Convection

Any good model of the Labrador Sea must accurately predict wintertime convection and the subsequent water mass transformation. The model convects in a small southwest corner of the Labrador Sea and is very sensitive of the initial T-S structure. Preliminary runs initialized with Levitus climatology did not reproduce convection with the right intensity or location. With the M. Visbeck climatology, mid-February marks the onset of deep convection in the model, which is within days of the observed process. The potential density of the newly-formed deep Labrador Sea water is within $.02 kg/m^{3}$ of the observed value. The model Labrador Sea water is at 2.75 degrees C, 34.85 salinity, while observations give values of 2.78 degrees C and 34.83 salinity. Labrador Sea water (LSW) has moved in a counter-clockwise path in T-S space over the last 70 years (Lab Sea Group, 1998), and the model LSW would unambiguously be categorized as a product of the 1990's. 1000-1500 meters of convection was observed in 1996-97; the mixed layer is approximately 1500 meters deep in a $200 km$ wide patch of the southwest interior Labrador Sea of the model. The quick restratification of the Labrador Sea after the convective season is also captured by the model, and will be discussed later in the study. Theory predicts maximum vertical velocities as large as 10 cm/s during convection (Jones and Marshall, 1993), but the largest values seen in the model are about 20 times smaller. This may be reasonable because convective plumes will not be explicitly resolved. Eddy kinetic energy shows a strong seasonal cycle with a dramatic increase in energy at all levels immediately after convection.