Infrasound production by bolides: A global statistical study

Abstract

We have examined a dataset consisting of 71 bolides detected by satellite sensors, which provide energy and location estimates, with simultaneous measurements of the same events on 143 distinct waveforms. These bolides have total source energies ranging from 0.02kt TNT equivalent yield to ≈20kt and probable diameters of order a few meters on average. We find that it is possible to detect large events with energies of ≈20kt or more globally. Infrasonic detections of these events for stratospheric arrivals have ranges between 350–17,000km and show clear wind-related amplitude modifications. We find that our period–yield relations are virtually identical to that found from AFTAC nuclear test data with the most robust period–yield correlation found for those events having multiple station averaged periods. We have also found empirical expressions relating maximum expected detection range for infrasound as a function of energy and low and high frequency cut-off as a function of energy. Our multi-variate fits suggest that 12 yield-scaling is most appropriate for long range bolide infrasound measurements with a distance scaling exponent of ≈1.1 best representing the data. Our best-fit wind correction exponent is a factor of ≈3 smaller than found by previous studies which we suggest may indicate a decrease in the value of k with range. We find that the integral acoustic efficiency for bolides is ≥0.01% with a best lower limit estimate nearer 0.1%. Finally, we conclude that a range independent atmosphere implementation of the normal-mode approach to simulate bolide amplitudes is ineffective at large ranges due to the large change in atmospheric conditions along source-receiver paths.