coefficient. In this study a new algorithm for the estimation of Kd averaged over the first attenuation depth (<Kd>_1) was developed for the NASA Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) satellite mission. The algorithm is based on the neural network framework described in Jorge et al. (Remote Sens. Environ, 263, 2021), but with the inclusion of additional bands of ocean reflectance available from the PACE Ocean Color Instrument (OCI) measurements. This new configuration enables the retrieval of <Kd>_1 from the ultraviolet (350 nm) through the near-infrared (750 nm) spectral region. The algorithm was trained using hyperspectral synthetic optical databases and validated using in situ measurements collected from a variety of marine optical environments. As a result of testing the number of input reflectance bands, 12 bands from the visible part of the spectrum were selected which represent a compromise between minimizing the number of required input bands and good algorithm performance for both a reference wavelength (490 nm) and the whole spectrum. The newly developed algorithm showed improved results when compared with other existing algorithms for both the 490-nm and hyperspectral retrievals. The performance of the algorithm is currently being assessed using newly released PACE images.

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

Cedric Jamet, Laboratoire d’Océanologie et de Géosciences (LOG), France

Hubert Loisel, Laboratoire d’Océanologie et de Géosciences (LOG), France

Matthew Kehrli, Scripps Institution of Oceanography, University of California San Diego, USA

Rick Reynolds, Scripps Institution of Oceanography, University of California San Diego, USA

Dariusz Stramski, Scripps Institution of Oceanography, University of California San Diego, USA

Code of Conduct
Privacy Policy