We have empirically estimated how often fireball shocks produce overpressure at the ground sufficient to damage windows. Our study used a numerical entry model to estimate the energy deposition and shock production for a suite of 23 energetic fireballs reported by US Government sensors over the last quarter century. For each of these events we estimated the peak overpressure on the ground and the ground area above overpressure thresholds of 200 and 500 Pa where light and heavy window damage, respectively, is expected. Our results suggest that at the highest overpressure it is the rare, large fireballs (such as the Chelyabinsk fireball) which dominate the long-term areal ground footprints for heavy window damage. The height at the fireball peak brightness and the fireball entry angle contribute to the variance in ground overpressure, with lower heights and shallower angles producing larger ground footprints and more potential damage. The effective threshold energy for fireballs to produce heavy window damage is ~5 - 10 kT; such fireballs occur globally once every one to two years. These largest annual bolide events, should they occur over a major urban centre with large numbers of windows, can be expected to produce economically significant window damage. However, the mean frequency of heavy window damage (overpressure above 500 Pa) from fireball shock waves occurring over urban areas is estimated to be approximately once every 5000 years. Light window damage (overpressure above 200 Pa) is expected every ~600 years.

We have detected and performed signal measurements on 78 individual bolide events as recorded at 179 infrasound stations between 2006 and 2015. We compared period-yield relations with AFTAC nuclear period-yield data, finding these to be similar with a slight offset. Scatter in period measurements for individual bolide is found to be caused in part by station noise levels and by attenuation effects with range. No correlation was found between the infrasound signal period and any of bolide height at peak brightness, entry speed or impact angle. We examined in detail three well constrained bolides having energy deposition curves, known trajectories and infrasound detections finding some evidence at shorter ranges that a component of station period scatter is due to varying source heights sampled by different stations. However, for longer-range stations in these three case studies, we were not able to assign unique source heights using raytracing due to large uncertainties in atmospheric conditions. Our results suggest that while source height contributes to the observed variance in infrasound signal periods from a given bolide, range and station noise play a larger role.