Deep-sea sound-speed profile estimation and prediction of sound propagation based on sparse depth sensing using towed temperature-depth sensors

Aiming at the problem of obtaining the sound speed profile in the deep sea, a method for orthogonal basis function extraction of deep-sea sound speed profiles is proposed based on the deep-sea hydrodynamic normal modes, which is able to reconstruct the sound speed profiles with less sample data compared to the empirical orthogonal functions. In addition, by combining the extracted orthogonal basis functions of the sound speed profiles with a limited number of temperature-depth measurements from the towed temperature-depth sensors, the spatial non-uniform sound speed profiles are estimated, which reduces the cost of measurements of the sound speed profiles. The method was validated by simulating the horizontal non-uniform variation of the deep-sea upper temperature profile. The sea trial data show that the average error between the reconstructed results of the sound speed profile and the measured results from the expendable bathythermographs is about 1 m/s, and that the deep-sea sound propagation losses predicted using the reconstructed results are in good agreement with the experimental data.