How many independent quantities can be extracted from ocean color?

Abstract

The reflectance of sunlight from the ocean can be observed from satellites and is used to derive many biologically relevant parameters, such as the concentration of chlorophyll in the upper ocean. Reflectances are currently observed at about 10 different wavelengths, but this will soon be expanded to hundreds with the upcoming launch of a new ocean color satellite, PACE, in early 2024. Many new algorithms are being proposed to make use of the wealth of ocean color data which will be provided. However, there are strong correlations between reflectances at different wavelengths; these correlations mean there will be far fewer products that can be independently derived than there will be reflectance wavelengths observed. Here, we use ship-based measurements similar to what will be provided from PACE to suggest that, on a global scale, only a few independent variables can be calculated from hundreds of reflectance wavelengths. Current and past satellites provide a similar amount of independent data to what is projected from PACE. We then show that, on a global scale, a set of six derived parameters only contains one independent piece of information, suggesting that more information exists in ocean color data than is being currently used.

Expanded Signal to Noise Ratio Estimates for Validating Next-Generation Satellite Sensors in Oceanic, Coastal, and Inland Waters

“…phytoplankton groups [68,69] or trace gasses such as methane [70,71], but a more common use is to locally tune band-ratio algorithms to achieve better performance [72,73] or to track spectrally dynamic features such as fluorescence line height [74] using a subset of fixed wavelengths. Recent analysis shows that global ocean hyperspectral data have comparable degrees of freedom compared to legacy sensors such as MODIS because of strong covariance across wavelengths [75]. While this may set an upper limit on the number of parameters that can be independently extracted from hyperspectral imagery, it does not negate the use-case for the identification of spectrally narrow features or for band-switching algorithms.…”

Extending aquatic spectral information with the first radiometric IR-B field observations

“…polarimetry). Hyperspectral sensing increases the number of data products available for extracting information, albeit at the expense of radiometric accuracy ( 29 ) and with high correlation between spectrally adjacent wavebands ( 12 , 44 ). Extraction of polarization information may improve atmospheric correction ( 45 ) and enable new applications ( 46 ).…”

Multidecadal changes in ocean transparency: Decrease in a coastal upwelling region and increase offshore

“…The high correlation among Chl a , Z SD , NPP, and other satellite‐derived variables is artificially increased when using satellite‐derived estimates. All these products are derived from remote sensing reflectances that are inter‐correlated and have few degrees of freedom (Cael et al 2023). With these caveats in mind, it is not surprising that Z SD may be better correlated with a model of primary production than different models of primary production using different input data with each other.…”