Feature-oriented acoustic tomography of a coastal thermal front

The continuous monitoring of coastal processes presents a great interest both from the environmental and economic viewpoints. Passive acoustic tomography can be a good candidate to provide synoptic measurements over wide areas while a range-dependent inversion scheme allows to achieve a reasonable spatial resolution along each vertical slice section. This work develops a feature-oriented parameterization scheme for acoustic tomography purposes, enabling the tracking of the main structure of a thermal front. A Kalman algorithm filters sequentially acoustic data recorded on a vertical array, in a frequency regime corresponding to a usefull part of ship noise spectrum.