The application of an evaporation duct model in operational and climatological assessments of propagation and the sensitivity of the model to meteorological measurements are examined. The unexpectedly frequent occurrence of evaporation duct heights greater than 40 m is related to thermally stable conditions in the atmospheric surface layer. The existence of stable conditions over the ocean is analyzed in terms of meteorological conditions and temperature measurement accuracies. Air-sea temperature differences measured by transiting ships are found to be biased toward stable conditions. The evaporation duct model is shown to be sensitive to this bias and yields high duct heights that are not supported by propagation measurements. A modified duct height calculation is proposed and applied to a radiometeorological data set.