Underwater sound propagation is sensitive to specific features, scales, and gradients in the ocean environment, from turbulent processes at acoustic wavelengths to large-scale circulations at ocean basin scales. However, due to the limited ocean observations, wide range of scales, and dynamic ocean processes, it is challenging to model and predict all these acoustics-relevant ocean features at sufficient levels of accuracy. In addition, the dominant sensitivities are themselves not always well known or understood, especially for strongly nonlinear effects. Finally, acoustics sensitivities depend on the sound frequency, source-receiver configuration, and many other operational and environmental factors. To further scientific understanding and augment acoustics modeling capabilities, both process studies of nonlinear sensitivities and stochastic modeling are useful. The former enables targeted studies of complex processes using data and models while the latter augments deterministic modeling with probabilistic environmental conditions and stochastic forcing inputs. The results can capture the environmental inputs that matter and organize them by their acoustic importance, both dynamically and probabilistically.

In this work, we demonstrate the use of our deterministic and stochastic Dynamically-Orthogonal Parabolic Equations (DO-ParEq) to determine and quantify the features of ocean fields that most affect underwater sound propagation. We illustrate results for idealized seamounts and for the New England Seamounts off the eastern US coastline. A long-term goal is to utilize our nonlinear stochastic DO analysis to quantify acoustics dynamical regimes and complete global dynamical analyses of ocean acoustics sensitivity.