Microwave techniques, such as the Scanning Multichannel Microwave
Radiometer (SMMR) and the Special Sensor Microwave Imager (SSMI), offer
the possibility of global monitoring of the latent heat flux, albeit
via a roundabout method. The SMMR measures the column integrated
water vapor (precipitable water), wind speed and sea surface
temperature, and the SASS provides surface wind (or stress)
vectors. Katsaros and Brown [1991] indicate accuracies of
2 - 3 kg m
, 
2 m s
, 
2 K, respectively,
based on the analysis of Seasat satellite SMMR data. The
integrated water vapor measurement is the most reliable,
comparable to estimates obtained from radiosonde ascents [
Katsaros et al., 1981]. Estimates of the surface fluxes,
however, require surface humidity. Liu [1986] derived a
relationship between the monthly men precipitable water and
specific humidity so that monthly mean bulk latent heat fluxes
can be derived on a global scale using precipitable water as a
proxy for specific humidity. The SSMI offers improved estimates
of wind speed and precipitable water, but lacks the low frequency
channels that are sensitive to variations in sea surface
temperature. Uncertainties remain; for example, it is difficult
to assess the affect of convergence of water vapor where the
surface air may be nearly saturated, but the total integrated
water vapor may be small [ Liu et al., 1991]. The frequent
very low wind speeds, highly variable sea surface temperatures
and specific humidity observed in the western Pacific Ocean using
in situ measurements pushes to the limit the ability to retrieve
fluxes by remote sensing. In these cases the integrated water
vapor does not reflect the vertical and horizontal humidity
variations that largely determine the surface latent heat flux.
Boundary layer processes cannot be readily determined by remote
sensing and more effort is required to understand the limitations
of these methods. The most attractive approach is the
combination of satellite measured parameters and numerical models
that contain sufficient physics to reduce most of the ambiguities
associated with the vertical integration and large spatial
averaging of the various parameters [ Atlas et al., 1987;
Atlas and Bloom, 1989].
Perhaps the most dramatic and readily documented large scale ocean-atmosphere interaction is the ENSO phenomenon. Large changes in sea surface temperature fields, water vapor, clouds, wind stress, precipitation and sea level can be inferred from satellites [ Njoku and Brown, 1993; Liu, 1988; Atlas and Bloom 1989; Arkin and Meisner, 1987]. This has lead to a better understanding of the air-sea interaction processes that may trigger the onset of an ENSO event [ Hirst and Lau, 1990].