© The Author(s), 2016. This is the author's version of the work and is distributed under the terms of the Creative Commons Attribution License. The definitive version was published in Deep Sea Research Part I: Oceanographic Research Papers 113 (2016): 57-79, doi:10.1016/j.dsr.2016.03.008.
Thorium is a highly particle-reactive element that possesses different measurable radio-isotopes in seawater, with well-constrained production rates and very distinct half-lives. As a
result, Th has emerged as a key tracer for the cycling of marine particles and of their chemical
constituents, including particulate organic carbon.
Here two different versions of a model of Th and particle cycling in the ocean are tested using an unprecedented data set from station GT11-22 of the U.S. GEOTRACES North Atlantic Section: (i) 21 228;230;234Th activities of dissolved and particulate fractions, (ii) 228Ra activities,
(iii) 234;238U activities estimated from salinity data and an assumed 234U/238U ratio, and (iv)
particle concentrations, below a depth of 125 m. The two model versions assume a single class
of particles but rely on different assumptions about the rate parameters for sorption reactions
and particle processes: a first version (V1) assumes vertically uniform parameters (a popular description), whereas the second (V2) does not. Both versions are tested by fitting to the
GT11-22 data using generalized nonlinear least squares and by analyzing residuals normalized
to the data errors.
We find that model V2 displays a significantly better fit to the data than model V1. Thus,
the mere allowance of vertical variations in the rate parameters can lead to a significantly better
fit to the data, without the need to modify the structure or add any new processes to the model.
To understand how the better fit is achieved we consider two parameters, K = k1=(k-1 + β-1)
and K/P, where k1 is the adsorption rate constant, k-1 the desorption rate constant, β-1 the
remineralization rate constant, and P the particle concentration. We find that the rate constant
ratio K is large (≥0.2) in the upper 1000 m and decreases to a nearly uniform value of ca.
0.12 below 2000 m, implying that the specific rate at which Th attaches to particles relative
to that at which it is released from particles is higher in the upper ocean than in the deep
ocean. In contrast, K/P increases with depth below 500 m. The parameters K and K/P
display significant positive and negative monotonic relationship with P, respectively, which is
collectively consistent with a particle concentration effect.
We acknowledge the U.S. National Science Foundation for providing funding for this study
(grant OCE-1232578) and for U.S. GEOTRACES North Atlantic section ship time, sampling,
and data analysis.
Woods Hole Open Access Server