Abstract/Summary

The standing stock of phytoplankton carbon is a fundamental property of oceanic ecosystems, and of critical importance to the development of Earth System models for assessing global carbon pools and cycles. Some methods to estimate phytoplankton carbon at large scales from ocean-colour data rely on the parameterization of carbon-to-chlorophyll ratio, which is known to depend on factors such as the phytoplankton community structure, whereas other methods are based on the estimation of total particulate organic carbon (POC), and rely on the assumption that a known fraction of POC is made up of phytoplankton carbon. The carbon-to-chlorophyll ratio is also used in marine ecosystem models to convert between carbon and chlorophyll, a common requirement. In this paper we present a novel bio-optical algorithm to estimate the carbon-to-chlorophyll ratio, and the standing stocks of phytoplankton carbon partitioned into various size classes, from ocean colour. The approach combines empirical allometric relationships of phytoplankton size structure with an absorption-based algorithm for estimating phytoplankton size spectra developed earlier. Applying the new algorithm to satellite ocean-colour data from September 1997 to December 2013, the spatio-temporal variations of carbon-to-chlorophyll ratio and phytoplankton carbon across various size classes are computed on a global scale. The average annual stock of phytoplankton carbon, integrated over the oceanic mixed-layer depth, is estimated to be ~0.26 gigatonnes, with the size-partitioned stocks of 0.14 gigatonnes for picoplankton, 0.08 gigatonnes for nanoplankton and 0.04 gigatonnes for microplankton. The root-mean-square error and the bias in the satellite-derived estimates of picoplankton carbon, when compared with corresponding in situ data, are found to be 36.23 mgC m-3 and -13.53 mgC m-3, respectively, on individual pixels. The regional uncertainties in the estimates of phytoplankton carbon are calculated to be less than the relative uncertainties in other satellite-derived products, for most parts of the global ocean, and can amplify only for certain oceanographic regions. Although the new estimates of phytoplankton are of the same order of magnitude as those based on existing models, our study suggests that a consensus is yet to be built on the accurate sizes of the phytoplankton carbon pools; improved satellite chlorophyll products, and better estimates of inherent optical properties would be essential pre-requisites to minimising the uncertainties.