An inverse modelling approach to constrain 7Be cycling in the subpolar North Atlantic

Beryllium-7 is a short-lived cosmogenic radionuclide that has been used as a tracer of atmospheric deposition at the sea surface and of physical processes in the upper ocean. These applications generally assume that (i) the fraction of marine ⁷Be in particulate form is negligible, and/or (ii) the interactions between the particulate and dissolved forms of ⁷Be in seawater can be neglected. In this study, we test different steady-state models of upper ocean ⁷Be cycling from measurements of total ⁷Be and particulate ⁷Be activities collected at two stations of the GEOVIDE cruise in the subpolar North Atlantic (May–June 2014). The most complete model includes vertical advection, vertical diffusion, gravitational settling, radioactive decay, atmospheric deposition, and the reversible exchange between dissolved ⁷Be (⁷Be_d) and particulate ⁷Be (⁷Be_p). This model reproduces the measured ⁷Be activities at both stations to within their uncertainties (±1 standard deviation). In the East Greenland-Irminger Current (station 51/60), models that do not consider adsorption and/or desorption can still reproduce the measured activities, while in the southern Labrador Sea (station 69) models that neglect reversible exchange or desorption poorly fit the data. Thus, ⁷Be_d at station 51/60 could have been supplied entirely by surface deposition, whereas ⁷Be_d at station 69 originated at least partly from the release of particulate ⁷Be into solution. The subsurface ⁷Be_p maxima, present at both stations, seem to require a flux of ⁷Be between particles and solution at station 69 but not at station 51/60. At both stations, most of the total ⁷Be deposited at the ocean surface was removed by radioactive decay, with at most 5% removed by sinking of ⁷Be_p, suggesting that reversible exchange can be neglected in applications of total ⁷Be as a tracer for atmospheric deposition. However, reversible exchange is important to consider in applications of ⁷Be_d as a deposition tracer, with a maximum bias of 40% when reversible exchange is neglected. The steady state assumption does not alter our results regarding the solid-solution exchange of ⁷Be, but it does result in a significant bias when deriving atmospheric ⁷Be_tot deposition fluxes from water column ⁷Be_tot inventories. Overall, our findings suggest that reversible exchange could significantly influence the oceanic cycling of ⁷Be at some locations, and should not be systematically neglected when using ⁷Be_d as an oceanic tracer.