Abstract:
The coastal environment presents some unique challenges for ray-based acoustic tomography. The need to resolve travel time effects of environmental changes over smaller distance scales pushes the experiment designer toward higher frequencies, but higher frequency signals are more susceptible to unmodeled features of the environment like surface and bottom roughness and micro-structure within the water column. At Haro Strait, it was found that these unmodeled features of the environment had a greater overall impact on the received acoustic signal than did the features of interest. Two techniques proved useful in improving the environmental estimates under these conditions. The first technique was a matched filtering of the received signal which emphasized those parts of the signal most sensitive to the environmental features of interest and least sensitive to unmodeled features of the environment. The second technique was adaptive repositioning of one of the tomography sources, made possible by mounting the source on an AUV. Tomographic resolution is dependent on the source/receiver geometry, and by repositioning a source, resolution can be concentrated at emerging features of interest. The improved resolution offered by these techniques is shown in the data from Haro Strait. [Work supported by ONR.]