Performance of techniques used to derive ocean surface winds

Various techniques used to derive analyses and forecasts of ocean surface winds were compared. These techniques are the geostrophic relation, a simple law, an Ekman slab model, NMC forecast model 1000 mb winds the Cardone (1969; model winds, the Clarke and Hess (1975) model winds, and the marine winds from Fleet Numerical Ocean Oceanographic Command (FNOC). Statistical comparisons of model winds with those from ships and buoys were made for wind speed, wind direction and the vector wind. The statistics suggest that none of the techniques was clearly better. The study did indicate that model wind speeds and inflow angles compare better with buoy data than ship data. For high wind speeds (>22.5 m/s) observed by ships, all models were too low. Overall, the Cardone model appears to produce slightly better verification results when both the analyses and 24 hour forecasts are compared with observations from the buoys. William H. Gemmill, Tsamn W. Yu, and David M. Feit. "April 1987." "This is an unreviewed manuscript, primarily intended for informal exchange of information." Also available online in PDF via the NOAA Central Library. Includes bibliographical references (pages 33-34).