During the 1984 Marginal Ice Zone Experiment, a coordinated ocean boundary layer and mixed layer experiment was carried out in concert with a complete heat and mass balance study in the summer marginal ice zone (MIZ) of the Greenland Sea. The measurements were made at two drifting ice stations and included observations of ocean boundary layer turbulence, profiles of water temperature, salinity, and velocity, ice ablation, ice concentration, solar radiation, and spectral albedos. Ocean conditions were found to be extremely variable. The ice and water velocities were heavily influenced by tidal and inertial motions with amplitudes from 6 to 12 cm s-1. In the course of the drift, a 30-km eddy was crossed wherein ice and water velocities exceeded 15 cm s-1. The northern extension of the East Greenland Polar Front was encountered, and a northward flow of Atlantic Water into the Arctic basin was observed below the eastern edge of the front, but the drift station never moved quite far enough west to be caught in the southward transport of Arctic Surface Water. Heat from the ocean was a major factor in ice melt, and when the ice was over Atlantic Water it was the dominant factor. The events that had the greatest impact on the ice were two ''outbreaks'' during which off-ice winds blew the ice across warm ice edge ocean fronts.

Once south of these fronts and over surface water as warm as 1¿C, the ice bottom melted at rates up to 100 kg m-2 d-1. The rapid changes in stratification associated with the outbreaks had a dramatic impact of the drag and heat transfer coefficients. The drag coefficient (relative to the square of the relative water velocity at 30 m) ranged from 0.002 to 0.02 and averaged 0.006. The heat transfer coefficient (relative to the friction velocity μ* and the mixed layer temperature elevation above freezing) ranged from 0.0025 to 0.009 and averaged 0.0038. In agreement with the idea that stratification inhibits turbulent exchange, the heat transfer coefficient decreased when stratification increased. In fact, the observed melt rates can be simulated well if molecular heat and salt transfer through a thin, near-surface, laminar boundary layer are accounted for, as well as transfer through the turbulent outer boundary layer. Also in agreement with turbulent boundary layer theory, the drag coefficient decreased when stratification increased slightly above neutral. However, for the large increases in surface stratification, the drag coefficient increased. This was due to the generation of internal gravity waves by ice bottom roughness. Because the surface exchange coefficients are so sensitive to surface stratification, which is in turn highly variable in the MIZ, accurate simulations of ice motion and ice melt will require realistic estimates of stratification and a good understanding of the relation between the ice-ocean interactions and upper ocean conditions. ¿ American Geophysical Union 1987