Search

As part of the French CLIPPER project, an eddy permitting model of the Atlantic circulation has been run for 22 years. The domain has open boundaries at Drake passage and at 30 degreesE, from Africa to Antarctica. The simulated mean circulation, as well as the eddy activity, is satisfactory for a 1/3 degrees model resolution, and the meridional heat transport at 30 degreesS is within the range estimated from observations. We use the "mixed" open boundary algorithm of Barnier et al. [1998], which has both a radiation condition and a relaxation to climatology. The climatological boundary forcing strongly constrains the solution in the whole domain. The model heat balance adjusts through the surface (heat flux retroaction term) more than the open boundaries. The radiation phase velocities calculated within the algorithm are analyzed. This shows, quite surprisingly, that both the eastern and western boundaries have a similar behavior, regardless of the preferred directions for advection (mainly eastward) and wave propagation (mainly westward). Our results confirm that open boundary algorithms behave differently according to the dynamics of the region considered. The passive boundary condition that Penduff et al. [2000] applied successfully in the north eastern Atlantic does not work in the present South Atlantic model. We emphasize the need for a careful prescription of the climatology at the open boundary, for which a new approach based on synoptic sections is implemented.

As part of the French CLIPPER project, an eddy permitting model of the Atlantic circulation has been run for 22 years. The domain has open boundaries at Drake passage and at 30 degreesE, from Africa to Antarctica. The simulated mean circulation, as well as the eddy activity, is satisfactory for a 1/3 degrees model resolution, and the meridional heat transport at 30 degreesS is within the range estimated from observations. We use the "mixed" open boundary algorithm of Barnier et al. [1998], which has both a radiation condition and a relaxation to climatology. The climatological boundary forcing strongly constrains the solution in the whole domain. The model heat balance adjusts through the surface (heat flux retroaction term) more than the open boundaries. The radiation phase velocities calculated within the algorithm are analyzed. This shows, quite surprisingly, that both the eastern and western boundaries have a similar behavior, regardless of the preferred directions for advection (mainly eastward) and wave propagation (mainly westward). Our results confirm that open boundary algorithms behave differently according to the dynamics of the region considered. The passive boundary condition that Penduff et al. [2000] applied successfully in the north eastern Atlantic does not work in the present South Atlantic model. We emphasize the need for a careful prescription of the climatology at the open boundary, for which a new approach based on synoptic sections is implemented.

As part of the French CLIPPER project, an eddy permitting model of the Atlantic circulation has been run for 22 years. The domain has open boundaries at Drake passage and at 30 degreesE, from Africa to Antarctica. The simulated mean circulation, as well as the eddy activity, is satisfactory for a 1/3 degrees model resolution, and the meridional heat transport at 30 degreesS is within the range estimated from observations. We use the "mixed" open boundary algorithm of Barnier et al. [1998], which has both a radiation condition and a relaxation to climatology. The climatological boundary forcing strongly constrains the solution in the whole domain. The model heat balance adjusts through the surface (heat flux retroaction term) more than the open boundaries. The radiation phase velocities calculated within the algorithm are analyzed. This shows, quite surprisingly, that both the eastern and western boundaries have a similar behavior, regardless of the preferred directions for advection (mainly eastward) and wave propagation (mainly westward). Our results confirm that open boundary algorithms behave differently according to the dynamics of the region considered. The passive boundary condition that Penduff et al. [2000] applied successfully in the north eastern Atlantic does not work in the present South Atlantic model. We emphasize the need for a careful prescription of the climatology at the open boundary, for which a new approach based on synoptic sections is implemented.