Airborne Measurements of Surface, Layer Turbulence over the Ocean during Cold Air Outbreaks

Abstract

Airborne measurements of atmospheric turbulence spectra and cospectra made at the 50 m level above the western Atlantic Ocean during cold air outbreaks have been studied. The data cover nearshore areas of cloud streets or roll vortices. In the inertial submnge, the results are in good agreement with previous measurements made over both land and sea, except for those of temperature. The high-frequency behavior is consistent with local isotropy. In the inertial subrange, a near 4/3 ratio is observed between velocity spectra normal to and those along the aircrat~ heading. Also, the normalized spectra and cospectra of vertical velocity, humidity, and temperature are independent of sampling directions. The normalized cospectrum of the humidity flux appears to have a -4/3 cospectral slope, while the normalized cospectrum of stress appears to have -4/3 and -5/3 cospectral slopes in the alongwind and crosswind samples, respectively. The shapes of the spectra and cospectra vary with sampling direction. Except for crosswind velocity, the normalized spectra and cospectra appear to have more energy in the crosswind samples at the dimensionless frequency (f) ~ 0.2 and more in the alongwind samples at f< 0. I. This is mainly due to the stretching action ofthe mean wind shear on convective elements and is in good agreement with previous aircraft measurements. For the alongwind samples, the normalized velocity spectra for f< 0.1 appear not to be in good agreement with the spectral models derived from the cloud-free, highly convective Minnesota data. According to the latter, the normalized horizontal velocity spectra are nearly equal for low frequencies, while our results show that the low-frequency convection is significantly suppressed in the alongwind direction by the vertical wind shear. The three dissipation estimates, derived from the high-frequency part of the velocity spectra, appear to be in good agreement with a 9% mean discrepancy. The normalized dissipation is systematically smaller than those derived from the convective Kansas and Minnesota data. The turbulent kinetic energy budget is also different. For the same total energy production by wind shear and buoyancy, energy is available for exporting out of the surface layer in this case, whereas it must be imported into the Kansas and Minnesota surface layers.