Particle Cycling in the Ocean based on Titanium and Particle Concentration Data from the US GEOTRACES North Atlantic Transect

As titanium (Ti) is very insoluble in seawater, particulate Ti (pTi) can be used as an inert tracer of particle dynamics. The partitioning of pTi between small and large particles is mainly driven by aggregation and disaggregation of particles. In this study, we estimated rates of those particle processes by combining pTi data (small size fraction, 0.8-51 um; large size fraction, >53um) from the US GEOTRACES North Atlantic Transect (GA03) with a simple 1-D particle dynamics model. The model is similar to the models previously used to describe oceanic thorium (Th) and particle profiles (e.g. Clegg and Whitfield, 1993), but with particle dynamic terms only. We assume Ti in the ocean is supplied by eolian Ti inputs to the surface mixed layer, and then take inversions of the pTi data using a generalized nonlinear least-squares technique to estimate aggregation (β₂), disaggregation (β-₂), and particle sinking (w) rate constants. A preliminary study using the data from station GT11-22 shows higher aggregation rate constant in the surface layer to explain the low small:large pTi ratio, and disaggregation rate constants increasing between the bottom of mixed layer and ~400m, where small:large pTi increases rapidly. Below ~400m, aggregation and disaggregation rate constants are relatively invariable reflecting constant small:large pTi ratios at depths, but particle sinking rate increases with depth, e.g. increasing by almost three-fold between ~400m and 3600m.

In this presentation, we will add small and large particle concentration data to this model. Unlike Ti, particle concentrations are also subjected to remineralization processes. We can thus separate the loss of particles by remineralization by considering pTi and particle concentration data together. Finally we will compare our results to previous studies using Th isotopes and/or particle concentrations, and discuss what information we can obtain from the Ti-based particle dynamics model.

Clegg, S.L., Whitfield, M. 1993. Application of a generalized scavenging model to time series 234Th and particle data obtained during the JGOFS North Atlantic bloom experiment. Deep-Sea Res. I 40(8), 1529-1545.