Autonomous sensors for gravitational carbon flux in the ocean are critically needed, because of uncertainties in the projected response of the biological carbon pump (BCP) to climate change, and the proposed, engineered acceleration of the BCP to sequester carbon dioxide in the ocean. Optical sediment trap (OST) sensors directly sense fluxes of sinking particles in a manner that is independent of, and complementary to, other autonomous, sensor-derived estimates of BCP fluxes. However, limited intercalibrations of OSTs with traditional sediment traps and uncharacterized, potential biases have limited their broad adoption. A global field dataset spanning 3 orders of magnitude in carbon flux was compiled and used to develop empirical models predicting particulate organic carbon flux from OST observations, and intercalibrating different sensor designs. These data provided valuable constraints on the uncertainty in the predicted carbon flux and showed a quantitative, theoretically consistent relationship between observations from OSTs with collimated and diffuse optical geometries. While not designed for this purpose, commercial beam transmissometers have been used as OSTs, so two models were developed quantifying the biases arising from the transmissometer’s housing geometry and optical beam diameter. Finally, an algorithm for the quality control of beam transmissometer-derived OST data was optimized using sensitivity tests. The results of this study support the expansion of OST-based gravitational carbon flux measurements and provide a framework for interpretation of OST measurements alongside other gravitational particle flux observations. These findings also suggest key features that should be included in designs of future, purpose-built OST sensors.

Particulate organic matter settling out of the euphotic zone is a major sink for atmospheric carbon dioxide and serves as a primary food source to mesopelagic food webs. Degradation of this organic matter encompasses a suite of mechanisms that attenuate flux, including heterotrophic metabolic processes of microbes and metazoans. The relative contributions of microbial and metazoan heterotrophy to flux attenuation, however, have been difficult to determine. We present results of compound specific nitrogen isotope analysis of amino acids of sinking particles from sediment traps and size‐fractionated particles from in situ filtration between the surface and 500 m at Ocean Station Papa, collected during NASA EXPORTS (EXport Processes in the Ocean from RemoTe Sensing). With increasing depth, we observe: (1) that, based on the δ15N values of threonine, fecal pellets dominate the sinking particle flux and that attenuation of downward particle flux occurs largely via disaggregation in the upper mesopelagic; (2) an increasing trophic position of particles in the upper water column, reflecting increasing heterotrophic contributions to the nitrogen pool and the loss of particles via remineralization; and (3) increasing δ15N values of source amino acids in submicron and small (1–6 μm) particles, reflecting microbial particle solubilization. We further employ a Bayesian mixing model to estimate the relative proportions of fecal pellets, phytodetritus, and microbially degraded material in particles and compare these results and our interpretations of flux attenuation mechanisms to other, independent methods used during EXPORTS.