Measurements of atmospheric fluxes of heat, moisture, and momentum were made simultaneously and coincidentally with microwave backscatter measurements from an airship flown over the Pacific Ocean in 1993. Because the airship required an air speed near 10 m/s in order to maintain altitude, averaging times at low wind speeds could be shorter than usually required for a given accuracy. Thus the measurement technique was well suited to measuring fluxes at very low wind speeds. The measurements show that very low wind speeds are always associated with very low microwave cross sections and very high air/sea drag coefficients. The occurrence of regions of very low wind speed is not usually correlated with either the sea surface temperature or the air/sea temperature difference. Nevertheless, these regions can remain in place for time periods of several hours. The rate of increase of the microwave cross section at very low wind speeds agrees with that predicted by Donelan and Pierson (1987) but the absolute value of the threshold wind speed appears to be lower than their prediction. The high drag coefficient at low wind speeds is due to the fact that the friction velocity is nearly constant for wind speeds below 4 to 5 m/s. Thus at these wind speeds, the increase of the microwave cross section follows the behavior of the wind speed rather than the wind stress. At higher wind speeds, however, the behavior is reversed with the cross section following the wind stress at a constant wind speed. We suggest that this behavior can be understood if momentum transfer across the air/sea interface is supported by both viscosity and the entire spectrum of waves on the surface, as many investigators have indicated. The proportion of stress supported by surface waves in various wavenumber regions depends on the shape of the wave height variance spectrum as a function of wavenumber. At very low wind speeds, the short waves which scatter microwaves have small slopes so that they carry only a small part of the stress. Thus while they are small, their growth causes only small changes in the friction velocity, which are essentially undetectable; most of the stress is carried by longer waves and viscosity, neither of which responds strongly to the local wind. At higher wind speeds, however, the small waves support a larger part of the wind stress so that changes in their amplitude have easily detectable effects on the stress. If the high-wavenumber part of the spectrum maintains a constant shape as wind speed increases and supports most of the stress, then the microwave cross section must follow the stress in this wind speed regime.
Return to list of journal publications