Whitecaps are the main sink of wave energy and their occurrence has been related to the steepness of the waves. Recent parameterizations of the wave dissipation in numerical models are based on this property, but wave models have seldom been verified in terms of whitecap properties. Here we analyze and adjust the breaking statistics used in two recent wave dissipation parameterizations implemented in the spectral wave model WAVEWATCH III (R) and now used operationaly at NOAA/NCEP. For dominant breaking waves, the reduction of breaking probabilities with wave age is well reproduced. Across the spectrum, the parameterizations produce a reasonable distribution of breaking fronts for wave frequencies up to three times the dominant frequency, but fail to reproduce the observed reduction in breaking front lengths for the shorter waves. Converted to whitecap coverage, the breaking parameterizations agree reasonably well with the classical empirical fits of whitecap coverage against wind speed and the global whitecap coverage estimated from space-borne radiometry.