Abstract:

In ocean acoustic tomography, the travel time of ray arrivals are used to infer the sound-speed field. In typically dispersive channels, the ray arrivals are followed by energy, which is more easily described by vertical modes. Using a vertical array, the analyst can separate the modal energy in order to construct a dispersion diagram; a measurement of mode group velocity versus frequency for each mode number. The difference between the measured and computed group velocities would then be used in a stochastic inverse for the sound-speed field. In the WKBJ approximation, the expression for mode group velocity can be written as a function of a single parameter, mode axial inclination angle instead of the two parameters, mode number, and frequency. This fact suggests that information from modes with the same inclination angle (or equivalently the same turning point depths), ought to be combined. A turning point filter can be built that does exactly this. Perhaps it is not surprising that nature constructs rays from modes with the same turning point depths also. There is some advantage in this method over standard linear beamforming. [Work supported by the Strategic Environmental Research and Development Program through ARPA.]