All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 107 7 0
PDF Downloads 16 4 1

Dynamical Forecast Experiments with a Barotropic Open Ocean Model

A. R. RobinsonDivision of Applied Sciences, Harvard University, Cambridge, MA 02138

Search for other papers by A. R. Robinson in
Current site
Google Scholar
PubMed
Close
and
D. B. HaidvogelDepartment of Physical Oceanography, Woods Hole Oceanographic Institution, Woods Hole, MA 02543

Search for other papers by D. B. Haidvogel in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The initial/boundary value problem for the barotropic version of a quasi-geostrophic open ocean model which requires normal flow everywhere on the boundary and vorticity on the inflow is studied. Parameter dependencies and sensitivities are determined for dynamical forecast experiments carried out over a 500 square kilometer domain with data simulated to represent the mid-ocean eddy field at 1500 m. The computational rms forecast error due to open boundary conditions is kept to 5% after one year of integration. Errors, gaps and noise are then introduced into the boundary and initial condition data. Objective analysis is introduced for mapping coarsely-distributed data onto the computational grid, and vorticity is derived from the streamfunction by several methods. Forecast error is sensitive to the frequency of updating of boundary data, but generally insensitive to vorticity errors. A simulated forecast experiment with composite error sources representative of feasible oceanic conditions is carried out for four months duration with rms error maintained to about 10%.

Abstract

The initial/boundary value problem for the barotropic version of a quasi-geostrophic open ocean model which requires normal flow everywhere on the boundary and vorticity on the inflow is studied. Parameter dependencies and sensitivities are determined for dynamical forecast experiments carried out over a 500 square kilometer domain with data simulated to represent the mid-ocean eddy field at 1500 m. The computational rms forecast error due to open boundary conditions is kept to 5% after one year of integration. Errors, gaps and noise are then introduced into the boundary and initial condition data. Objective analysis is introduced for mapping coarsely-distributed data onto the computational grid, and vorticity is derived from the streamfunction by several methods. Forecast error is sensitive to the frequency of updating of boundary data, but generally insensitive to vorticity errors. A simulated forecast experiment with composite error sources representative of feasible oceanic conditions is carried out for four months duration with rms error maintained to about 10%.

Save