Uncertainty Estimation for Atmospheric Acoustic Propagation Prediction

The propagation of acoustic signatures from vehicles, aircraft, boats and artillery can have a significant impact on military operations. Specifically combatants need to reliably detect enemy acoustic signatures while at the same time concealing their own. The physical propagation environment, including surface and atmospheric parameters, may drastically alter the range at which an emitting object may be detected. There are currently several models for predicting these acoustic propagation effects based on atmospheric profiles and/or ground topography. Unfortunately these models are sensitive to errors in environmental inputs. We propose a method for estimating the uncertainty (variance) of the acoustic propagation prediction given the uncertainty in the environmental inputs. Advantages to this new method are (1) it fits within current Parabolic Equation (PE) marching algorithms and does not require a second stand-alone code and (2) it is computationally efficient. This would allow our forces to simultaneously predict both the acoustic propagation and its uncertainty. The algorithm used to calculate the uncertainty may also be used to determine the sensitivity of the sound field to the various individual environmental parameters which could aid (1) in acoustic environmental parameter inversion and (2) in the intelligent allocation of meteorological measurement resources.