An Optimization Method for Sound Speed Profile Inversion Using Empirical Orthogonal Function Analysis

Abstract

In this paper, the empirical orthogonal function (EOF) analysis of historical Argo data and sea surface parameters were used to invert the sound speed profile (SSP), the inversion results cannot restore the details of the SSP, so an optimization method was proposed to improve the inversion results by using historical water temperature. Set the Argo SSP as reference, the root mean square error (RMSE) of inversion SSP is 1.4502, and the RMSE of optimized inversion SSP is 0.6302. It shows that the optimized inversion SSP is closer to the actual results than the non-optimized ones. In order to furtherly verify the effect of the optimized inversion SSP, the Bellhop model were adopted to calculate the acoustic propagation characteristics under three kinds of SSP (Argo SSP, non-optimized inversion SSP, optimized inversion SSP). The comparing results show that the optimized inversion SSP can accurately reflect the acoustic channel feature better than the inversion SSP.