Infrasonic observations of the June 2009 Sarychev Peak eruption, Kuril Islands: Implications for infrasonic monitoring of remote explosive volcanism

of 15

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
15 pages
0 downs
Infrasonic observations of the June 2009 Sarychev Peak eruption, Kuril Islands: Implications for infrasonic monitoring of remote explosive volcanism
  This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institutionand sharing with colleagues.Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third partywebsites are prohibited.In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further informationregarding Elsevier’s archiving and manuscript policies areencouraged to visit:  Author's personal copy Infrasonic observations of the June 2009 Sarychev Peak eruption, Kuril Islands:Implications for infrasonic monitoring of remote explosive volcanism Robin S. Matoza a, ⁎ , Alexis Le Pichon a , Julien Vergoz a , Pascal Herry a , Jean-Marie Lalande a , Hee-il Lee b ,Il-Young Che b , Alexander Rybin c a CEA/DAM/DIF, F-91297 Arpajon, France b Earthquake Research Center, Korea Institute of Geoscience and Mineral Resources, Korea c Sakhalin Volcanic Eruptions Response Team, Institute of Marine Geology and Geophysics, Yuzhno-Sakhalinsk, Russia a b s t r a c ta r t i c l e i n f o  Article history: Received 1 June 2010Accepted 25 November 2010Available online 7 December 2010 Keywords: explosive volcanisminfrasound atmospheric propagationinfrasound source locationvolcano monitoringvolcano seismology Sarychev Peak (SP), located on Ostrov Matua, Kurils, erupted explosively during 11 – 16 June 2009. Whereasremoteseismicstationsdidnotrecordtheeruption,wereportatmosphericinfrasound(acoustic wave~0.01 – 20 Hz) observations of the eruption at seven infrasound arrays located at ranges of ~640 – 6400 km from SP.The infrasound arrays consist of stations of the International Monitoring System global infrasound networkand additional stations operated by the Korea Institute of Geoscience and Mineral Resources. Signals at thethree closest recording stations IS44 (643 km, Petropavlovsk-Kamchatskiy, Kamchatka Krai, Russia), IS45(1690 km, Ussuriysk, Russia), and IS30 (1774 km, Isumi, Japan) represent a detailed record of the explosionchronology that correlates well with an eruption chronology based on satellite data (TERRA, NOAA, MTSAT).The eruption chronology inferred from infrasound data has a higher temporal resolution than that obtainedwith satellite data. Atmosphere-corrected infrasonic source locations determined from backazimuth cross-bearings of  fi rst-arrivals have a mean centroid ~15 km fromthe true location of SP. Scatter insource locationsof upto~100 kmresult fromcurrentlyunresolved detailsof atmospheric propagation andsource complexity.Weobservesystematictime-variationsintrace-velocity,backazimuthdeviation,andsignalfrequencycontentat IS44. Preliminary investigation of atmospheric propagation from SP to IS44 indicates that these variationscan be attributed to solar tide variability in the thermosphere. It is well known that additional informationaboutactivevolcanic processescanbelearned bydeploying infrasonicsensorswithseismometers ateruptingvolcanoes. This study further highlights the signi fi cant potential of infrasound arrays for monitoring volcanicregions such as the Kurils that have only sparse seismic network coverage.© 2010 Elsevier B.V. All rights reserved. 1. Introduction Sarychev Peak (SP), an andesitic stratovolcano (summit elevation1446 m a.s.l.) on the northwest side of Ostrov Matua (Matua Island),Kurils (Fig. 1), erupted explosively during 11 – 16 June 2009. Theeruptionwas fi rstindicatedbysatellitedataacquiredon11 June2009that showed a thermal anomaly and weak ash emissions (SVERT,2009). Subsequently, at ~22:16 UT 12 June 2009, spectacularphotographs of an eruption column issuing from SP were taken byastronauts aboard the International Space Station (ISS) (Fig. 1, inset).These photographs also captured ash dispersed at altitude frompreviouseruptions,andpyroclastic fl owsintheprocessofdescendingthe mountain (Fig. 1, inset). Due to the remote location of the Kurils,ground-based observations are sparse. In particular, no seismicnetwork was in place on SP at the time of the eruption and theeruption did not register on any remote seismic stations (e.g., seismicstations on Paramushir, Iturup and Sakhalin at distances of 352 km,512 km and 800 km from SP, respectively). Therefore, there are noseismic dataconnected withthis event. Consequently, previous to thecurrent study, the chronology of the eruption has been constructedprimarily with satellite data (TERRA, NOAA, MTSAT; SVERT, 2009).Although the Kurils are sparsely populated, they are located within aheavily travelled air corridor linking Europe, North America, andnorthern Asia. Effective monitoring of Kuril volcanism is thereforeimperative for aviation safety (Neal et al., 2009).Acoustic waves with frequencies ~0.01 – 20 Hz are named  infra-sound . Here we report atmospheric infrasound observations of the June2009SPeruption.Energeticvulcanianandplinianexplosionscanradiate large-amplitude infrasound directly into the atmosphere (e.g.,Garces et al., 2008; Matoza et al., 2009; Fee et al., 2010a, 2010b). In contrast,seismicity(eruptiontremor)recordedondedicatedvolcano-seismic networks during vulcanian and plinian explosions may resultfromsubsurfaceprocessesand/orlimitedair-groundacoustic – seismiccoupling (Matoza, 2009). Consequently, large-amplitude infrasound  Journal of Volcanology and Geothermal Research 200 (2011) 35 – 48 ⁎  Corresponding author. Tel.: +33169264279. E-mail address: (R.S. Matoza).0377-0273/$  –  see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.jvolgeores.2010.11.022 Contents lists available at ScienceDirect  Journal of Volcanology and Geothermal Research  journal homepage:  Author's personal copy signalsmaydelineatetheexacttimingofvolcanicexplosions,whereaseruption seismicity may be relatively weak or not necessarilycorrelated with the timing of eruption into the atmosphere.Infrasound can propagate over large distances in the atmospheredue to low attenuation (Sutherland and Bass, 2004) and due to the formation of waveguides by temperature and wind variations withaltitude (e.g., Garces et al., 1998). Whereas remote seismic stationsdid not record the SP eruption, we report infrasound signalspropagatingasfaras6433 kmfromSPinthestratosphericwaveguide.However, the details of infrasound propagation in the atmosphereremain a subject of active research (Le Pichon et al., 2010a). Foratmospheric studies, volcanoes can represent repetitive sources of infrasound from known and  fi xed source locations, making themessential ground-truth sources for assessing models of infrasoundpropagation and atmospheric speci fi cations (Le Pichon et al., 2005). TheexplosivephaseoftheeruptionsequenceatSPlastedfor5 – 6 daysin duration. We use array processing to estimate infrasoundwavefrontparameters,e.g.,backazimuthandtrace-velocity(apparentvelocity of the wavefront across the array), as a function of timeduring the eruption sequence. We show that the estimated back-azimuth, trace-velocity and signal frequency content of infrasonicsignals from repetitive explosions exhibit systematic variations withtime that can be explained by atmospheric variability.This paper is organized as follows. In Section 2 we describe thesequence of observations at seven infrasound arrays deployed atranges of ~640 – 6400 km from SP. We show how these infrasoundobservations permit the reconstruction of a more detailed eruptionchronology (i.e., with a higher temporal resolution) than is possiblewith satellite data alone. We then highlight some observed signalcharacteristics resulting from atmospheric propagation and illustratetheeffectsofatmosphericpropagationoninfrasoundsourcelocations.In Section 3, we model the infrasound propagation using 3D ray-tracing and realistic atmospheric speci fi cations and attempt a sourcelocation including atmospheric corrections. We then illustrate withparabolic equation modeling how diffraction and scattering mayin fl uence infrasonic propagation from explosive volcanic eruptions.Sections 4 and 5 consist of discussion and conclusions. 2. Observations  2.1. Data The International Monitoring System (IMS) includes a globalnetwork of infrasonic stations designed to detect atmosphericexplosions anywhere on the planet (Christie and Campus, 2010).Each infrasound station consists of an array of at least 4 infrasonicsensors with a  fl at response typically from 0.01 to 8 Hz (sampled at20 Hz) and a sensitivity of about 0.1 mPa per count. Fig. 1 shows theIMS infrasound stations and an additional station, YAG, operated bythe Korea Institute of Geoscience and Mineral Resources (KIGAM),used in this study. Signals from SP were also recordedat other KIGAMinfrasoundstations,however,weselectedonestation(YAG)foruseinthis studysincethe KIGAMstations fall atsimilar rangesfromSP. IS44(Kamchatka) is the closest station to SP at a range of 643 km.Fig. 2 shows the results of applying Progressive MultiChannelCorrelation (PMCC) array processing (Cansi, 1995; Le Pichon et al.,2010b) to the infrasound stations labeled in Fig. 1. PMCC estimates wavefrontpropertiesofcoherentacousticenergyasafunctionoftimeatanarraybyconsideringcorrelationtime-delaysbetweensuccessivearray element triplets (Cansi, 1995). A grid search is performed over successivetimewindowsandfrequencybands.Acoherentarrivalinaparticular time window and frequency band is registered as a  “ pixel ” .Pixels are then grouped into  “ families ”  of pixels sharing commonwavefront properties. PMCC processing was performed in 15 log-spaced frequency bands between 0.02 and 9.5 Hz (window lengthvaried from 120 s to 30 s with overlaps 90% of window length). Fig. 2shows all PMCC pixels coming from an azimuth corresponding toSP +/ − 15° as viewed from each array (i.e., for each station, we showall PMCC detections that have an azimuth +/ − 15° of the azimuth of the great-circle path from the station to SP). In order to align thedetections in time and facilitate association of the recordings at thevariousstations,wehaveapplieda timeshifttothedetectionsateacharray in Fig. 2. The time shift corresponds to the range divided by aconstant celerity of 0.33 km/s. Celerity is de fi ned as the total rangetravelled divided by the propagation time. The celerity of 0.33 km/s Fig. 1.  Map showing location of Sarychev Peak (SP, red triangle), infrasound arrays that recorded signal from SP (blue inverted triangles), and infrasound arrays that did not recordsignal from SP (black inverted triangles). Signals are observed at long-range to the west of SP corresponding to the stratospheric downwind direction in June 2009. Inset: Astronautphotograph of SP eruption column taken at 22:16 UT 12 June 2009 from the International Space Station (ISS). Image credit: NASA's Earth Observatory.36  R.S. Matoza et al. / Journal of Volcanology and Geothermal Research 200 (2011) 35 – 48  Author's personal copy usedinFig.2istypicallyappropriateforinfrasoundpropagatinginthetroposphere. For infrasonic propagation in the stratosphere, a slowercelerity of 0.3 km/s is more appropriate. Thus, tropospheric arrivalsfrom SP align vertically on Fig. 2, whereas stratospheric arrivals willnot align perfectly in the vertical. This is barely visible for the time-scale of data shown in Fig. 2.Fig. 2 shows that a long sequence of infrasonic signals arrived fromthedirection of SP atIS44, IS45,IS30, and YAG betweenJuliandays 162and1672009(11 – 16June2009).Inaddition,severalofthesignalswithlarger amplitude propagated to longer range. IS31 (Kazakhstan) is thefurthestrecordingstationatarangeof6433 km.WenotethatIS53,IS39,and IS59 did not record infrasound from SP despite being closer to thesource than IS31 (Fig. 2). The pattern of signal detectability on Fig. 1 is consistentwith propagationofthesignals to longrange in theseasonalstratospheric duct. Stratospheric winds blow westward (from east towest) in the summer at midlatitudes, leading to favorable signalreception tothewestof a source. IS53,IS59,and IS39are located to theeast and south of SP where stratospheric propagation is not generallypredictedatthistimeofyearattheselatitudes.Itisknownthat~80%of all infrasonic signals in the 0.2 – 2 Hz bandrecorded globally onthe IMSnetwork correspondtosignals received in the stratospheric downwinddirection (Le Pichon et al., 2009).We note in Fig. 2 the presence of some coherent noise sourcesfalling withinthe chosenazimuthbounds, but thatarenot likely to besignal from SP. This coherent noise is not associated across multiplestations. In particular, continuous coherent signal is recorded atbackazimuths greater than the true backazimuth of SP at IS30 andYAG. This signal is likely microbarom noise (e.g., Willis et al., 2004)given its continuous nature, frequency content (not shown on Fig. 2),and direction pointing towards the ocean (approximately samedirection as SP for these stations located near eastern coastlines).  2.2. Observed signal chronology at IS44 IS44 is the closest station to SP and recorded the greatest numberof detections associated with the eruption. The chronology of signalsrecorded at this array therefore gives the best indication available of the chronology of the SP eruption sequence. We note in Fig. 2 that alldetections observed at IS45 and IS30 are also observed at IS44 but notvice versa. Fig. 3 shows the infrasonic waveforms at IS44. Thewaveforms correspond to a time-domain beam (e.g., DeFatta et al.,1988) at the great-circle azimuth of SP and a trace-velocity of 0.34 km/s (typical trace-velocity for  fi rst-arrivals and majority of detectionsatIS44,seeSection2.3).ThewaveformsinFig.3areshown attheirtrueobservedtimeandarenotcorrectedbyceleritybacktoanassumed srcin time at SP. At the beginning of the sequence there arethree impulsive signals between 6 am and 9 am UT on 11 June 2009(day 162). These signals appear to represent earlier, less-vigorousexplosion precursors to the main phase. Following these explosionsthere is a repose in observed signal for ~17 h. Then, beginning at~02:10:44 UT 12 June 2009 (day 163), there is a long sequence of energetic coherent infrasonic signal srcinating from SP (Fig. 2). Forthe  fi rst ~10 h following 02:10:44 UT 12 June 2009, the signal is aconvoluted superposition of broadband infrasonic tremor and moreimpulsive explosion signals, typical of vulcanian – plinian eruptions(Matoza et al., 2009; Fee et al., 2010b). This then transitions gradually into a series of more isolated explosion signals by ~12:00 UT 12 June2009. Fifteen such repetitive explosions occur between 12:00 and24:00UTon12June2009withtheinter-eventtimespacinggraduallyincreasing during this time (indicated by arrows between the mid-pointandtheendofthesecondtraceonFig.3).Fromthebeginningof 13June2009(day164)totheendof15June2009(day166)wecountten large explosions srcinating from SP. These signals consist of broadband, high-amplitude infrasonic tremor signals lasting tens of minutes to over an hour in duration each. The amplitude of theseexplosion signals are generally of the same amplitude or larger thanthe sequence recorded on 12 June (day 163). Consequently, thesesignals were recorded further from the source (Fig. 2). The explosionon Fig. 3 between 17:00 and 18:00 UT on 15 June (day 166) wasrecorded at IS31 (Fig. 2). The waveforms in Fig. 3 are further complicated by local wind noise and a regional seismic event(travelling at seismic velocity, azimuth not associated to SP). Thelevel of detail visible in the waveforms at IS44 at a range of 643 km isremarkable and points to the utility of remote infrasonic arrays formonitoring erupting volcanoes.ThesignalchronologycanbeanalyzedinmoredetailbyconsideringtherateofPMCCdetectionsasafunctionoftime.AwaveformplotsuchasFig.3showsclearlythosesignalsthathavehighsignal-to-noiseratio, Fig. 2.  PMCC processing of 10 infrasound arrays deployed globally (see Fig. 1). Arrays are shown in order of increasing range from SP. Detections are displayed in terms of theirbackazimuth deviation from the true great-circle path from each array to SP (azimuth scale is constant for all stations). Color bar corresponds to  log  10  ( N  ) where  N   is the number of PMCC pixels in a bin of size 0.1° in azimuth and 3.5 min in time. Detections at each station have been aligned in time back to an assumed srcin time at SP by time-shifting theinfrasonic arrivals by a celerity of 0.33 km/s. Clear infrasonic detections from the entire SP sequence are recorded at IS44 (range from SP  r  =643 km), IS45 (1690 km), IS30(1774 km), and YAG (2310 km). Infrasonic detections from various individual explosions within the entire sequence are also recorded further away at IS34 (3450 km), IS46(4655 km) and IS31 (6433 km) (indicated by arrows). No detections are recorded at IS53, IS39 and IS59.37 R.S. Matoza et al. / Journal of Volcanology and Geothermal Research 200 (2011) 35 – 48  Author's personal copy but no distinction is made between coherent acoustic signal andincoherent noise (e.g., wind noise). In addition, it is dif  fi cult in Fig. 3 toidentifysignalpacketdurationsforindividualarrivalswithinthe~10 h,highly convoluted sequence at the onset of the main phase between~02:10:44 UT and 12:00 UT on 12 June 2009 (day 163). Fig. 4 shows abar graph of the number of PMCC detections at IS44 occurring perminute from an azimuth +/ − 15° of SP. Since these detections onlycorrespond to coherent signal srcinating from SP, it is easier in Fig. 4than in Fig. 3 to separate the individual signal arrivals. From thisinformation we pick automatically the onset time  t  on  and end-time  t  off  ofthesignalpacketswithanaccuracyof1minute.Table1showsthelistof signals identi fi ed with this method. The signal onset times have alsobeen corrected back to the srcin time at SP assuming a celerity of 0.33 km/s. Considering that the true celerity for  fi rst arrivals may varybetween 0.30 and 0.34 km/s we estimate that the srcin times at SPlisted in Table 1 would be accurate to ~5 min. Following theidenti fi cation of   t  on  and  t  off   for all signals in the sequence, we applied a fi nal procedure to identify the average dominant period (sec) andamplitude (Pa) of each signal received from SP. The raw data for IS44were beamformed sequentially at the azimuth and trace-velocity of  Fig. 3.  Infrasonic waveforms recorded at IS44 (Kamchatka) for the entire SP eruption sequence (5 days: 11 – 16 June 2009, or Julian days 162 – 167 2009). Each waveform shows onefull day of data. E.g., Origin time of upper trace: 0000 UT 11 June 2009 (day 162). Origin time of lower trace: 0000 UT 15 June 2009 (day 166). Infrasound array data have beenbeamformed(waveformsalignedandstackedforazimuthofSPandacousticvelocity)usingatime-delaybeamformerand fi ltered0.5 – 5 Hz.Notethatsomeofthefeatures observedon this plot are local wind noise and not coherent infrasonic signal (refer to Fig. 2). Fig. 4.  Chronology of coherent infrasonic signal at IS44 compared to the eruption chronology inferred from satellite data by SVERT. Lower plot: number of PMCC detections perminute at IS44 srcinating from the direction of SP. Time of arrivals are corrected back to an inferred srcin time at Sarychev assuming a celerity of 0.33 km/s. Black horizontal barsaboveplotrepresentbeginningandendtimesofcoherentsignalpackets.Greybarsrepresentexplosiononsettimes+/ − 15 mininferredfromsatellitedatabySVERT.Verticalextentof each grey bar is scaled relative to the maximum plume altitude inferred by SVERT. Upper plot: an expanded view of lower plot between Julian days 163 and 164.5 2009.38  R.S. Matoza et al. / Journal of Volcanology and Geothermal Research 200 (2011) 35 – 48
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks