Ocean acoustic tomography

Ocean acoustic tomography (OAT) was proposed in 1979 by Walter Munk and Carl Wunsch as an analogue to x-ray computed axial tomography for the oceans. The oceans are opaque to most electromagnetic radiation, but there is a strong acoustic waveguide, and sound can propagate for 10 Mm and more with distinct multiply-refracted ray paths. Transmitting broadband pulses in the ocean leads to a set of impulsive arrivals at the receiver which characterize the impulse response of the sound channel. The peaks observed at the receiver are assumed to represent the arrival of energy traveling along geometric ray paths. These paths can be distinguished by arrival time, and by arrival angle when a vertical array of receivers is available. Changes in ray arrival time can be used to infer changes in ocean structure. Ray travel time measurements have been a mainstay of long-range acoustic measurements, but the strong sensitivity of ray paths to range-dependent sound speed perturbations makes the ray sampling functions uncertain in real cases. In the ray approximation travel times are sensitive to medium changes only along the corresponding eigenrays. Ray theory is an infinite-frequency approximation, and its eikonal equation has nonlinearities not found in the acoustic wave equation. We build on recent seismology results (kernels for body wave arrivals in the earth) to characterize the kernel for converting sound speed change in the ocean to travel time changes using more complete propagation physics. Wave-theoretic finite frequency kernels may show less sensitivity to small-scale sound speed structure.