Moving Source/Receiver Underwater Acoustic Tomography

Remote sensing of the interior ocean environment by acoustic signals promises to be a cost effective approach to obtaining high resolution measurements over large areas. In this paper, an approach to remote sensing the interior sound speed of the ocean using historical sound speed data, spatially sparse sampling of sound speed at points within the ocean valume, and acoustic measurements is described. The approach is based on the technology developed in the latter 1970's to produce 3-dimensional images of portions of the human body for use in medical diagnosis, known as Computer Aided Tomography. A recent paper by Munk and Wunsch [1.] describes a system using moored acoustic sources and receivers for remote sensing the ocean interior. Since the acoustic signals respond to sound speed, the result of the tomographic procedure is an estimate of the sound speed field. This paper presents specific results for remote sensing a vertical slice of the ocean sound speed field. The ocean sound speed field is a three-dimensional time dependent process with fluctuations on space and time scales of many different orders of magnitude. There are very intense processes which produce fluctuations equal to 1% of the nominal sound speed over time scales of tens of days, horizontal space scales of 100 kilometers, and vertical space scales of one kilometer. These intense oceanographic processes can be resolved spatially by systems with horizontal resolution of tens of kilometers and vertical resolution of hundreds of meters. Since the processes change on time scales of tens of days, the measurement procedure can occur over a period of days to produce a single image valid for the entire time period. The remote sensing system must provide estimates which are sensitive to fluctuations in sound speed of 1% of the nominal sound speed. The tomographic estimate is based on historical sound speed data, spatially sparse samples of sound speed, and acoustic measurements. The acoustic measurements are derived from broad band transmissions which resolve the acoustic multi-paths between source and receiver. The path travel times, amplitudes, and arrival angles contain information related to the ocean sound speed field. In a vertical slice of the ocean along a line connecting source and receiver, at each range there is a different set of ray paths connecting the source and receiver.The rich variety of ray path structures obtained when the range varies provides a means of resolving small spatial variations of the sound speed field. A number of tomographic inversion techniques can be applied. The discrete time Kalman filter appears to be the most appropriate technique. The results of a numerical experiment show that moving source/receiver tomography is feasible, can provide a horizontal spatial resolution on the order of tens of kilometers and is sensitive to sound speed fluctuations of
10^{-3}
times the nominal sound speed in a vertical slice of the ocean.