New ocean color algorithms for estimating the concentration of particulate organic nitrogen from remote sensing in oceanic surface waters

Abstract

The concentration of particulate organic nitrogen (PON) in seawater plays a central role in ocean biogeochemistry. Limited availability of PON data obtained directly from in situ sampling methods hinders our ability to better characterize the spatio-temporal variability of PON within the global ocean. Tight relationships have recently been developed between in situ measurements of seawater inherent optical properties (IOPs) and PON. Knowing that IOPs can now be estimated from ocean color remote-sensing, these relationships could then be used to assess PON from semi-analytical algorithms applied to satellite ocean color observations. The present study aims at evaluating which IOPs, as estimated from space, can be used as the best proxy for the remote sensing retrieval of PON. The different considered IOPs are the absorption coefficients of total particulate matter, ap(λ), phytoplankton, aph(λ), and non-algal particles, ad(λ), as well as the particulate backscattering coefficient, bbp(l). IOPs have been derived from satellite ocean remote-sensing reflectance, Rrs(l), using different available inverse methods. The validation of the algorithms is based on matchup between an extensive dataset of 156 concurrent measurements of in situ PON and satellite-derived particulate IOPs. Our results show that reasonably strong PON vs satellite-derived IOPs relationships hold across a range of diverse oceanic and coastal environments. aph(443) shows the best ability to serve as a PON proxy over a broad range of PON from open ocean oligotrophic to coastal waters (MdAPD of 25.39 %). bbp(555) can also be considered as a good proxy of PON in open ocean environments (MdAPD of 22.03 %). Comparison with in situ time series over ten years shows also the good performance of the algorithm to reproduce the seasonal variation of PON. The application of this algorithm to Moderate-Resolution Imaging Spectroradiometer (MODIS; 2002-present) observations provides global PON distributions pattern which agree with in situ PON expected geographical distribution. High PON concentrations are observed in turbid shelf and coastal regions as well as in upwelling areas; while low PON are observed in oligotrophic regions. The presented relationships demonstrate a promising means to assess long-term trends and/or budget of PON in specific areas of the ocean or at the global oceanic scale.