HR: 1340h
AN: S23A-1109    [Abstracts]
TI: Resolution Analysis for Experiment Planning of a Nonlinear Seafloor Acoustic Inverse Problem
AU: * Ganse, A A
EM: aganse@apl.washington.edu
AF: Applied Physics Laboratory, 1013 NE 40th Street, Seattle, WA 98105, United States
AU: * Ganse, A A
EM: aganse@apl.washington.edu
AF: Department of Earth and Space Sciences, Box 351310, University of Washington, Seattle, WA 98195, United States
AU: Odom, R I
EM: odom@apl.washington.edu
AF: Applied Physics Laboratory, 1013 NE 40th Street, Seattle, WA 98105, United States
AU: Odom, R I
EM: odom@apl.washington.edu
AF: Department of Earth and Space Sciences, Box 351310, University of Washington, Seattle, WA 98195, United States
AB: Geoacoustic inversion is the estimation of physical properties of the ocean bottom as a continuous function of position (or depth) in the seafloor given acoustic receptions in the water column. It is closely related to marine reflection seismology but also has features of refraction seismology, and uses sonar equipment and less than ideal geometries because accurate scientific determination of the seafloor may not be the primary goal of the experiments. However, the authors show how a pre-measurement inverse theory resolution analysis can be used as part of experiment planning regarding sensor placement and ship tracks, such that a desire for an experimental configuration giving the most information in bottom inversion can be quantitatively balanced with that for other needs like tracking and communication. This nonlinear geoacoustic inverse problem is ill-posed, so that one can only estimate the continuous function of seafloor properties to a limited resolution. This limited resolution varies with experiment geometry, frequency, and other such factors, and can be quantified in either a frequentist or Bayesian framework. Given statistics of the measurement noise (but without any new measurements themselves), the resolution can be quantified exactly for a linear inverse problem, and compared between different experiment geometries. Nonlinear problems complicate this picture, but if the problem can be transformed into a weakly nonlinear form then the resolution may still be explored in an approximate sense and used as a tool in the planning phase. The ideal situation is when previous seafloor estimates exist for the same region in which a new experiment with new geometry and configuration is being planned. For the scenario without previous results, a somewhat more ad-hoc approach can still compare changes in resolution across different seafloor models. This presentation demonstrates the technique for a synthetic problem involving a single stationary source and a single vertical array, but the formulation can be adapted to virtually any other sensor configuration as well.
DE: 3025 Marine seismics (0935, 7294)
DE: 3045 Seafloor morphology, geology, and geophysics
DE: 3094 Instruments and techniques
DE: 3260 Inverse theory
DE: 3275 Uncertainty quantification (1873)
SC: Seismology [S]
MN: 2007 Fall Meeting