Technical Contribution How Loud (Pascals) is a “Loud” Explosion Airblast? What Happened to TWA Flight 800?

Abstract

Various witnesses to the destruction of TWA Flight 800, off Long Island, New York, on Tuesday, July 26, 1996, reported up to five bangs, as “loud”, “very loud”, even “incredibly loud”. Some reported that their houses shook from the blast. These reports led to an early estimate that an explosion of at least 1-tonne TNT equivalent was required to give around 20 Pa (120 dB) peak airblast pressure at their minimum 15 km distance. None of the publicized causes for this accident, however, seem capable of causing such a large explosion. Thus, some confirmation of this yield is necessary, but that depends on calibrating loudness words of witnesses into wave overpressures for translation through known explosion overpressure-distance-yield relations [1]. The acoustics community holds continuing debates about effects of amplitudes, wave frequency contents, and repetition rates on neighbourhood acceptability of impulsive noises (as opposed to continuous sounds). But practically nothing has been published about loudness perceptions of single airblasts. Extensive studies of lightning bolts and common thunder claps have seemed to be exclusively concerned with electricity and close-in shock wave formation which could contribute to damage. Since nothing can be done about thunder it doesn’t warrant research and remediation, other than for improving lightning rods. A great deal is known and written about sound and its audibility, over frequency ranges from 40 kHz down to about 200 Hz [2]. Fundamental frequencies for even tiny fireworks explosions, however, fall below about 100 Hz. Measurements were made of 60-mg TNT equivalent explosions in an attempt to define perceptions of their loudness. At 95 dB (1.125 Pa) in a single cycle of 67 Hz, it was barely heard near 2 km distance when the exact arrival time was known to the second. At closer ranges, one observer felt that 115 dB (11.25 Pa) was “loud”. Another, biased from many years of explosions testing, estimated that 125 dB (35.6 Pa) was loud enough to gain attention of an unprepared witness. Thus, that early 120 dB estimate might appear reasonable, except that a 1-ton TNT explosion wave at 15 km range from TWA-800 could have a somewhat different waveform than simple yieldscaling would predict for a classical explosion, in result of air attenuation and weather differences. Considering other information on lower frequency sound gives results shown by curves in Figure 1, where numbered references are shown in brackets. A curve for minimum audibility (0 phons) [3] is shown along with higher valued phon curves, but these were not accompanied by loudness perception