Frequency- and signal type dependence of the performance of broad-band geoacoustic inversion in a shallow water environment with soft sediments
Siemes, K.; Hermand, J.-P.; Snellen, M.; Simons, D.G. (2011). Frequency- and signal type dependence of the performance of broad-band geoacoustic inversion in a shallow water environment with soft sediments, in: Papadakis, J.S. et al. 4th international conference and exhibition on Underwater Acoustic Measurements: Technologies and Results. pp. 8 In: Papadakis, J.S.; Bjørnø, L. (2011). 4th international conference and exhibition on Underwater Acoustic Measurements: Technologies and Results. FORTH/IACM: Kos. , more |  |
| Available in | Authors | | Document type: Summary
|
| Keyword | | | Author keywords | geoacoustic inversion; broad-band; seabottom classification; frequency dependence |
| Authors | | Top | - Siemes, K., more
- Hermand, J.-P., more
- Snellen, M.
- Simons, D.G.
| | |
| Abstract | Geoacoustic inversion techniques are an attractive means for estimating physical properties of underwater environments. These techniques aim, at least partly, at a substitution of the costly methods of probing the seabottom by grab samples or cores. However, geoacoustic inversion comes at the price of high computational efforts. Especially, in cases in which large numbers of parameters need to be inverted for, finding the best fit between the measurements and a predicted model requires hundreds of iterations. Efficient global optimization tools exist that help reducing these efforts. One of these methods is the differential evolution method, which is employed in this paper. Beside the time needed for the optimization, another issue is the computational effort needed for establishing the forward model. It highly depends on the number and magnitude of frequencies employed. In general, high frequency calculations are more computational intensive. It is therefore investigated, for a given soft-layer bottom model, which frequencies are beneficial for the estimation of seabottom parameters and which frequencies only increase the computational time. Employed are frequencies in the bands of 300–800Hz (low-frequency) and 800–1600Hz (mid-frequency) for creating broad-band signals. Both, signals composed of tones at discrete frequencies (multi-tones) and frequency modulated waveforms (chirps) are compared. These signals are observed at a 4-element vertical line array. The measurements were performed during the Maritime Rapid Environmental Assessment / Blue Planet (MREA/BP'07) experiments [Le Gac & Hermand, 2007], which were carried out in the Mediterranean Sea in 2007, to address novel concepts of characterizing the continental shelf environment. The data originate from a shallow-water location, west of Italy and south-east of Elba Island, which is known to be composed of very fine grained sediments and an underlying silty clay bottom. |
|
