Performances of the ensemble Kalman filter on the tracking of time-evolving sound speed profiles

The performances of the Ensemble Kalman Filter (EnKF) on the tracking of time-evolving Sound Speed Profiles (SSPs) is studied.Firstly,based on the empirical orthogonal decomposition,the SSPs provided by the Princeton Ocean Model (POM) at the background of an actual experiment in the South China Sea are formulated as the form of state-space with the first three coefficients and the state equation is modeled as a three-orders auto-regress process.Then,according to the theory of Kalman filter,the dynamic tracking is implemented by the correction step where Range-dependent Acoustic Model (RAM) simulated acoustic filed is applied to correct the predicted state.Simulations in range-independent,range-dependent and geoacoustic parameters mismatched environments show good tracking results,which verifies the feasibility of the algorithm.In addition,the performances of EnKF with respect to Signal Noise Ratio (SNR),ensemble number,number of hydrophones and mismatched geoacoustic parameters are investigated.It shows that the increase of observations can efficiently reduce the effects of errors in both observing and modeling,which is confirmed by actual experiment data and provides an important reference for the realistic applications.