Observation and model-based analyses of ocean biological carbon fluxes and ecosystem dynamics

Abstract

This dissertation examines the ocean biological carbon pump, a complex coupling of biophysical processes that exports carbon from the surface to the deep (5-15 Pg C yr-1), by focusing on the largest component, the gravitational sinking flux (4-9 Pg C yr-1). In the first chapter, I examine the impact of interannual variability, specifically marine heat waves, on ecosystem and export production in the Northeast Pacific using a suite of observational data and an ocean biogeochemical model. The model shows that warming-induced strati- fication during marine heat waves relieves winter light limitation while decreasing nutrient supply increasing small phytoplankton production at the expense of large phytoplankton, in agreement with observations. This shift in the phytoplankton assemblage is propagated through the food web, leading to a smaller zooplankton assemblage, and weaker export. I also show that previous observation-based estimates of ecosystem production misattributed a spatial gradient in nutrients to a rapid decline in productivity during a recent marine heat wave. These results highlight the difficulty in disentangling spatial and temporal variability when interpreting sparse observations. In Chapters two and three, I examine the export proxy, thorium-234, used to estimate particle fluxes in situ. I detail the collection of over 1500 individual water samples collected in the Northeast Pacific and the North Atlantic during the 2018 and 2022 NASA EX- PORTS field campaigns. These data were analyzed to produce dozens of particle sinking flux profiles using 1-dimensional (1D) thorium budgets that rely on the assumption that the influence of physical dynamics is negligible relative to the flux of thorium adsorbed on sinking particles. Using observed and simulated velocities, I find that physical processes could bias observed flux estimates by 30% in both regions (Chapter 2). In Chapter 3, using an idealized double gyre model to examine physical thorium transport in fine-scale struc- tures I show that coherent mesoscale eddies (lasting > 20 days) are well-suited for thorium sampling due to the decreased likelihood of strong physical transport. In contrast, frontal regions exhibit large vertical velocities that decouple the sinking flux and 1D thorium esti- mate, suggesting that thorium-based particle sinking fluxes across frontal regions should be interpreted cautiously.