Direct measurements of the surface fluxes are limited to a few specialized platforms so extensive, global-scale measurements must rely on the parameterization of the surface fluxes that can use easily measured quantities obtained either in situ or remotely [e.g., Katsaros et al. 1994]. The basic premise is that we can relate fluxes in the surface layer to logarithmic profiles of the mean quantities. The surface fluxes can then be determined from the mean wind, temperature and humidity at a single height by introducing bulk transfer coefficients of heat, moisture and momentum. This rationale applies directly to gas exchange also.
A major limitation of many bulk parameterizations of the surface
fluxes is that they are not well defined in light wind conditions
where the sea surface becomes very smooth [ Bradley et al.,
1991; Miller et al., 1992]. Bradley et al. [1991]
compared several parameterization methods with observations they
obtained in relatively light winds in the western Pacific. They
noted generally very good agreement between the model of Liu
et al. [1979] and the observations below wind speeds of about 4
m s
. The observations indicate, however, that the latent heat
flux is nonzero even at zero wind speed, a factor not considered
in most existing parameterizations. To account for this effect
it is necessary to add a gustiness factor to the mean wind that
is proportional to the convective scaling velocity (a velocity
scale derived from the surface buoyancy flux and depth of the
mixed layer) (C. W. Fairall et al., The TOGA COARE bulk flux
algorithm submitted to J. Geophys. Res., [1994]). The
stability dependence of the profiles of temperature, moisture and
momentum in highly unstable conditions also should depart from
the traditional over land functions. Based on preliminary data
from COARE and the results of other tropical cruises, models are
being developed to include these effects (e.g., C. W. Fairall et
al., ibid.). Interpretation of surface temperature measurements
also remains a problem and has been addressed recently. Bulk
parameterizations of the fluxes require estimates of the skin
temperature of the ocean that are not routinely available, except
from a few infrared sensors flown on aircraft or mounted on
research ships. Space-borne measurements provide skin
temperatures, but uncertainties exist in the accuracy of the
retrievals primarily due to the large absorptivity of long wave
radiation by the high water vapor content of the atmosphere in
the tropics. In situ measurements rely on temperatures obtained
from about 2 m below the surface that must be corrected for the
effect of overlying water temperature and the evaporation at the
sea surface that produces the so-called cool skin effect (C. W.
Fairall et al., The cool skin and warm layer in bulk flux
calculations submitted to J. Geophys. Res., [1994]). Abrupt
temperature gradients close to the surface are observed frequently
in light wind conditions in the tropics where the daytime solar
flux approaches 400 W m
. Roughly half of this heat is absorbed
within the top two meters of the ocean. The results presented by
C. W. Fairall et al. (ibid.) indicate that a model of the surface
warm layer, ollowing Price et al., [1986], is essential to
properly determine the temperature at the surface of the ocean
and differences in excess of 2 K are possible between 2 m depth
and the surface. This temperature stratification indicates that
the sea surface temperature is very sensitive to wind-driven
mixing of the upper few meters of the ocean.
Improvements in the parameterization of fluxes are likely as a result of the very extensive sets of direct measurements obtained by ships and boundary layer aircraft over the open ocean in the past few years (e.g., SOFIA/ ASTEX and TOGA COARE) and from equivalent measurements in relatively shallow coastal waters that in addition to ships and aircraft utilize instrumented towers and fixed platforms (e.g., the Marine Boundary Layer (MBL) experiment, a recent initiative from the Office of Naval Research).