Characterization of the range dependence of an ocean environment to reduce acoustic estimation time

As computer capabilities continue to increase, characterization of the oceanic environment improves both spatially and temporally. This oceanographic information (specifically sound velocity) is often used to make acoustic predictions, but wide-area acoustic predictions have been limited by computation time. Currently, in order to characterize an area acoustically, propagation loss estimates are computed for equally spaced grid points over a large area. Acoustic models have become fairly fast, however, when doing multiple source and receiver configurations, multiple frequencies and several range dependent bearings for many grid points, the task can become computationally prohibitive. Range independent predictions are significantly faster than range dependent predictions, but can under-represent the acoustic propagation due to features in the sound speed structure where the environment is non-adiabatic (range dependent). Adiabatic normal mode calculations can also be done fairly quickly (faster than range independent transmission loss predictions) but for the same reason, still do not provide the best estimate of the acoustics. A method is developed and presented here to compute normal modes for an environment and use them to determine how adiabatic that environment is. This is done by comparing the number of modes and the wave numbers computed for each sound speed profile, to those of neighboring profiles. If the nature of the mode functions changes significantly, the profile is flagged as non-adiabatic. Once the adiabatic determination is made, range dependent runs are computed for the areas that are shown to be non-adiabatic and range independent runs are computed for the adiabatic areas. This process reduces the amount of run time required to assess an area, while maintaining a high level of accuracy of the acoustic characterization of the whole area. This run time reduction can be significant for many areas of the world. Estimates of acoustic coverage (that is, - the area for which acoustic transmission loss is below a threshold) of the area using the adiabatic measure are computed and compared to the full grid computation. This shows that taking advantage of the adiabatic areas to reduce the number of acoustic predictions required for an area still provides an acceptable estimate of the wide-area acoustic environment.