Abstract
The mesoscale variability of surface heat fluxes induced by atmospheric convection is studied by using 3D cloud explicit simulations and surface observations. Two convective cases observed during the Coupled Ocean–Atmosphere Response Experiment are simulated (26 November 1992 and 17 February 1993) corresponding to different ambient surface wind conditions, namely, light and moderate winds.
Numerical results in the first case are successfully compared to surface observations. Local enhancements of two times for the latent heat flux and three times more for the sensible one are produced in the rainy area. Intense wind gusts generated by convective outflow are found mainly responsible for these increases. For the second case, the simulated surface fluxes are found to vary greatly, although they are structured in response to an organized convective system.
At the domain scale (90 km × 90 km) corresponding to a general circulation model (GCM) grid box, it is shown that convective activity significantly enhances the averaged surface heat fluxes. This effect is important since the preconvective wind is weak. To compute these surface fluxes with a bulk formula using fields defined on the domain scale, special attention must be given to the determination of the mean wind speed.
In GCMs, gusts generated by downdrafts are subgrid scale and are hence unresolved. This study suggests that flux enhancement due to clouds may be parameterized in GCMs by extending to deep convection the gustiness correction previously proposed for free convection by other authors. Analysis of both model simulations and observed time series suggest that once convection increases above a rather small threshold value, gustiness saturates at about 3 m s−1, whereas surface air humidity varies only slightly. These are the major elements of a proposed new parameterization of evaporation from the tropical ocean.
