ON THE POTENTIAL ROLE OF SEDIMENTS IN ARCTIC SLOPE-C0NVECTION
Jochen Kampf*, Hermann Fohrmann**, and Jan O. Backhaus*
*Institut fur Meereskunde der Universitat Hamburg, Hamburg, Germany
**Sonderforschungsbereich 313, Christian-Albrechts-Universitat zu Kiel, Kiel, Germany
1. INTR0DUCTI0N
The role of slope convection, or cascading, of dense shelf bottom water in formation of water masses in the Arctic Ocean is well accepted and explains properties of deep and intermediate waters (Backhaus et al., 1997).
The potential role of sediments in water mass formation, however, has as yet not been considered in much detail although substantial amounts of recent sediments of either glacial or fluvial origin, or from ice melting are available in the Arctic Ocean. Slope convection, enhanced by suspended sediments, therefore, appears to be an issue worth further investigation.
Process studies with a sediment-plume model (Fohrmann, 1996) in which typical water masses were prescribed show that sediments which are initially suspended in a gravity plume, or later due to erosion on its descent, may account for a larger negative buoyancy in comparison to classical water mass plumes. Their dynamics may become largely ageostrophic allowing for a very rapid descent, a deeper penetration and a different entrainment rate of water masses. Localized areas of high accumulation of recent sediments at the base of the continental slope and intermediate and bottom nepheloid (high attenuation) layers in the Norwegian Sea are considered as evidence of sediment plumes (Fohrmann et al., Sediments in bottom arrested gravity plumes - numerical case studies, submitted to JPO, 1997). Nepheloid layers were also observed in the vicinity of continental slopes of the Arctic Ocean. Once the sediment-plume has reached its equilibrium density horizon reduced turbulence allows for a setting of sediments. The plume water masses, void of the sediment load, may then become lighter than ambient waters and initiate upward directed convection in the water column originating from its intrusion level. This was already observed in the tropics (Quadfasel et al,, 1990) and later confirmed by laboratory experiments (Kerr, 1991).
An idealized model experiment on a lateral, sediment-laden intrusion from a slope was conducted with a non-hydrostatic numerical model (Kampf and Backhaus, 1998). This model was applied for investigation of both sediment plumes and sedimentation-driven convection.
2. M0DEL DESCRIPTION
The nonhydrostatic model is defined in a vertical plane. It is based on nonlinear Boussinesq equations for an incompressible fluid and contains prognostic conservation equations for both salinity and sediment concentration. Coriolis forces are neglected, assuming that sediment- driven bottom boundary currents are in first order approximation ageostrophic. A simple equation of state is assumed where the bulk density depends on both salinity and sediment concentration, For simplicity the effect of temperature on density is neglected. Simple sediment dynamics are being considered, consisting of setting, deposition and erosion of sediments. Only the sediment fraction of silt with a typical grain diameter of 20 μm is used since (a) larger particles are deposited very quickly because of their weight, and (b) smaller particles develop into larger aggregates. The plume-induced turbulence near the sea floor determines whether sediments are deposited or existing bottom material is eroded. For simplicity we assume that sediments are eroded when the bottom currents exceed the critical speed of u* = 10 cm/s. The settling velocity of sediment particles is determined from the relation w = w* {1-min (1,u/u*)} where u is the plume speed and w* =-1.82 mm/s.
3. M0DEL SETUP
The model domain contains a 100 m shallow shelf, a continental slope with an inclination angle of 1.9゜,and an abyssal plane with a water depth of 1000 m (Figure 1a). Initially, beneath a 100 m thick well-mixed surface layer, density increases linearly with depth (Figure 1b). On the shelf a fixed sediment bottom layer of 0.4 cm in thickness is prescribed, being mixed throughout the shelf water column through a fictitious storm event. There are no further sediments available along the continental slope. The numerical grid has a size of 500 m x 20 m. The simulation time is 10 days, resolved by a time step of 15 minutes.
* Corresponding authors addresses: Jochen Kampf and Jan O. Backhaus, Institut fur Meereskunde der Universitat Hamburg, Troplowitzstr. 7, D.22529 Hambur9, Germany, e-mail: kaempf@dkrz.de, backhaus@dkrz.de;
Hermann Fohrmann, Sonderforschungsbereich (SFB) 313, Christian-Albrechts.Universitat zu Kiel, Heinrich-Hecht-Platz 10, D-24118 Kiel, Germany, e-mail: hermann@sfb313.uni-kiel.de
