Table Of Contents

This Page

Notes from “Eddy-driven sources and sinks of nutrients in the upper ocean: Results from a 0.1 deg. resolution model of the North Atlantic” by McGillicuddy et al. (2003)

Things to do:

  • Annual cycle of nitrate at 5 and 100 m compared to observations at station ALOHA.
  • Time series of horizontal and vertical advective flux of nutrients at 120 m depth.
  • Convective flux is the largest term at BATS: Ho to compute it?

Additional notes (07/09/2010)

The author discuss the various mechanims and their respective timescale to restore nutrient in the aphotic zone. The first mechanism is the ambient flux of sinking particulate organic matter (POM) with a timescale of about 10–100 days. The second one is the isopycnal transport of nutrient with a timescale inversely proportional to the horizontal diffusivity coefficient and varying from 2 months to 2 years. The last one is the diapycnal diffusivity with a timescale much larger than one year. Several runs are performed with varying horizontal diffusivity coefficient and thus varying restoring timescale. All runs have a similar skill in reproducing the observations suggesting that the model results are not too sensitive to the restoring timescale. The best simulation is the one with the shortest restoring with a timescale of 10 days.

Eddy advection in this model is mostly positive across the subtropical gyre while in the model of Oschlies (2002b) that has about 1/9-deg. resolution in the horizontal, the eddy advection is found strongly negative along the southeastern flank of the gyre. Furthermore, at BATS, although in the present model, annual new production reaches about 0.63 mol N/m^2/year with the vertical advection acounting for about 0.12 mol N/m^2/year, in Oschlies’ model, the annual new production is only 0.16 mol N/m^2/year with vertical advection acounting for only 0.02 mol N/m^2/year. The low value of annual new production in Oschlies’ model is mainly due to the forcing period corresponding to a high value of the North Atlantic Oscillation and a shallower convective winter mixing. According to McGillicuddy et al., the reduced contribution of vertical advection in oschlies’ model is due to the underestimated relationship between nutrient and density in that model:

“the magnitude of the eddy-driven flux is sensitive to the nitrate concentrations at the base of the euphotic zone, which are largely controlled by remineralization. This process is implemented very differently in the two models. Remineralization is computed prognostically in the NPZD model in Oschlies’s [2002b] study: That approach turns out to yield correlations between nitrate and density that are less than observed just below the euphotic zone. In contrast, the present model restores nitrate to climatology as a function of density and therefore preserves the observed correlations. Clearly, the issue of how to handle remineralization in such models is an important avenue for future research.”