Modelling the Oceanic Nd Isotopic Composition With a North Atlantic Eddy Permitting Model

The oceanic water masses differ by their temperatures, salinity, but also a number of geochemical tracers characterized by their weak concentrations and their ability to quantify oceanic processes (mixing, scavenging rates etc). Among these tracers, the Nd isotopic composition (hereafter epsilon-Nd) is a (quasi) conservative tracer of water mass mixing in the ocean interior, far from any lithogenic inputs. It has been recently established that exchange of Nd at the oceanic margins could be the dominant process controlling both its concentration and isotopic composition distribution in the ocean. This was demonstrated using in situ measurements and budget calculations and has recently been confirmed by a low resolution (2°) modeling approach (Arsouze et al., 2006). However, the currents flowing on the ocean margins are not correctly represented in coarse ocean models. It is the case in the North Atlantic ocean, which is of particular interest since i) it is the area of deep water formation and ii) these deep waters are characterized by the most negative epsilon-Nd values of the world ocean, which are used as "imprint" of the present and past thermohaline circulation. It is therefore essential to understand how these water masses acquire their epsilon-Nd signature. We propose here the first results of the modeling of oceanic Nd isotopic composition at eddy-permitting resolution, with the North Atlantic 0.25° version of the NEMO model used for the DRAKKAR European project. A 150 years off-line experiment and a shorter on-line experiment are performed. Simulated Nd distributions are compared to the present-day data base, vertical profiles, and the results of the low resolution model (in the North Atlantic). The eddy permitting model generally provides improved results, provided a high enough exchange rate is imposed in the deep ocean. Deficiencies of the simulated distribution in the Nordic Seas and the subpolar gyre are explained by errors in the input function on the margins of the Iceland-Scotland Ridge and south of Iceland. The model does not account for all the complexity of the Nd profiles observed downstream of Denmark Strait and in the Labrador Sea. This is partly due to the representation of the deep western boundary currents, still marginally resolved in this eddy-permitting model. However, these promising results with an eddy permitting model pointed out the deficiencies of the parameterization of the input function, and confirm the potential of epsilon-Nd as a powerful dynamic tracer, already presumed with the low resolution model. Reference: Arsouze, T., Dutay, J.-C., Lacan, F. and Jeandel, C., 2006. Modeling the neodymium isotopic composition with a global ocean circulation model Chemical Geology, in press.