Robust Algorithm Based on the Reflectance Curvature for Estimating Particulate Organic Carbon and its Spatiotemporal Variability in the Global Ocean

Estimation of particulate organic carbon (POC) is essential for the studies of biological carbon export from the surface to the deep ocean, carbon-based net primary production, phytoplankton growth rate, and global carbon cycle. Despite the number of regional and global algorithms reported in earlier studies, an accurate estimation of POC and its spatiotemporal variability from satellite ocean color data are often hampered by biases associated with the algorithm parameterizations and a lack of in situ data for the coastal oceans with complex physical and biogeochemical processes (such as physical mixing, biological production, horizontal and vertical transport of POC through the ocean currents and circulations, and POC sinking fluxes). In this study, we developed a simple maximum band ratio index (MBRI) algorithm based on global in situ POC and remote sensing reflectance data and validated and intercompared with existing algorithms using independent in situ and MODIS-Aqua POC data in global oceanic waters. In general, the POC products estimated by the MBRI algorithm have greater accuracy with a mean relative error of 0.218, a root mean square error of 34.07, and a correlation coefficient of 0.88. The MBRI approach was further applied to time-series satellite data to analyze the spatiotemporal variations and trends in POC in regional/global oceanic waters as well as the Arctic Ocean (AO) region. This study highlighted a substantial change and an increase in POC fields in the AO region in response to global climate change over the recent decades.