Hybrid Turbulent Fluxes Evaluation

A gridded hybrid turbulent heat flux time series is created by applying the state-of-the-art COARE v3.0 bulk algorithm to state variables (sea surface temperature, winds relative to currents, air temperature, and air specific humidity) derived from either Numerical Weather Prediction (NWP) reanalysis (NCEP1, NCEP2, and ERA40) or satellite retrievals (QuikSCAT winds and TMI Microwave sea surface temperature (SST)). The most accurate source for each state variable was determined by comparing variables to tropical Pacific TAO buoy observations for years 2000 - 2001. The selected sources were: Quikscat for winds relative to currents, ERA40 for air temperature and specific humidity, and TRMM Microwave Imager Fusion product for SST. Sensitivities to errors in state variables were analyzed. Specific humidity errors contributed the most to errors in latent heat flux (LHF). Overall, the hybrid LHF product had a bias of -5.8 W m-2 and a standard deviation of difference of 16.2 W m-2, which is comparable to the accuracy of LHF derived from TAO measurements. [Download manuscript]

---

	Mean bias of 10-m relative wind speed (Ur ) from NCEP1, NCEP2, ERA40, and QuikSCAT relative to TAO at 38 buoys in the equatorial Pacific. The buoy at 9oN, 140oW is grouped with the 8oN buoys. Zonal ocean currents estimated from the altimeter are subtracted from TAO winds. Except for the higher TAO winds near the ITCZ (8oN), QuikSCAT winds are much better than either the NCEP or ERA40 winds. (Click on figure to enlarge.)
	Same as figure above, but for Latent Heat Flux. The hybrid heat flux product is the best in the eastern Pacific, but too high in the western Pacific and near the ITCZ. (Click on figure to enlarge.)