Controls on Ocean Color Spectra Observed During the North Atlantic Aerosols and Marine Ecosystems Study (NAAMES)

Satellite ocean color remote sensing is the primary method to retrieve synoptic measurements of the optical properties of the ocean on large spatial and regular time scales. Through bio-optical modeling, changes in ocean color spectra can be linked to changes in marine ecosystem and biogeochemical properties. Bio-optical algorithms rely on assumptions about the covariance of marine constituents as well as the relationships among their inherent and apparent optical properties. Validation with in situ measurements of in-water constituents and their optical properties is required to extrapolate local knowledge about ocean color variations to global scales. Here, we evaluate seasonal and spatial relationships between optical constituents and their inherent and apparent optical properties throughout the annual cycle of the North Atlantic plankton bloom using bio-optical data from four cruises conducted as part of the North Atlantic Aerosols and Marine Ecosystems Study (NAAMES). Our results show ocean color variability, quantified using field observations of the remote sensing reflectance spectrum at each NAAMES station, is driven by colored dissolved organic matter (CDOM) absorption in the ultraviolet wavelengths, phytoplankton absorption in the blue wavelengths, and total particulate backscattering in the green wavelengths. Results from a recently storm-mixed station at the height of the spring bloom demonstrate that significant changes in bio-optical properties can occur on daily scales. By testing the effects of variations in lighting conditions and solar geometries, we also demonstrate that, for this data set, remote sensing reflectance should be considered a quasi-inherent optical property. We find that the temporal and spatial chlorophyll concentrations and the magnitudes of inherent optical properties can be accurately assessed using previously published ocean color algorithms. However, changes in the spectral slopes of the inherent optical properties are often poorly retrieved, indicating the need for improvements in the retrieval of optical constituent composition. The characterization of such a dynamic environment provides beneficial insights for future bio-optical algorithms.