Role of convection in winter mixed layer formation in the Gulf of Maine, February 1987

Moored hourly observations of temperature and salinity from the Wilkinson Basin in the western Gulf of Maine, augmented with gulf‐wide hydrography surveys, document the winter mixed layer evolution between 4 and 18 February, 1987. Wind stresses and air‐sea heat fluxes were estimated from the Gulf of Maine (44005) National Data Buoy Center buoy winds and temperatures using bulk formulae. During the study period, a pair of strong cooling episodes due to offshore (southeastward) winds bracketed an even stronger cooling event due to a strong northeasterly storm. The 0–165 m water column cooled during the study period. Heat budgets based on the observations show that local air‐sea heat loss could only explain cooling in the upper 35–60 m of the Wilkinson water column. Lateral advection must have caused the deeper cooling. The Price, Weller and Pinkel [PWP] one‐dimensional bulk mixed layer model [ Price et al. , 1986], forced with observed surface fluxes, exhibited a mixed layer that deepened from about 60 m to 120 m during the initial stages of the nor'easter in general accordance with observation. Diagnostic PWP model experiments showed that convective overturning was required to produce the observed mixed layer depth of 120 m. Direct wind mixing alone could deepen the model mixed layer to only about 80 m. However, the model property profiles produced by a 10‐day PWP model run differed significantly from observations, which reflected advection effects. Observed hydrographic property distributions combined with horizontal velocity estimates from a Dartmouth linear, three‐dimensional, finite‐element circulation model of the Gulf (FUNDY5) help to explain differences between the PWP model and observations in terms of lateral advection in the upper 65 m. The observed cooling between 65 m and 165 m must be related to advection.