Hyperspectral optical absorption closure experiment in complex coastal waters

Abstract

Accurate measurements of absorption data are required for the development and validation of inversion algorithms for upcoming hyperspectral ocean color imaging sensors, such as the NASA Phytoplankton, Aerosol, Cloud, and ocean Ecosystem mission. This study aims to provide uncertainty estimates associated with leading approaches to measure hyperspectral absorption coefficients in complex coastal waters. Absorption spectra were collected at 12 different stations, all located in the Indian River Lagoon, Florida, USA, between 09 January 2017 and 13 January 2017. Measurements included spectral absorption coefficients in the visible range (400–700 nm) associated with dissolved, aCDOM, total particulate, ap, and total nonwater, anw, fractions, and were made both in situ and from discrete samples. Discrete sample approaches included dual-beam spectrophotometer, liquid waveguide capillary cell, point-source integrating cavity absorption meter (PSICAM) for dissolved matter absorption samples, and quantitative filter technique ICAM measurements and the dual-beam spectrophotometer with center-mounted integrating sphere filter pad technique, while the Turner Designs ICAM, and WET Labs AC-s, and AC-9 instruments were used to determine absorption coefficients in situ. The Gershun approach, determining absorption from measurement of the irradiance quartet with respect to depth was also assessed in situ. Measurement uncertainties and relative accuracies were quantified for each of these approaches. Results showed generally strong agreements between different discrete sample methods, with average percent absolute error %δabs < 7% for aCDOM and < 9% for ap. In situ approaches showed higher variability and reduced accuracy. For anw, %δabs deviation relative to PSICAM data was on average 12% to 20%. Results help identify remaining technological gaps and need for improvements in the different absorption measurement approaches.