Icebergs, Ice Shelves and Sea Ice: A ROMS Study of the Southwestern Ross Sea for 2001-2003

Michael S. Dinniman (1), John M. Klinck (1), and Walker O. Smith, Jr. (2)

(1) CCPO, Old Dominion University, USA
(2) Virginia Institute of Marine Science, College of William and Mary


Satellite imagery shows that there was substantial variability in the sea ice extent in the Ross Sea during the period 2001-2003. Much of this recent variability is thought to be due to several large icebergs that have calved of the Ross Ice Shelf and moved through the area. In March 2000 the iceberg B-15 was formed and a large fragment of this, B-15A grounded near the east end of Ross Island in January 2001 and did not begin to break up until October 2003. Throughout this time there were substantial changes observed in the ice concentration and distribution in the McMurdo Sound area. During the austral summer of 2002-2003 there was also unusually high ice cover in much of the Ross Sea due to the restricted advection of the ice by the presence of iceberg C-19. It would be difficult to accurately simulate the highly variable sea ice concentrations from 2001-2003 with just a dynamic sea ice model since much of the variability was due to the effects from the floating icebergs in the area. Here, a model is used where the sea ice and icebergs are imposed as boundary conditions in order to study the ocean conditions.

A high resolution (5 km) ROMS model of the Ross Sea was created which included the Ross Ice Shelf and simulates the mechanical and thermodynamic effects of the Ice Shelf on the water in the cavity underneath. After a several year spin up period, the circulation model is forced with daily winds for the period September 2001 - September 2003, as well as heat and salt fluxes calculated from atmospheric climatologies by bulk formulae. Instead of a dynamic sea ice model, sea ice concentration is specified from satellite observations which then modify heat and salt fluxes through the melting and freezing cycle. To examine the effects of the extra sea ice due to C-19, simulations were performed using either the climatological ice concentrations or the observed ice for that particular time period. Another simulation was performed where a fixed iceberg (representing B-15A which was nearly stationary throughout this period) was placed in the model.

The area-average annual basal melt rate beneath the Ross Ice Shelf in the simulation with the climatological ice cover (12 cm/yr) is within the range of estimates (12 - 22 cm/yr) from observations. The simulation with the observed sea ice reduces the annual mean melting by about 1 cm/yr. There is reduced advection of warm surface water during austral summer from the Ross Sea polynya into McMurdo Sound in the B-15A simulation, but this effect is small in the austral summer of 2002-2003 since the surface water in the polynya was much cooler than typical because of the extra sea ice due to C-19.