Scale dependence of air-sea fluxes over the western equatorial Pacific

The goal of this study is to examine the horizontal scale dependence of vertical eddy flux in the tropical marine surface boundary layer and how this scale dependence of flux relates to the bulk aerodynamic relationship and the parameterization of subgrid-scale flux. The fluxes of heat, moisture, and momentum are computed from data collected from 27 NCAR Electra flight legs in TOGA COARE (The Tropical Ocean Global Atmosphere Coupled Ocean-Atmosphere Response Experiment) with flight elevations lower than 40 m and flight runs longer than 60 km. The dependence of the fluxes on two length scales are studied: the cutoff length scale, defining the averaging length over which mean components are obtained in order to partition field variables into mean and perturbation components; and the flux averaging length scale, defining the length over which products of perturbations are averaged in order to estimate vertical fluxes. Based on the characteristics of the scale dependence of fluxes, the total flux of each flight leg is partitioned into "turbulent," "large eddy," and "mesoscale" fluxes due to motions smaller than 1 km, between 1 and 5 km, and between 5 km and the flight leg length, respectively. The results show that fluxes are sensitive to the choice of cutoff length scale in the presence of significant mesoscale activity and in the weak wind case where the turbulent fluxes are small. The turbulent momentum flux decreases with increasing flux averaging length scale due to mesoscale modulation of the turbulent stress vector. Mesoscale heat, moisture, and momentum fluxes for individual flight legs can reach 20% of the turbulent fluxes in the presence of well-organized convective cloud systems even at 35 m above the sea surface. The mesoscale flux is less correlated to the wind speed and bulk air-sea difference than turbulent fluxes. The local mesoscale flux can be upward or downward, and therefore, its average value is reduced when averaging over a single flight leg and reduced further when compositing over all of the legs. The mesoscale momentum flux is less systematic than the turbulent stress and is more sensitive to the flux averaging scale than the turbulent stress. Sampling and instrumentation problems are briefly discussed, particularly with respect to mesoscale motions.