SLAMS-2.0: a stochastic, Lagrangian aggregate model of sinking particles for the ocean’s biological carbon pump, version 2

Preprint

Sinking biogenic particles drive the ocean biological carbon pump by transferring carbon from the surface ocean to depth, yet the processes governing their formation, transformation, and fate remain poorly constrained by observations. Here we present SLAMS-2.0, a stochastic, particle-resolving Lagrangian modelling framework that simulates the life histories of sinking marine particles across the full size spectrum using fractal scaling and explicit particle-particle and particle-environment interactions informed by observations. Particulate organic carbon (POC) flux emerges from mixed particle formation pathways whose relative contributions evolve with depth. The model is evaluated at six open-ocean time-series sites against seasonal- and depth-resolved observations of POC and biogenic mineral fluxes, particle abundances, and size distributions. Across sites, the model reproduces observed magnitudes and vertical attenuation patterns of POC flux for mechanistically consistent reasons. We show that carbon packaging, porosity, mineral ballast, and aggregation efficiency —-rather than particle size or sinking velocity alone—- govern POC flux transfer to depth. Large, porous, ballast-rich aggregates disproportionately drive deep-ocean POC flux, while small, abundant particles dominate near-surface inventories but contribute less to deep flux. By integrating diverse observational constraints with explicitly simulated particle scale dynamics, SLAMS-2.0 reconstructs unobserved particle processes and provides a platform for investigating how particle physicochemical attributes and interactions shape emergent carbon fluxes.

Authors: Rufas, A; Khatiwala, S; Jokulsdottir, T

• Publication status: Published
• Peer review status: Not peer reviewed
• Preprint server: ESS Open Archive
• Publication date: 2026-03-10
• DOI: 10.22541/essoar.177316670.08373510/v1
• Server owner: Wiley
• Language: English