Global sensitivity of radar wave propagation power to environmental variables for a parabolic equation numerical simulation in maritime regions

Numerous environmental factors impact radio wave propagation in the marine atmospheric boundary layer (MABL) through effects such as scattering and refraction. Furthermore, environmental parameters can interact with each other to either compound or reduce the impact of each individual parameter. Thus, in order to properly assess the sensitivity of radar wave propagation power to environmental variables a global sensitivity approach is needed. In this study, we examine the global sensitivity of propagation power to a number of environmental variables using a parabolic equation (PE) numerical simulation for maritime regions. The sensitivity analysis is performed using the Extended Fourier Amplitude Sensitivity Test, which is a global variance-based method that can account for multi-degree interaction effects. The method is ideal for complex nonlinear models and permits computation of both leading order and total order sensitivity for each parameter. The study examines 16 environmental parameters, 8 sea state and 8 atmospheric, that are used to generate inputs for the Variable Terrain Radiowave Parabolic Equation (VTRPE) numerical simulation. This model uses a split-step rotated Green´s function parabolic wave equation solution to the scalar wave equation for transverse field components derived from Maxwell´s equations. The simulation accounts for effects of refraction (including ducting) as well as variable boundary conditions on water surfaces, such as wind seas, swell, and variable dielectric properties. Vertical profiles of atmospheric refractivity, which are assumed homogenous in range for this study, are generated using a simplified refractivity model that depends on the 8 atmospheric parameters and can generate linear refractivity profiles as well as evaporation, surface, and elevated ducts, including any combination of these ducts. Sea state parameters can be grouped into 3 general categories: surface dielectric properties, directionality, and surface roughness. Atmospheric parameters can be grouped into 4 general categories: evaporation layer, mixed layer, inversion layer, and upper layer parameters. We examine results in the context of these parameter groupings.