S001-0016
Nonlinear acoustic propagation single-point spectral indicator: Comparative performance analyses for Sakurajima, Yasur, Popocatépetl, Augustine, Nabro, Reventador, Tungurahua, Villarrica, Cotopaxi, and Calbuco explosion waveforms
Nonlinear acoustic propagation single-point spectral indicator: Comparative performance analyses for Sakurajima, Yasur, Popocatépetl, Augustine, Nabro, Reventador, Tungurahua, Villarrica, Cotopaxi, and Calbuco explosion waveforms
Monday, 7 December 2020
Poster
Abstract:
Volcano infrasound (acoustic waves below 20 Hz) is becoming an increasingly popular tool for understanding erupting volcanoes due to its ability to enable rapid estimates of eruption source parameters which are important inputs to forecast models of volcanic hazards. Source parameter estimates are commonly made assuming linear acoustic propagation; however, nonlinear propagation (e.g., shock-wave formation) may cause changes to the wavefield that distort recordings made at a distance. These changes include upward transfer of acoustic energy in the power spectrum that can occur propagation at high amplitudes near the source or at low ambient pressures above the tropopause. If these effects can be quantified, their influence on acoustic recordings can be assessed during post-processing. We investigate the suitability of a previously developed, frequency-domain, single-point, quadspectral density nonlinearity indicator to quantify nonlinear changes to volcano infrasound over a range of recording distances (0.4 – 264 km), peak amplitudes (0.15 – 449.48 Pa) and eruption styles (Strombolian to Plinian). We use infrasound data from at least ten volcanoes including Sakurajima (Japan), Yasur (Vanuatu), Popocatépetl (Mexico), Augustine (Alaska), Nabro (Eritrea), Calbuco (Chile), Reventador (Ecuador), Tungurahua (Ecuador), Villarrica (Chile) and Cotopaxi (Ecuador) to evaluate the indicator performance. Preliminary results suggest that cumulative energy transfer of up to a few decibels (e.g., 2.1 dB at Augustine) can be accurately detected when receivers are close to the source (<3 km) with good line-of-sight. Conversely, the method does not accurately recover nonlinear energy transfer from recordings are affected by multiple topographic reflections or wavefield interactions with atmospheric structures (e.g., at Sakurajima). These results suggest that nonlinear changes commonly occur near the source of high-amplitude volcanogenic acoustic wavefields, but these changes are easily obscured by first-order propagation effects. For example, wavefield interactions with topography and variable sound speed gradients inhibit the performance of the nonlinearity indicator.