Estimating particulate organic matter flux from in-situ optics: A framework for correcting for suspended particles and incorporating depth-dependent degradation

Abstract. Accurate quantification of the particulate organic matter (POM) flux from settling particles is crucial for understanding the efficiency of oceanic CO2 sequestration. Recent advancements in in-situ camera systems (ISCs) have enabled high-resolution estimates of the size distribution of particle concentration (PSDc) in the global ocean, both vertically and horizontally. When calibrated against corresponding sediment trap flux observations, these PSDc data serve as a basis for estimating downward POM flux. However, these estimates are subject to several uncertainties. A central unresolved issue lies in differentiating settling particles from suspended particles in ISC data. Additionally, the conventional method for converting PSDc to POM flux – which estimates a particle's contribution to the total flux based on a power-law relationship with its size – does not explicitly integrate particle sinking velocities and degradation rates, two key factors regulating POM export. Overcoming these limitations requires incorporating spatiotemporally aligned supplementary data and refined methodological approaches. Here, to address these uncertainties, we introduce a process-based, three-step framework that improves flux estimates relative to the empirically optimized conventional method. First, to correct for the contribution of non-sinking particles (suspended particles and zooplankton), we calibrate ISC data against particle flux measurements from gel traps using a measured size–velocity relationship, thereby yielding a robust estimate of the sinking particle flux. Second, leveraging the output from the first step, we build upon the conventional power-law approach by developing a mechanistic particulate organic carbon (POC) flux model that incorporates size-specific sinking velocities and a depth-dependent degradation term modulated by water temperature and oxygen concentration. Third, with the POC flux estimates from the second step as input, we extend the framework to estimate particulate organic nitrogen (PON) flux using a simple dynamic model for particulate organic C : N stoichiometry. By linking the C : N ratio to particle age, the model reproduces the increase in C : N with depth – a key biogeochemical trend driven by preferential nitrogen remineralization over carbon that is not resolved by conventional, static-ratio modeling approaches.