Sound-speed field tracking in a range-dependent shallow water environment by ensemble Kalman filtering
Carrière, O.; Hermand, J.-P.; Candy, J.V.; Le Gac, J.-C.; Rixen, M. (2007). Sound-speed field tracking in a range-dependent shallow water environment by ensemble Kalman filtering, in: Rixen, M. et al. Rapid Environmental Assessment (REA) - Coastal processes: challenges for monitoring and prediction, Lerici, 25-27 September 2007.
In: Rixen, M. et al. (2007). Rapid Environmental Assessment (REA) - Coastal processes: challenges for monitoring and prediction, Lerici, 25-27 September 2007. NATO Undersea Research Centre: [s.l.]. , meer
|
| Auteurs | | Top |
- Carrière, O., meer
- Hermand, J.-P., meer
- Candy, J.V.
|
- Le Gac, J.-C.
- Rixen, M., meer
|
|
| Abstract |
In the framework of the European Seas Observatory Network (ESONET) a research objective is to couple physical oceanography and ocean acoustics into a data inversion and assimilation scheme for monitoring dynamic processes in shallow waters at multiple scales. In this paper the use of ensemble Kalman filtering scheme is investigated for tracking the time variations of a sound-speed field in a vertical section of a shallow water environment, taking into account the seafloor and sub-seafloor acoustic properties. The measurement setup consists of a broadband, multi-frequency sound source and a vertical receiving array spanning a significant part of the water column. In this preliminary study the state-space variables represent the main features of the sound-speed profiles in a low dimensional parameterization scheme, i.e., thickness of mix layer and thermocline vs. range. To test the algorithm full-field acoustic data are synthesized from multi-model super ensemble ocean predictions obtained in support of the BP07 MREA experiment southeast of Elba, Italy. Bottom geoacoustic properties obtained from previous geoacoustic inversion are input to the propagation model as a background dataset. Sea-surface temperature (SST) data from satellite and in-situ CTD observations provide a priori rough information about the rangedependent subsurface structure and an estimation of the sea-surface sound speed. The evolution of the entire sound-speed field is then sequentially estimated in an ensemble Kalman filtering scheme to handle correctly the non-linearity of the inverse problem. |
|
