How fast does ice melt from below?

• Published in: 67th Annual Meeting of the APS Division of Fluid Dynamics - Gallery of Fluid Motion
• Main Authors: Keitzl, T., Notz, D., Mellado, J.
• Format: Other/Unknown Material
• Language: English
• Published: APS Division of Fluid Dynamics GFM 2014
• Subjects: Sea ice
• Online Access: http://hdl.handle.net/11858/00-001M-0000-002A-C260-B; http://hdl.handle.net/21.11116/0000-000B-10F2-1

This study uses laboratory experiments and direct numerical simulation to investigate the turbulent water movement beneath the ice--ocean interface. An ice--ocean interface is found for up to 7% of the ocean surface and is therefore one important component of the climate system. Heat and salt fluxes across this interface are of fundamental interest to properties of both of its constituents. On the one hand, temperature and salinity affect the density of water masses. In general, colder and more saline waters are less buoyant and will sink. In that way heat and salt fluxes across the ice--ocean interface can drive the thermohaline circulation of the ocean. On the other hand, they determine ablation and freezing processes of the ice on top and hence its thickness. Observation and model predictions do not agree on the rate of ablation and relative uncertainties are easily of order one. This indicates that the details of heat and salt exchange at the ice--ocean interface are not well understood. We want to improve the predictability of sea-ice behaviour by investigating the turbulent motion of the water beneath the ice. To do so, we study an idealised ice--water system in the laboratory and with simulations. Here we present and compare data obtained from both. Qualitative and quantitative agreement allows to investigate the underlying physical mechanisms via the detailed flow structure of the simulations. With that we improve our understanding of how increasing ocean temperatures influence the ice--ocean heat flux and the sea-ice melting. As a result one can incorporate it more realistically into climate models.