A technique for predicting the expected number of shipping‐produced lines in an observed underwater acoustic ambient noise spectrum is described. The method is based on a mathematical model which includes the sonar equation, available information on the acoustic signatures of individual ships, and data concerning the oceanic distribution of shipping. The average number of detectable spectral lines in a given frequency band (line density) is shown to depend on the location of the sensor, sea state, and frequency. The line density also depends on the parameters of the sensor system. The results of calculations based on the model imply, for example, that the spectra observed with more directive receivers will contain more spectral lines. The results and methods described have application to passive sonar systems engineering.