Characterization of explosion signals from Tungurahua Volcano Ecuador

Characterization of explosion signals from Tungurahua Volcano, Ecuador David Fee and Milton Garces Infrasound Laboratory Univ. of Hawaii, Manoa dfee@isla. hawaii. edu Robin Matoza Laboratory for Atmospheric Acoustics (L 2 A) Scripps Institution of Oceanography

Overview • Tungurahua Volcano • Array(s) • Explosion Algorithm and Events • Examples – March 2007 Sequence – May 2006 • Explosions Source • Cross-Correlation • Conclusions

Tungurahua Volcano • 5023 m high, 3200 m of relief • Frequent eruptions characterized by pyroclastic flows, lavas, lahars, as well as tephra falls • Over 30, 000 people live in close proximity, evacuated in 1999 • Significant ash ejections resulting from nearly constant tremor and explosions • Motivation: – Understand dynamics and evolution of explosions – Aid general understanding and monitoring Images Courtesy Instituto Geofisico

ASHE Arrays - RIOE • 4 Element Array, ~100 m aperture • Chaparral 2 Microphones • Flat between 0. 1 -200 Hz • Collocated BB seismometer • Porous hoses in open field • Recorded signals from Tungurahua and Sangay Volcanoes 37 km 33° 43 km 132°

Explosion Detection Algorithm • Time period: 2/15/06 -11/1/2007 • High-pass filter data >. 5 Hz • STA/LTA event onset and end time – 2/5 secs, 3/40 secs – Detection must be on all 4 channels • Run PMCC between 0. 5 -4 Hz – 10 bands, 10 sec windows – Families with correct azimuth (± 7°) during event time – Minimum RMS amplitude >0. 02 Pa RMS – Minimum family size >15 pixels

Amplitude and Number Events • 9331 Events detected • >400 per day during peak • Events clumped during periods of high activity • Amplitudes: 0. 018 -24. 4 Pa Mean = 0. 64 Pa 4/1/06 7/1/06 10/1/06 1/1/07 4/1/07 7/1/06 10/1/06 • Durations: 0. 1 -16. 5 s Mean = 3. 95 s

Acoustic Source Energy • EAcoustic=2πr 2/ρc ∫ΔP(t)2 dt r=source-receiver distance ρ=air density • Energy normalized by reference event – Removes geometric spreading and topographical effects – Reference Event: 1. 30 x 107 J C=sound speed ΔP=change in pressure • Assume spherical spreading • More energy more eruptive material?

Energy Release • Energy ratios: 7. 5 x 10 -5 -502 Mean = 0. 81 Largest explosions follow 7/14/06 VEI 3 Eruption 4/1/06 7/1/06 10/1/06 1/1/07 4/1/07 7/1/06 10/1/06 • Group eruptive activity: – – – Background tremor May 06 July 14 -15 th, 2006 August 16 -17 th, 2006 March 07 10/1/06 1/1/07 4/1/07 7/1/06 10/1/

Effective Yield • • Convert Explosion Energies to Effective Yield 1 ton of TNT = 4. 184 GJ Largest explosion=1. 56 ton, most around. 001 ton (~1 kg of TNT) Volcanic explosion in fluid, relationship may not hold 4/1/06 7/1/06 10/1/06 1/1/07 4/1/07 7/1/06 10/1/06

March 2007 Sequence - Example • • Moderate-High Activity resumed between 2/15 -4/15 Significant number of explosions and associated ash Seismic Tremor and LPs returned 2/23/07 Significant number of explosions starting 2/24

February 24 th Event • • 2/24/07 Impulsive Onset Signal lasts ~5 mins Sustained amplitude ~± 1 Pa • Ash >40’ 000 ft • Jetting? Similar spectrum

Acoustic Source Energy Example Explosion: 3/8/07 0745 UTC ~10 Pa at 36. 89 km 368, 900 Pa at 1 m 205 d. B (re 20 μPa)! Effective Yield: 0. 115 ton (105 kg) TNT

March 2007 Explosion Energy • Most energetic explosions during middle of sequence • Cloudy weather hampered visual monitoring for much of sequence • Energy and number of explosions correlate well with heightened volcanic activity

Observation vs. Recording: April 4 th, 2007 • Good recording and viewing conditions. Selected day for eyewitness, satellite, infrasound correlations • Observation: 0450 UTC Explosion. Vibration of windows in Banos (7 km) and heard at observatory (13 km). Clear weather and constant emission reaching 8. 5 km asl (~28, 000 ft) • Infrasound: 2007 -04 -04 04: 51: 03, 2. 91 Pa, 6. 4 sec, 1. 571 energy ratio

Explosions Infrasonic Harmonic Tremor • Mid-May 06: Explosions trigger gliding harmonics lasting up to 30 mins • Very little ash during these explosions/tremor • New Injection of Magma?

Explosions Seismic Harmonic Tremor • Band-limited sustained seismic tremor • Similar frequencies, but harmonics not very apparent (low SNR as well)

Explosion Source • Ruiz et al. 2005: analyzed travel times of seismic and acoustic first arrivals (ΔT =Tacoustic-Tseismic) • Large variations in ΔT source location variability? • Concluded explosions events originate <200 m, followed by outflux of gas, ash, and solid material ~1 s later • May 06 Explosions similar to acoustic recordings from Arenal Volcano, Costa Rica (Garces et al. , 98) • Explosion in low sound speed, low density magma-gas mixture would couple better into the atmosphere acoustic impedance match • Then decompression front propagates into conduit and create resonance • Substantial pressure perturbation could destabilize the melt and initiate flow • Explosion near surface of a gas-rich conduit creates a resonance that transmits into the atmosphere and couples into earth through the conduit walls

Infrared Video 2006/7/31 Explosion ~300 m 0. 1 0 -0. 1 • Somewhat emergent onset, relatively low amplitude • Long duration • Liquid magma ejected

Cross-Correlation • • Pick “master” waveform for subset of events Cross-correlation for each event Look at evolution of correlation value? Parameters: 0. 1 -5 Hz, Window: -3 s, +8 s from onset, amplitude >. 5 Pa Test Master Waveform

Cross-Correlation Results • Subset data between 5/11 -5/16 2006 • 385 Explosions • Waveforms very similar on 5/14 Master

Cross-Correlation – Families • • 37 km away so atmosphere effects may affect waveform similarity Possible solution: compare waveforms with similar atmospheric conditions Use 0. 02 -. 1 Hz as a proxy for wind speed (Fee and Garces, 2007) Sort explosions by waveform similarities (Green and Neuberg, 2006; Umakoshi et al. , 2003) 5/11 Waveform Family 5/14 Waveform Family

Conclusions • Significant number of high s/n explosions recorded from Tungurahua Volcano • Similarities and differences exist: amplitude, duration, energy, correlation, ash content, harmonics • Other data sources necessary to understand effect of explosions – Some ash rich, some ash poor • Understanding explosions key to hazard monitoring and dynamics at Tungurahua • Future Work – Correlate explosions with observed activity: ash, pyroclastic flows, incandescent blocks – Waveform cross-correlation by families: group by correlation and similar atmospheric conditions – Model acoustic/seismic explosion source and determine how it relates to tremor