Probabilistic infrasound propagation using realistic atmospheric perturbations

This study demonstrates probabilistic infrasound propagation modeling using realistic perturbations. The ensembles of perturbed analyses, provided by the European Centre for Medium-Range Weather Forecasts (ECMWF), include error variances of both model and assimilated observations. Ensemble spread profiles indicate a yearly mean effective sound speed variation of up to 8 ms−1 in the stratosphere, exceeding occasionally 25 ms−1 for a single ensemble set. It is shown that errors in point estimates of effective sound speed are dominated by variations in wind strength and direction. One year of large mining explosions in the Aitik mine, northern Sweden, observed at infrasound array IS37 in northern Norway are simulated using 3-D ray tracing. Probabilistic propagation modeling using the ensembles demonstrates that small-scale fluctuations are not always necessary to improve the match between predictions and observations.