Sound speed estimation from sparse environmental measurements with application to mode filtering

Acoustic mode filtering requires accurate knowledge of the mode shapes and wavenumbers at the receiver, which are functions of the sound speed profile. During a year-long deployment, environmental measurements, hence measurements of sound speed, often consist of data from a small set of temperature sensors mounted on the vertical receiving array. This paper describes a space-time Kalman filter framework for estimating the sound speed profile from a spatially-sparse set of temperature measurements. The approach assumes that the changes in sound speed associated with slow variations in the environment can be written in terms of the quasi-geostrophic dynamic modes. The paper analyzes the sound speed errors associated with the method and considers how those errors propagate through the mode computations. Examples using environmental measurements made during the North Pacific Acoustic Laboratory experiment are provided