LONG-TERM GOALS. To understand the convective process in the ocean, develop numerical models to study it, and so improve the representation of convection in ocean circulation models. OBJECTIVES. To understand the relative importance of vertical mixing by convection and lateral advection by geostrophic eddies in setting the volume and properties of water-masses formed by intense winter storms blowing over the open ocean. The insights gained will inform and improve parametric representations of convection used in large-scale ocean models. APPROACH. We are interpreting data gathered in the ONR Labrador Sea Convection Experiment, making use of process models of convection and a high-resolution model of the Labrador Sea developed at MIT. We have developed non-hydrostatic, hydro-dynamical models exploiting parallel computational architectures and languages (see publications Marshall et al. [1997 a, b], Hill and Marshall , Adcroft et al. , Marshall et al. ) that are capable of explicitly resolving the convective scale in the ocean. The model has been used to study convection modified by rotation in idealized settings; these have led to much insight into the nature of the convective process, have provided input to the planning of the field program, and are helping in the interpretation of the measurements that have been taken.