Emery: discussed operational SST products, derived from infrared radiances. Water vapor and aerosol variations, as well as cloud clearing algorithms, remain as difficult problems in remote sensing of SST. Drifting buoys are used to calibrate SST, with buoys having precision of 0.1o C and accuracy of 0.3-0.5o C for measurements at 1-2 m depth. These errors are not accounted for in the calibration of the satellite SST products. Buoys are predominantly in the tropics and subtropics, with ships dominant at high northern latitudes. A key issue in using the buoys in calibrating the satellite-derived SST is the spatial representativeness of the buoys. Wick/Emery/Schleussel model for skin-bulk temperature difference was described, that depends on wind speed and net heat flux. During night, this relationship could be inverted so that known values of DT and wind speed could be used to obtain the net heat flux. Finally, a global map was described of monthly-average night/noon skin-bulk temperature difference, which was obtained from an ocean mixed layer model with skin layer submodel that was forced with ECMWF surface fluxes.
Wick: GOES multichannel retrievals are providing more accurate atmospheric correction. Bias errors in general for IR products are seen to vary over the diurnal cycle, which can seriously mess up any attempt to use these data to infer the diutnal cycle of SST. New TMI SST product (Frank Wentz) was described using low channel microwave observations (similar channels on AMSR, EOS PM, ADEOS 2). 10 GHz has a penetration depth of 1 mm. Microwave SST retreivals are affected by near surface wind and direction, but are insensitive to clouds, water vapor, aerosol. The TMI has limited space/time coverage, but this will improve with EOS PM, etc.
Clayson: Emphasized that skin SST is what is desired for turbulent flux calculations, although operational SST products provide bulk SST. Because infrared methods can't "see" beneath clouds, space/time sampling of IR-derived SST products is nowhere near adequate for determining the diurnal cycle. An indirect technique was described whereby predawn skin SST values are interpolated at a given location to provide daily values of predawn skin temperature. A parameterization for the diurnal amplitude of the skin SST is then applied, which is derived from an ocean mixed layer model and depends on peak daily insolation, average daily wind speed and precipitation. Using this model allows for very high time/space resolution in the skin SST field, even under cloudy conditions. This method has been extensively applied in the TOGA COARE region with successful results, and also has been successfully applied to the subtropics (ASTEX). Global application of the algorithm will require additional mixed-layer model simulations in other regions to test the validity of the algorithm (or derive alternative coefficients) in different climatic regimes.
Rossow: Emphasized the importance of diurnal variations in SST. Use of IR-derived SST values (from clear-sky situations) will give you a field where the errors in fluxes are are correlated with the meteorology: not random, but systematic errors. Without including the diurnal cycle, you can have errors of 1-3o C in the skin SST. Recommends physical retrieval rather than regression equation, so that effects of aerosols, etc are explicitly considered.
Taylor: Discussed the diurnal thermocline, and how skin-bulk relationships must be careful to account for which bulk temperature is being referred. States that most bulk flux algorithms use bulk SST, and that most drag coefficients are calibrated using bulk SST. Bulk temperature is alot less sensitive than skin SST, is a more stable parameter to analyze for flux calculations
Discussion:
Curry, Liu: emphasize that research quality flux algorithms all use skin temperature, which is the physical variable that the fluxes depend upon.
Xeng: Considerable uncertainties exist in the operational SST products, arising from water vapor, aerosols, clouds. Since the calibration of the satellite SST values is based on buoys, theere is a bias from sampling clear sky. Use of bulk instead of skin SST can even change the sign of the atmospheric surface layer stratification. Describes a new method for determining skin-bulk temperature difference as a function only of wind speed and the time series of bulk temperature. This is being used to infer the skin SST at the TAO buoys.
Discussion:
Emery: Is compiling in situ observations of skin SST on a CD. Also, has been proposing to make skin measurements on ships of opportunity
Rossow: To get the atmosphere to respond correctly in NWP models, need to put more energy into the mesoscale; including the diurnal cycle will help this.
Weller: Some high spatial variability of skin SST is observed in clear, low wind conditions in the tropics.
Liu: There is merit in re-examining the idea of obtaining nighttime net heat flux from measurements of skin-bulk difference.
