Finite-Difference Time-Domain Simulation for Interpreting Regional Infrasound Arrivals and Estimating Propagation Path Effects

New full-wave infrasound simulation tools were developed based on an optimized finite-difference time-domain solution of 2D axisymmetric linear and nonlinear fluid dynamic equations. These tools were applied to estimate propagation path effects and enhance phase interpretation for regional infrasound arrivals. Middle-atmosphere full-wave predictions for a range of idealized atmospheres show that shadow zone arrivals and fast and slow arrival pair components from multiple ducts can be delineated using a combination of full-wave modeling and geometrical acoustics. Relationships between phase velocity and effective sound profile were established for each arrival type. Implications for source characterization were that amplitude and spectrum variation with distance was unique for each arrival type which documents that path effects must be estimated separately for each arrival type, and that the source spectrum was maintained at all ranges beyond first bounce distance only for the fast arrival. Full-wave predictions for 8 surface explosions in Nevada using high-resolution, event-specific atmospheric specifications showed that the phase interpretation methods developed previously for idealized atmospheres are also applicable to these more realistic atmospheres. These phase interpretations could be linked to the observations in many cases, but this relationship was dependent on the accuracy of the simulation result as well as requiring that phase velocity could be reliably estimated for the observations using array data. Source scaling factors were derived by matching waveform integrals for equivalent arrival types in the observations and synthetics. Application of these scaling factors produced a strong relationship with explosive weight, despite a small range of source sizes (less than a factor of two). The combined linear and nonlinear full-wave modelling methods developed in this thesis were applied to predicting stratospheric and thermospheric infrasound observations from the September 2017 North Korean underground nuclear test. Linear modeling was used to predict stratospheric arrivals using arbitrary source amplitudes which were subsequently scaled to the observations by matching amplitude integrals over a window centered on the stratospheric arrivals. A range of absolute source amplitudes centered on the value determined by the stratospheric scaling factor were then used to initialize nonlinear simulations for predicting thermospheric arrivals. The nonlinear effects on the amplitude and period of thermospheric arrivals as a function of source size were analyzed to explore the suitability of thermospheric arrivals for source characterization. These modeled effects matched the thermospheric observations from the explosion when the source amplitude from stratospheric scaling was used, illustrating the importance of including both nonlinear and linear propagation effects into a single numerical code in order to assess the complete suite of atmospheric arrivals from large underground nuclear explosions.