This paper presents an objective assessment of four methods for estimating sea surface friction velocity u* from wind speed at height z, uz, and air-sea temperature difference. The methods are compared by using the computed friction velocity as the normalization factor in parametric correlations of wind-wave paramters show that (1) wind profile parameters obtained form the four methods are generally comparable, (2) parametric correlations with parameters normalized by uz lead to significantly decreased percentage deviations over correlations normalized by and (3) correlations based on u* derived from the four methods show nearly percentage deviations. The conclusions are that (1) u* normalization acts to eliminate the effect of atmospheric stability and (2) any of the four methods can be used effectively in practical applications. ¿American Geophysical Union 1987