Title :
Geoacoustic profile estimation using empirical orthogonal functions for propagation applications
Author :
Jaye, Michael ; Badiey, Mohsen ; Siegmann, William L.
Author_Institution :
Dept. of Math. Sci., Rensselaer Polytech. Inst., Troy, NY, USA
fDate :
10/1/2001 12:00:00 AM
Abstract :
The use of empirical orthogonal functions (EOFs) for the direct estimation of sediment sound speed, density and attenuation profiles is described, when extensive data sets such as those from the Atlantic Generating Station (AGS) site are available. Interpolations are generated based on minimizing mean square errors and their construction is modified here to incorporate lateral terrain orientations as occur at the AGS site. Uncertainties in the geoacoustic fields can then be modeled by stochastic variations of the EOF coefficients. The field representations are used to produce geoacoustic input for propagation calculations for fluid or elastic sediment models in the AGS region. Statistics of transmission loss (TL) variations are discussed in terms of geoacoustic variability, source frequency and resolution provided by the number of interpolating EOFs. Preliminary results show that sediment sound speed uncertainties can produce a significant uncertainty envelope for TL predictions in the AGS area
Keywords :
interpolation; least mean squares methods; oceanographic techniques; seafloor phenomena; sediments; stochastic processes; underwater acoustic propagation; attenuation profiles; density profiles; empirical orthogonal functions; geoacoustic interpolation; geoacoustic profile estimation; lateral terrain orientations; minimizing mean square errors; propagation model; random uncertainties; sediment sound speed; shallow water; stochastic propagation; transmission loss statistics; Acoustic propagation; Attenuation; Frequency; Interpolation; Mean square error methods; Propagation losses; Sediments; Statistics; Stochastic processes; Uncertainty;
Journal_Title :
Oceanic Engineering, IEEE Journal of