A sensitivity study of weather data inaccuracies on evaporation duct height algorithms

The sensitivity of five evaporation duct height algorithms to errors in shipboard meteorological data is examined, and the algorithms are intercompared. The data set used was generated parametrically so a large variation of environmental conditions could be considered. The errors in the meteorological data consisted of two types: uncorrelated random errors associated with sensor inaccuracies and both random and systematic errors due to the influence of ship‐induced distortions. Before considering any errors, however, algorithm‐to‐algorithm differences of 10%–70% in the computed duct height are demonstrated and related to the assumptions and simplifications used during algorithm development. The sensitivity study shows that although the evaporation duct height algorithms have different genealogies, they have similar sensitivities. The sensitivities to errors caused by sensor inaccuracies range from 10% to 50% relative uncertainty in the calculated duct height, except for extremely low duct heights where the uncertainties are greater. During the daytime the relative uncertainties due to errors caused by ship‐induced distortions are approximately 10%–20% higher than those due to the sensor errors, and they are about 50%–75% smaller at night because of a lack of solar heating of the ship. These conclusions represent a best‐case scenario for Navy operational applications because of the omission of some sources of error, the optimistic ship error characteristics used, and the assumption of horizontal homogeneity in the near‐surface refractivity field.