The Inversion Of Equivalent Sound Speed Profiles

At the present time, it is a very hot field to inverse the marine environment parameters by using acoustic methods in the world. The great progress has been made in the past 40 years with many scientists' efforts around the world. The focus of work in this field is how to obtain the parameters exactly and fast. Any improvement in this area will help people understand and monitor the oceans simply and quickly.The real ocean acoustic propagation environment is complex and variable, especially the perturbation of sound speed profiles have huge effects on the spread of the acoustic signal. Compared with the ray theory, the normal wave theory is more useful to the real ocean environment. In this paper, the equivalent sound speed profiles or temperature, salinity profile are inversed by using the matched mode processing method combined with empirical orthogonal function method, empirical mode decomposition method and genetic algorithm, basing on the sophisticated normal wave theory. The numerical simulation are carried out to verify the reliability of the inversed scheme by using the observation in the Luzon Strait and the Yellow Sea. Finally, the equivalent inverse temperature profiles are inversed by using real acoustic experimental data.After the simple description of the inversed method, the basic scheme of genetic algorithm are modified to improve its performance, such as that the real code method are applied to improve accuracy, the best solution among each generation are kept to speed the convergence velocity, the similarity threshold of the individuals are set to ensure the population diversity. The improved genetic algorithm is more stable than the basic algorithm from the numerical simulation.The monthly average sound speed profiles in 1991 and 1997 in the Luzon Strait are inversed numerically, due to the scarcity of the historical observation, the inversion error are large relatively among the upper layer. The empirical mode decomposition method is performed before the empirical orthogonal function analysis, in order to remedy the lack of the observation and pick up the information of temperature and salinity structure effectively. The effects of normal mode information on the sound speed profile inversion are discussed briefly using the data of Yellow Sea numerically.By considering the real acoustic observation, the subsection-equivalent method is adopted to inverse the sound speed profile when the topography is variable and the range-dependency of temperature field is strong. The segmentation location is also inversed simultaneously by applying the genetic algorithm. The agreement between the inversion and observed sound speed profile are very good.