MODELING RIVER INPUT ALONG THE ARCTIC SHELVES
R. Newton1, D. Martinson1* , P. Schlosser1, W. Maslowski2, and A. Semtner2 1Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY
2Naval Post-Graduate School, Monterey, CA
1. INTRODUCTION
Shelf-derived plumes of cold, salty water are critical in the large-scale thermodynamics of the Arctic ocean. They feed the halocline through horizontal interleaving along isopycnals, and thereby modify and maintain the halocline which segregates the warm, salty Atlantic layer from the surface waters. If the halocline were to deteriorate, there is good reason to believe that the Arctic ice cover would be threatened through vertical mixing of heat from the Atlantic Layer, which in turn would drive significant climate impacts through the ice-albedo and ice-cloud feedbacks. To the extent that the plumes break past the surface and intermediate waters, they can be driven to great depths, where, on long timescales, they influence the properties of abyssal waters. Horizontal and temporal transients in these properties provide critical evidence about the history of large-scale patterns of convection and abyssal flow. In addition, recent measurements show that the horizontal density structure of the Arctic surface waters, below the ice cover, are strongly influenced by plumes of shelf-water which have been freshened by river runoff or relatively fresh inflow from the Bering Strait. The shelf waters have a strong influence over the strength and structure of the baroclinic component of the Arctic circulation. Unfortunately, direct measurements of plume sizes, volumes and densities are very incomplete. Even the points of separation of surface "streams" of fresh or dense water from the shelf are poorly documented.
The Arctic shelf break is an area of steep and complicated topography, strong along-shelf currents and significant tidal mixing. The associated complex dynamics make the explicit modeling of shelf-plumes very difficult. However, recent advances in tracer measurements for the Arctic are making it possible to track their influence far from the source regions. Oxygen 18/16 ratios can distinguish between sea-ice (with a delta-O-18 value very close to the ocean value) and precipitation, most of which enters the Arctic as river runoff (with a value of approximately -21). In addition, chemical tracers can further discriminate between fresh waters from the Canadian rivers (using barium), the Bering Strait (using silica and dissolved nutrients) and perhaps even between the Russian rivers (using industrial or agricultural pollutants).
We are engaged in a program of closely integrated tracer field studies and model experiments. Measurements of oxygen-isotope, tritium, helium, chloroflourocarbons (freons) and noble gases in the Arctic have been used, for example, to: (l) identify the source of freshwater pooling in the surface of the Beaufort Gyre (river runoff); (2) quantify changes in open water convection in the Greenland-Iceland-Norwegian seas; (3) measure water-mass gradients in the abyssal Arctic basins. In the model runs, we add several freshwater tracers to a coupled ice-ocean general circulation model in order to diagnose the model results against the tracers, and ultimately determine the mechanisms responsible for the tracer (and river) distribution and its sensitivity to changes.
2. MODEL AND EXPERIMENTS
The ocean model (Maslowski et al., 1997) is based on a free-surface version of the Parallel Ocean Program (Dukowicz and Smith, 1994), which originates from the Bryan/Cox/Semtner codes. This is coupled to, and forced by, an ice model (Zhang et al., 1997) based on Zhang,s implementation of Hibler,s plastic-viscous rheology (Zhang and Hibler, 1997) and Semtner,s 0-layer thermodynamics (Semtner, 1976). The atmospheric forcing is taken from the ECMWF re-analysis of its operational forecasts. The model grid offers good resolution (though not eddy-resolving) for a basin scale model; 18-km horizontal resolution and 30 vertical levels. For each river, long term measurements of flow and temperature are used to establish average monthly flow-rates and heat inputs (zero salinity is prescribed for all rivers). These are interpolated to the model time step (40 minutes), so that a source for freshwater and heat is introduced at the boundary. At the Bering Strait, three years of current, temperature
* Corresponding author address. Douglas G. Martinson, Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY 10964, USA; e-mail: dgm@ldeo.columbia.edu; FAX (914) 365.8736
