The largest dome collapse of Soufriere Hills Volcano removed ~210x10 July 12/132003 Eruption Timing 20:00 LT 12 July ~3kmis from SHV. LT18:00 12July pyroclastic flows build in intensity throughout the day. 07:00-09:00 LT 12 July rainfall occurred two daysfor priorto 12 July. time in between each event decreasing as the eruption recorded by small seismic events withthe before the eruption, beginning July 8, increasing seismicity was 04:00 LT 13 July 23:45 LT 12 July 23:39 LT 12 July collapse. volcanianSeveral short explosionsresult continue 23:34 LT 12 July 23:00 LT 12 July pyroclastic flowintensity. dense units of pyroclastic are also flows entering the ocean.The record seismic showsincreasean in mount andradome, and single component, very-broad-band Sacks-Evertson dilatometer. Technologies series-5000 tiltmeter, Ashtech U-Z CGPS receiver with choke ring antenna with SCIGN CIW (Duke Hz-1kHz), Pinnacle Each CALIPSO site is equipped with a three component seismometer CALIPSO Borehole Sites (Mattioli et al., 2002) 2002) etal., (Mattioli band Sa thevery-broad- of Schematics collapse. theJuly 12/13 from accumulation ash of cm ~11 Note 2003. 14, July AIRSCGPSsite Figure 3. envisioned in the design of the installation at SHV. correlate with observations and demonstrate strainmeter sensitivity to geophysical processes not simulated ocean waves were compared toobserved strain atTrants. Simulated wave heights These tsunami arehypothesized to be signal the strainmeter is observing. The spectral power test that pyroclastic flows and dome collapse into the ocean at the TRV delta stimulated tsunami. initiation and propagation to the TRV delta and the coastline. GEOWAVE was used to model wave packets with periods ~250-500 The s. strongest os 10 dome collapseconsisting ofx210 time (LT)12July. The eruption occurred 23:30 LT 12 July, with Pyroclastic flows were observed entering the ocean at the TarRiver Valley (TRV)delta at18:00 local up to12/13 July 2003 dome collapse eruption of SoufriereVolcano,Hills , Indies. West offset by 7-12 minutes from theseismic signals leading (CALIPSO) borehole sites recorded dilitation Strainmeters atthree Caribbean Andesite Lava Abstract 6 m Valley delta Pyroclastic flows entering the ocean at the Tar River 3 (DRE). Dilatometers at three sites recorded complex, correlated long period oscillatory wave cks-Evertson dilatometer eruption withends a decreasing seismic signal. - GOES-12 satellite observes an eruption cloud at ~16 km. - HARR seismometer andmicrophone the record impact of ballistic clasts. - peak seismicity occurs during the major dome collapse. - pyroclastic flows are observed reaching the ocean at the Tar River Valley delta, which - pyroclastic surges are observed 2-3 km offshore, suggesting that largevolumes of - transmission from MVO instrument at YardWhites is lost during dome collapse. - . rainstorm begins and initiates rockfalls and small pyroclastic flows. The - ing from the unroofing of the conduit during dome iltr inlwsa rns h ienearest to cillatory was at signal Trants, the site apoce (Voight et al., 2004). Intense tropical approached Island Precision-Seismo-geodetic Observatory al., 2004). al., 2004). et (Mattioli SHV of side eastern the on deposits flow pyroclastic grey is the delta TRV The by ASTER. taken image Infrared Color Right: SHV. around stations monitoring other and squares) (red sites borehole Map ofCALIPSO Left: Figure 4. SHV Monitoring Sites 6 (2), Alan Linde (4), Peter Malin (5), Jurgen (5), PeterMalin (4), (2), AlanLinde Elizabeth Van Boskirk (1), Barry Voight (2), (1),Barry Voight Boskirk Elizabeth Van m 3 DRE of material.DRE of Several days Strainmeter during Massive DomeCollapse atSoufriereHills Volcano, Montserrat,WestI Pyroclastic FlowInducedTsunamiCapturedby Borehole nasa geometry the of TRV delta. inincluded the model (Watts and Waythoma water waves. The model treats the flow as a solid block that decelerates from an initial velocity. Underwater deformation of t are exactly the same used as to generate a wave from a subarial bathometry the regionfor of Montserrat is smoothed to reenact the underwater . The equations to generate apyr and 2003;Waythomas, etal.,Watts 2003; and Kirby,simulateWei 1995). the wave initiation and development of atsunami (Watts intotheprogram FUNWAVE, to model for wave propagation combined with TOPICS, which is input GEOWAVE uses a Boussinesq GEOWAVE b) Taken August 12, 2003 from the same location. location. the from same 2003 12, August Taken b) View 2003. 31, May taken (SHV) Volcano Hills Soufriere a) Figure 1. c) July 12-13 2003 dilatometer and vertical seismometer seismometer vertical and dilatometer 2003 12-13 July c) Approximately 210 M 210 m Approximately SSE. al., 2004). et (Mattioli 00:44 12 July at 0 with inseconds is Timescale SHV. from 5.8km is which site, TRNT from records collapse. eruption/dome 12/13 July 12/13 July 2003 Eruption 2003 July 12/13 Neuberg (6), PilpWts() hitn iiiaat ( Widiwijayanti (3),Christina PhillipWatts 3 of the dome collapsed during the the during collapsed dome of the (Watts and Waythomas, 2003). 2003). and Waythomas, (Watts wave water acoherent produce and water displace would flow debris pyroclastic The shear. plume and pressure plume c) plume; and flow debris pyroclastic b) explosion. steam a) lefttoright: from Traveling flow entering ocean. the apyroclastic of section Cross 6a. Figure al., 2003). et (Watts underwater downslope traveling landslide a as acting block GEOWAVE Schematics offorces and kinematics GEOWAVE block ofmodel. Figure 6b. water waves. water waves. directly proportional to the the to proportional directly is strain volumetric the since Trants offshore area km aseveral over load the examine will We space. half elastic uniform a on load a for Functions Green’s the calculate to is step next The time. afixed for space in are averaged the results and Trants around aradii as placed are (orange) gauges Several in GEOWAVE. used gauges represent circles orange and Blue delta. TRV the at entry the of geometry the torecreate sources pyroclasticflow has five entry GEOWAVE. The simulated in used sources entry represent circles Green eruption. the during had operating strainmeters AIRS and GERS, TRNT, and sites borehole CALIPSO are Red stars Bathymetry. Montserrat 7. Figure Selwyn Sacks (4), Eylon Shalev (5), R. (5), Shalev Selwyn Sacks(4),Eylon s, 2003). GEOWs, ndies: ObservationsandTheory F E B C D A

Cumulative Normalized AVE simulated r AVE simulated Frequency (Hz) Nanostrain Frequency (Hz) 6 3 0.000 0.002 0.004 0.006 0.008 d -400 Vo l um e ( 10 m ) Amplitude 400 100 5 200 0.4 x 10 0.2 0.6 0.8 -1 Vof the 6 2 4 8 0 0 1 0 debris flow. GEOWAVE handles frequency dispersion, simulating de HARR seismicenvelope TRNT spectrogram HARR spectrogram 1 00 02:00 00:00 5 I 10 1 1 1 1 6.7¡N 6.8¡N 6.7¡N 6.8¡N Time [hourUT on 13July 2003] Time 15 Fi northeast. northeast. the to is Antigua and northwest the to is Nevis July. 12 LT 23:30 at peak eruption the with consistent is which 2004), al., et (Pelinovsky July 13 LT 1:00 and July 12 LT 23:00 between in water rise abrupt an reported witnesses several southeast, tothe km ~70 Guadeloupe, 2003 July 14 taken image Satellite 2. NASA Figure ( HH:MM UTC 13 Jul HH:MM UTC13 300 sec 50 sec II -2.0 -1.5 -1.0 -0.5 -6.0 -4.5 -3.0 -1.5 g 0.0 0.5 1.0 1.5 2.0 0.0 1.5 3.0 4.5 6.0 ure 4.Mattiolietal.,2006,Geolo meters meters 20 nu and enough allows entry tosources recreate the unnup a 62.2¡W 62.1¡W 25 III 04:00 b y 2003 HARR Z-cmp seismogram c 30 ) ) lnMtil 1,DrkEsot (2 Elsworth Derek Mattioli(1), 2), Glen TRNT dilatation IV 35 06:00 1000 2000 3000 38 Z max 100 sec -4.0 -3.0 -2.0 -1.0 to higher frequencies (Mattioli et al., 2006). et al., (Mattioli frequencies to higher spread is III,energy event collapse peak during and Hz, 0.006 and 0.004 between is power of Most E. in dilatation TRNT for Spectrogram F: HARR. at energy seismic as recorded are events same after min 7–10 occur that dilatation in amplitudes Peak Hz. at>0.002 filtered collapse events. E: TRNT dilatation highpass significant during is observed energy frequency higher that and band Hz 1–3 in is ofpower most C.Note that in seismogram for Spectrogram C: HARR vertical component seismogram. D: inA. amplitude HARR-normalized from volume Roman numerals. B: Cumulative collapse as shown are divisions major and envelope onseismic annotated are flows pyroclastic Individual 2003. July 12 (UTC) at23:00 start records All seismometer. broadband surface for HARR amplitude seismic A: Normalized and Dilatation. Figure 5. Eruption Seismicity 0.0 1.0 2.0 3.0 4.0 0.0 1.0 2.0 3.0 4.0

meters meters 62.2¡W 62.1¡W Guadeloupe Geology Mattioli etal.,200 Figure 3. Steven J Sparks (7), Simon R Young R StevenJSparks(7),Simon gy Montserrat 6 for 7Mmfor are Simulations extremes. model to scaled are level) wave heights below sea for blue and above heights Colors (red wave for GEOWAVE simulations Figure 8. of 16700 m of 16700 ratio towidth volume with ms–1 at ~70 sea entering et al., 2006). (Mattioli toTRF. proximal localized m 14 as high as heights with meters, maximum waves of few experienced Bay, Trant’s to TRF from Montserrat, Eastern domain. over entire temporal maximum wave height 3 he flow is not not is he flow pyroclastic flow oclastic flow . 2 . Z max is is max . Z Detailed ep- f) Average of Trants 250 m m 250 of Trants Average f) 250m Trants, Offshore e) Airport d) Point Spanish c) Ghaut Bottom White’s b) River Tar at Entry a) amplitude Wave 11. Figure Analysis series. time gauge the synthetic all in Hz 0.01 at peak broad a with associated power at spectral correlation of increased the signify lines The files. gauge 6 using radii m 500 of Trants Average e) files 5 gauge using radii m 250 Trants of Average d) TRNT from m 250 Trants, Offshore c) Airport b) Point Spanish a) dilatation. the TRNT with outputs amplitude wave GEOWAVE compare to used was of GMT Spectrum1d output. GEOWAVE simulation for Spectra 9. Figure causing ocean excitation should be simulated. half-plane. Also, to further recreate the actual collapse events of 12/13 July 2003,multiple pyroclastic flows tsunami, however, one needs to solve Green’s function, tsunami (Mattioli etal., 2006). Toquantitatively compare thesignals ofthe strainmeterto thesimulated unique strainmeter observations may be explained by ocean loading from pyroclastic flow generated collapse andpyroclastic flow events leading up tothepeak collapse. Detailed simulations confirmthatthe pyroclastic flow entering the TRVdelta inwhich to create tsunami similar tothe12/13 July2003dome GEOWAVE, along with the methods applied by Voight and provideWatts parameter limitations to simulate a Conclusions Herd, T. Syers, P. Williams, andD. Williams, 2006,Unique andremarkable strainmeter measurements of pyroclastic flow-generate Walder, mass J.S., P. O.E. Watts, flows, Sorenson, andK. Janssen (2003), Tsunami generated by subarial Record Lava-Dome Collapse onMontserrat byBorehole CALIPSO Instrumentation Project, draft manuscript submission.for Voight, B.,G.S. Mattioli, S.R. Young, S.J. Arkansas. Van Boskirk, E., 2006, Pyroclastic Hills Flow Generated Volcano, Tsunami Recorded by CALIPSO Borehole Strainmeters atSoufreiere Herd, T. Syers, P. Williams, andD. Williams (2006), Unique andRemarkable Dilatometer Measurements Pyroclastic of Flow-Generat Mattioli, G.S., B. Voight, A.T. Linde,I.S. Sacks, P.Watts, C. Widiwijayanti, S.R. Young, D. Hidayat. D. Elsworth, P.E. Malin, PBO Instrumentation of CALIPSO Hills Captures Volcano, World-Record Montserrat Lava (Fea Dome Collapse of July 2003on Soufrière Mattioli, G.S., S.R. Young, B. Voight, R.S.J. Sparks, E. Shalev, S. Sacks, P. Malin, A. Linde, W. Johnston, D. Hidayat, D. Elsw Mattioli, G.S., C. A.T. Widiwijayant Linde,I.S.Sacks, P.Watts, Systems (Nov. 2002), Heathman, WA Workshopand continuous onActive Magmatic GPS:The CALIPSO Project planand goals, EarthScope Mattioli, G., D. Elsworth, A. Linde,P. N. Malin, McWhorter, D. Nielson, S. Sacks,E. Shalev, Voight, andS. B. Young, 2002,Ma Linde, A.,G.S. S.R. Young, Mattioli, B.Vo Hidayat, D., Linde,A.,G.S. S.R. Mattioli, at . ..Wyhms(2003), Theoretical analysi P.,C.F.Watts Waythomas Wei, G., andKirby,Wei, J.T., 1995, Time-dependent numerical code Sciences, v. 3,p.391-402. P.,Watts, Grilli, S.T., Kirby,G.J., J.T., andTappin, Fryer, D. phases of CALIPSO. Students from our institutions and the University of the West Indies made valuable contributions during drilthankWe the CD and IF programs NSF-EAR at and the UK NERC CIW,for UARK, and support. cost-shares. significant MV provided PSU ), Dannie Hidayat Dannie ), radii using 6 gauge files. 6 gauge using radii m 500 of Trants Average g) files 5gauge using radii TRNT from GEOWAVE simulation. for time in height (2) Young, B. Voight, S.J. Sparks, E. Shalev, S. Sacks, P. Malin, W. Johnston, D. Elswo W. P.Malin, S.Sacks, E.Shalev, Sparks, Young, S.J. B.Voight, ight, S.J. Sparks, E. Shalev, S. Sacks, P. S.Sacks, E.Shalev, Sparks, S.J. ight, Sparks, E. Shalev, S. Sacks, P. Malin, W. W. P. Malin, S.Sacks, E.Shalev, Sparks, s of tsunamis generation of by pyroclastic flows, . 03 adld snm aesuisuigaBusns mod R., 2003, Landslide tsunami case studies using a Boussinesq i, S..R Young,i, S..R B.Voight,D. o xeddBusns equations: Journal of Waterway, Port, Coas extendedfor Boussinesq Photo: Voight Photo: Voight Photo: Voight Photo: (7) , Department of Earth Sciences Sciences Earth of Department Bristol, of University (7) University Leeds Sciences, Earth Environment & Earth of School (6) University Duke Sciences, andOcean Earth of Division (5) Institution Carnegie Magnetism, Terrestrial of Department (4) Inc. Engineering, Fluids Applied (3) University State Pennsylvania Geosciences, of Department (2) Arkansas of University Geosciences, of Department (1) Malin, W. Johnston, W. Malin, D. Hidayat, D. Elswor Johnston, P.Dun D. Hidayat,Elsworth, D. Hidayat,Elsworth, P.E.Malin D. Acknowledgements Journal ofGeophysical Research References Journal of Geophysical Research which determines the energy transfer elasticanfor , E. Shalev, E. Van Boskirk, W. Johnston, R.S.J. Johnston, W. Boskirk, Van E. , E.Shalev, E. Shalev, E. Van Boskirk, W. Johnston, R.S.J. Sparks, J. Neuberg, V. Bass, P. Dunkley, R. R. Dunkley, P. V.Bass, Neuberg, J. Sparks, R.S.J. Johnston, W. Boskirk, E. Shalev, E.Van t,P uke,R ed .Nueg .Nro,adC iiiaat (2004), unpublishedrth, P. Dunkley, R. Herd, J.Neuberg, G. Norton, and C. Widiwijayanti h .Dnly .Hr,J ebr,G otn n .Wdwjyni(2004), unpublished.th, P. Dunkley, R. Herd, J.Neuberg, G. Norton, and C. Widiwijayanti orth, P. Dunkley, R. Herd, G. Norton, J.Neuberg, andC. Widiwijayanti, 2004, Prototype gma reservoirgma andconduit dynamics as revealed by aborehole geophysical observatory el and a fully nonlinear tsunami generation model: Natural Hazards and Earth System ly .Hr,J ebr,G otn n .Wdwjyni(2004), Capture of World- kley, R. Herd, J.Neuberg, G.C. Norton, Widiwijayanti and d tsunami, , 108(B12), 2563. ed Tsunami Waves, Geology. tal and Ocean Engineering, v. 121, p.251-261. amplitudes. changes in the gaugemaximum wave of the edge waves (S1 and S2) cause allIn cases, the values for thesources shoreline wave amplitudes are: parameters that significantly affect the GEOWAVEThe simulation entry Sensitivity Test Results ture article), Montserrat During Massive Dome Collapse, Masters Thesis, University of of University Thesis, Masters Collapse, Dome Massive During Montserrat •runout time •landslide volume •runout time •final depth and runout length , 108(B5), 2236. Geology linginstallation. and EOS transactions, American Geophysical Union , in press . O staff aided installation and initial operation operation initial and installation staff aided O Sparks, J. Neuberg, V. Bass, P. Dunkley, R. P. Neuberg, V.Bass, J. Sparks, d) sample 4 e) sample 5. sample e) 4 sample d) 3 sample 2c) sample 1 b) eruption. the throughout packets wave different data Table 1. Strainmeter signal. unobserved sampling of the a is strain ofthe sample Each data. eruption 2003 July 12/13 Unfiltered strainmeter. TRNT Figure 10. Spectra of was sampled at at sampled was , 85 , 317-328.