Yixuan Song , Melissa Omand, Colleen Durkin, Margaret Estapa , Ken Buesseler

While particle-intercepting traps remain a dominant method for quantifying the contribution of sinking particles to the biological carbon pump, fluxes are typically integrated over days to months. Observations of time-varying particle flux over shorter durations are very limited. To this end, we prototyped a camera system called “GelCam” which captures a rapid time-lapse image sequence of particles that settle into a polyacrylamide gel layer located at the base of a sediment trap tube. Here, we describe the system design, post image processing, and results from nine deployments during the EXPORTS campaigns in 2018 and 2021 in the North Pacific and Atlantic ocean. Because wave-driven oscillations of the surface-tethered traps produced a lateral motion of the settled particles, we applied a cross-correlation method for tracking individual particles over time. All particles were subsequently classified into one of six categories based on visual traits, and then quantified into a particulate organic carbon (POC) flux. Using this image-based approach, we are able to distinguish differences in depth-based transfer efficiency among groups, and detect a diel variation in fecal pellet flux. Additionally, the GelCam resolved flux events on timescales shorter than days, allowing for the investigation of covariance among different particle types over short timescales. Paired with the direct recovery of samples to quantify carbon content and high resolution particle images, this approach will enhance our ability to resolve sinking “events” that occur episodically and may be missed when integrating over a traditional trap deployment.