Sensitivity of Cyclogenesis to Sea Surface Temperature in the Northwestern Atlantic

Abstract

During the Intensive Observation Period 15 (13?15 February 1997) of the FASTEX Experiment, a major cyclone crossed the Atlantic Ocean from the Newfoundland Basin to southern Iceland. Its surface low center deepened by 17 hPa in 7 h when the perturbation crossed the North Atlantic Current (NAC) from cold (3°C) to warm water (15°C). To elucidate the role of sea surface temperature (SST) and air?sea fluxes in the dynamics of oceanic cyclones, three nonhydrostatic mesoscale simulations were performed. The first one is a control experiment with a realistic SST field describing in detail the oceanic front associated with the NAC system. The two following simulations are sensitivity experiments where the SST front is removed: the first one uses a uniformly cold SST equal to 3°C and the second one uses a uniformly warm SST equal to 15°C. The frontogenetic function and the vertical velocity sources in the lower-atmospheric layers of the three simulations were diagnosed. In the control simulation, the surface heat fluxes were found to be negative in the perturbation warm sector and positive in the region behind the cold front. As reported by numerous authors, this pattern of surface heating and cooling did not intensify the cyclone, except in the occlusion when the phasing with the SST front occurs. This configuration enhances the horizontal gradient of surface buoyancy flux across the occlusion, which increases the buoyancy flux source of vertical velocity (w). When the SST front is removed, the surface heat fluxes are strongly affected in magnitude and in spatial variability. The marine atmospheric boundary layer (MABL) stability, the convective activity, the warm advection in the core of the wave, and the heating depth are strongly affected by the different surface flux fields. There are several consequences: (i) the uniform SSTs tend to decrease the cold front intensity of the wave, (ii) a weaker buoyancy flux source of vertical velocity is found above a uniform cold SST across the occlusion in comparison with the control case, and (iii) surprisingly, a weaker w buoyancy flux source is also found above a uniform warm SST because of a higher heating depth. Vertical velocity depends not only on the buoyancy flux forcing but also on the thermal wind, the turbulent momentum, and the thermal wind imbalance forcings. The thermal wind forcing and the thermal wind imbalance forcing were the most sensitive to the SST compared to the turbulent momentum forcing. This result means that (i) the feed back of the ageostrophic circulation induced by the surface is greater on the kinematic forcings than on the turbulent forcings and (ii) the turbulent momentum forcing does not play a crucial role in cyclogenesis. Above a uniform warm SST, the strongest intensity of the occlusion is due to the strongest w thermal wind forcing and w thermal wind imbalance forcing in the MABL, in spite of a weaker w buoyancy flux forcing than in the control case. This result is explained by the convective activity that increases the low-level convergence and vorticity spinup. This point means that latent heat release and baroclinicity are in tight interaction. In the first 12 h and at the scale of the simulation domain, the three cyclones evolve similarly, but at a small scale their internal structures diverge strongly and rapidly. The scale at which the surface turbulent fluxes act on the dynamics of marine cyclones is therefore important. Finally, the cyclone simulated in the warm SST case developed more rapidly than those simulated in the real and the cold SST cases. This behavior is attributed to the strong positive surface heat fluxes because they preconditioned the MABL by moistening and heating the low levels during the incipient stage of the cyclone development.