High precision relocation of earthquakes at Iliamna Volcano, Alaska
Patrick Statz-Boyer a, Clifford Thurber a,⁎, Jeremy Pesicek a, Stephanie Prejean b
a Department of Geology and Geophysics, University of Wisconsin–Madison, Madison, WI 53706, United Statesb USGS Alaska Science Center, Alaska Volcano Observatory, Anchorage, AK 99508, United States
In August 1996, a period of elevated seismicity commenced beneath Iliamna Volcano, Alaska. This activitylasted until early 1997, consisted of over 3000 earthquakes, and was accompanied by elevated emissions ofvolcanic gases. No eruption occurred and seismicity returned to background levelswhere it has remained since.We use waveform alignment with bispectrum-verified cross-correlation and double-difference methods torelocate over 2000 earthquakes from 1996 to 2005 with high precision (~100 m). The results of this analysisgreatly clarify the distribution of seismic activity, revealing distinct features previously hidden by locationscatter. A set of linear earthquake clusters diverges upward and southward from themain groupof earthquakes.The events in these linear clusters show a clear southward migration with time. We suggest that theseearthquakes represent either a response to degassing of themagma body, circulation of fluids due to exsolutionfrom magma or heating of ground water, or possibly the intrusion of new dikes beneath Iliamna's southernflank. In addition, we speculate that the deeper, somewhat diffuse cluster of seismicity near and south ofIliamna's summit indicates the presence of an underlyingmagma body between about 2 and 4 kmdepth belowsea level, based on similar features found previously at several other Alaskan volcanoes.
Iliamna Volcano is a glacially carved stratovolcano located adjacentto the Cook Inlet in south central Alaska, about 220 km from the cityof Anchorage (Fig. 1). It has a cone-shaped edifice at the north end of a5-km-long, north–south oriented ridge (Miller et al., 1998). No his-torical eruptions have been documented; however, the volcano hasseveral fumaroles that regularly emit steam and gas (Waythomas andMiller, 1999). The volcano experiences background seismicity (typi-cally 1–3 events per day) consisting mainly of shallow (b5 km) low-magnitude (bML 2) earthquakes beneath the summit region (e.g.,Dixon et al., 2005). A field study in 1999 reported evidence of at leasttwo minor eruptions in the last 300 years (Waythomas et al., 2000).Iliamna is one of six volcanoes in the Cook Inlet region known to havehad eruptions during the past 10,000 years.
In 1996 and 1997 there were two main periods of elevatedseismicity within a volume of crust beneath the edifice of Iliamna.Roman et al. (2004) hypothesized that this period of unrest, which didnot lead to an eruption, was probably caused by an episode ofmagmatic intrusion. The 1996–1997 activity was recorded by a sparsenetwork of six stations located within 15 km of the volcano's summit(Fig. 1), operated by the Alaska Volcano Observatory (AVO). Most ofthe seismicity occurred in two swarms. The first swarmoccurred in themonth of May 1996 and consisted of about 90 events. The secondswarm, consisting of over 2800 events, began in early August 1996,
er).
ll rights reserved.
peaked late thatmonth, and tapered off in early 1997. The largest eventin the swarmwasML 3.2, and the rate of detected earthquakes reacheda peak of about 75 per day during the second swarm (Fig. 2). Mostearthquakes were shallower than 6 km depth (below sea level, hereand elsewhere), and virtually all of themwere reported to be volcano-tectonic (VT) in nature with very few long-period (long period, LP) orhybrid (VT plus LP) earthquakes identified (Roman et al., 2004).Although increased emissions of CO2 and SO2 were measured during1996, no other signs of unrest were observed (Roman et al., 2004).
Roman et al. (2004) examined seismicity patterns and focal mechan-isms associated with the 1996 swarm and showed that the swarmcoincided with a period of increased degassing of CO2 and SO2. Based onthese observations, they postulated that a new dike was emplacedbeneath the southern flank of Iliamna at this time. However, the locationsused in their studyare quite scattered anddonot provide a clear picture ofthe seismic activity beneath Iliamna. Here we use waveform alignmentand hypocenter relocation using double-difference (DD) techniques toobtain much more precise locations. These methods have been success-fully applied to other Alaskan volcanoes, such as Mount Spurr, MountRedoubt, andGreat Sitkin (Brownet al., 2004;DeShonet al., 2007; Pesiceket al., 2008). At Iliamna, our precise relocations provide a clearer picture ofthe seismogenic processes beneath the volcano, and show a detailedimage of the nature of the 1996–1997 earthquake swarms.
2. Methods
In standard earthquake catalogs, seismic wave arrival-time pickshave varying accuracy largely due to event-to-event variability in signal-
Fig. 1. Station map showing the 6 closest AVO stations (triangles) and all AVO catalog events from 1994 to 2005 (dots) (Jolly et al., 2001; Dixon et al., 2002, 2003, 2004, 2005, 2006).The diamond symbol represents the summit of Iliamna. Topographic contours are shown at 500 m intervals. Inset — index map showing location of Iliamna (star) and nearby CookInlet volcanoes (triangles).
324 P. Statz-Boyer et al. / Journal of Volcanology and Geothermal Research 184 (2009) 323–332
to-noise ratios. Such errors can result in significant errors in calculatedhypocenter locations. A significant portion of this error can be eli-minatedbyaligning thewaveformsof similar events (close in spacewithsimilar mechanisms) in order to determine precise differential timesand/or adjust the arrival-time picks (Got et al., 1994; Dodge et al., 1995;Rubin et al., 1998; Waldhauser and Ellsworth, 2000; Rowe et al., 2002).
Fig. 2. Number of events recorded beneath Iliamna per day
We used the BCSEIS algorithm of Du et al. (2004a), a bispectrum (BS)verified cross-correlation (CC) method, to align waveforms for pairs ofearthquakes observed at a given station (station 1 with stations 2through 8, station 2 with stations 3 through 8, etc.). This algorithmincludes analignment stepusing cross-spectral phase forhighCCvalues,allowing for subsample precision (Poupinet et al.,1984). Standard cross-
by the AVO network from August 1996 to April 1997.
Table 1BCSEIS selection criteria values used for Iliamna.
Criterion Value Explanation
Δlim 1 sample Maximum alignment difference allowed between CC and BSCCliml 0.40 Minimum CC value considered for verificationCClimu 0.90 If CCmaxNCClimu, all delays above CCliml are verifiedCClim 0.75 If CCmaxbCClimu, then only verify stations with CCNCClim
325P. Statz-Boyer et al. / Journal of Volcanology and Geothermal Research 184 (2009) 323–332
correlation (CC) methods can produce inaccurate differential times dueto correlated noise and cycle skips. In contrast, BCSEIS uses a higherspectral order correlation technique that is less sensitive to correlated orhigh-skew noise to provide an independent estimate of the waveformalignment and hence verify the CC results. The BCSEIS method typicallyresults in more high-quality correlations than the standard CC method(Du et al., 2004b). The main control parameters for BCSEIS are (1) themaximum allowed alignment difference between the CC and BC esti-mates foracceptance (Δlim), (2) a lowCC thresholdCCliml for consideringa waveform pair for verification, (3) an upper CC threshold CClimu forverifying all waveform pairs with CC above CCliml, and (4) an inter-
Fig. 3. Original AVO catalog locations (map view and cross-sections) for the earthquakes at IliFig. 1.
mediate threshold CClim for verifying waveform pairs if the maximumCC value (CCmax) is less that CClimu. The parameter values used forBCSEIS are shown in Table 1. Once aligned, the relative arrival times forIliamna earthquakeswere obtainedwithmuch greater precision. Unlikesome Alaskan volcanoes, such as Shishaldin (Caplan-Auerbach andPetersen, 2005) and Augustine (DeShon et al., in press), we did not findevidence for repeating earthquake families at Iliamna. An example ofsimilar waveforms is shown in Appendix A Fig. A1.
We initially used the DD location algorithm hypoDD (Waldhauserand Ellsworth, 2000; Waldhauser, 2001) to relocate the Iliamna earth-quakes using Alaska Volcano Observatory (AVO) catalog data between1994 and 2005 (Jolly et al., 2001; Dixon et al., 2002, 2003, 2004, 2005,2006), with particular attention to the period of elevated activity be-tweenmid-1996andearly 1997. Inour relocationsweused theAVOone-dimensional (1D) velocity model for Iliamna Volcano (Roman et al.,2004). We successfully relocated 2041 out of 3343 events (the failuresbeing mainly “airquakes” discarded by hypoDD), with the relocationsshowing visually improved clustering. These results were encouraging,but we elected to try to improve upon them further using tomoDD(Zhang and Thurber, 2003). TomoDD has similar relocation capabilities
amna (Jolly et al., 2001; Dixon et al., 2002, 2003, 2004, 2005, 2006). Other symbols as in
Fig. 4. (a) 2014Well-located events from the tomoDD results color-coded by year, with clusters A, B, C, E, and F identified. Note that the vast majority of the events are from 1996. (b) Locations of well-located events from all of 1996 and January1997, with clusters D and F identified. Other symbols as in Fig. 1.
326P.Statz-Boyer
etal./
Journalof
Volcanologyand
Geotherm
alResearch184
(2009)323
–332
Fig. 5. TomoDD results (map view and cross-sections) for well-located events from (a) August 1996, (b) September 1996, (c) October 1996, (d) November 1996, (e) December 1996, and (f) January 1997. Other symbols as in Fig. 1.
327P.Statz-Boyer
etal./
Journalof
Volcanologyand
Geotherm
alResearch184
(2009)323
–332
Fig.
5(con
tinu
ed).
328 P. Statz-Boyer et al. / Journal of Volcanology and Geothermal Research 184 (2009) 323–332
Fig.
5(con
tinu
ed).
329P. Statz-Boyer et al. / Journal of Volcanology and Geothermal Research 184 (2009) 323–332
330 P. Statz-Boyer et al. / Journal of Volcanology and Geothermal Research 184 (2009) 323–332
as hypoDD and is designed to be used to simultaneously determineseismic wave velocity structure (i.e., seismic tomography). For reloca-tion, a key advantage of tomoDD over hypoDD is the fact that it takesstation elevations into account, which improves the ability to locateevents above sea levelwithin the volcanic edifice.We note that an effortto use tomoDD to derive a three-dimensional (3D) velocity model wasnot fruitful, which we attribute to the unfavorable earthquake-stationgeometry (Fig. 1). With most of the earthquakes in a single zone andvery few stations, there are insufficient crossing rays to provide sig-nificant resolution.
3. Results and discussion
The hypocenter relocations obtained with tomoDD provide a clearpicture of seismicity beneath Iliamna, especially for those events thatwere part of the second 1996–1997 earthquake swarm. Following acareful assessment of event location quality, we selected a threshold of10 differential times per event, which resulted in retaining 2014 of whatwe will term “well-located” events. The vast majority of these eventswere fromthe secondswarm.The chosen threshold shows thepattern ofseismicity verywell without discarding toomany events. The significantimprovement achieved in the locations of the hypocenters can be seenby comparing Figs. 3 and 4. TomoDD yielded a substantial reduction inthe root mean square residuals (RMS) for these events. The initialweighted RMS values for catalog and CC data were 0.413 s and 0.179 srespectively. After relocation, these values were reduced to 0.050 s forcatalog data and 0.007 s for CC data.
Earthquakes beneath Iliamna during the 1996–1997 swarmwere concentrated primarily in a shallow region (1–5 km depth)about 5–10 km south of the summit (Figs. 4a, 5). The northernportion ofthis group of events (Cluster A, Fig. 4a), which was most active inSeptember and October 1996, forms a diffuse cloud with little dis-cernible structure.We speculate that this feature indicates the presenceof an underlying magma body, as it is similar to seismicity featuresassociated with magma chambers at the Alaskan volcanoes Redoubt(DeShon et al., 2007), Great Sitkin (Pesicek et al., 2008), and MountSpurr (Brown et al., 2004). All of these have somewhat diffuse, dippingclusters of earthquakes overlying an inferred magma body. At Iliamna,the diffuse cluster dips approximately 50° to the south. The northernedge of Cluster A merges into a smaller cluster (Cluster B, Fig. 4a) at adepth of 1–2 km. The southern boundary of Cluster A merges intoCluster F, discussed below. Cluster B events occurred mostly in Augustand September of 1996. A horizontally elongated cluster, Cluster C, iscentered below and just to the northeast of the summit between 0 and1 km depth (Fig. 4a). Cluster C contains events that occurred almostexclusively between 1998 and 2005. This “background” seismicity maybe a result of fracturing caused by hydrothermal activity, indicated bythe fact that steam plumes are often observed at Iliamna's summit(Roman et al., 2004), or it could simply be a response to regional tectonicstress (Moran, 2003). No other signs of unusual activity have beennoticed at Iliamna since mid-1997 (McGimsey et al., 2007).
Among the most interesting features of the relocations is a systemof linear clusters (Cluster F, Fig. 4b) that emanate from Cluster A to thesouth, away from the edifice of the volcano, forming twomain groups.These clusters resemble linear features observed at other volcanoes,such as Kilauea, Hawaii, and Krafla, Iceland (Rubin and Pollard, 1987).We interpret these earthquakes to be the result of either a magmaticintrusion (Roman et al., 2004), magma degassing (D. Roman, pers.comm., 2008), or circulation of fluids due to exsolution from magmaor heating of ground water. Fig. 5 shows all the relocated eventsplotted by month, from August 1996 to January 1997. They show arelatively clear southward migration with time, away from thespeculated location of the magma body. We note that migration ofvolcanic earthquakes is not a common observation. The long timeframe for this migration (about 5 km over 6 months) suggests thatthese earthquakes may have been caused by migration of degassed
volatiles from magma, as intrusion-related migration typically occursin a shorter time period. For example, for the 1998 intrusion off theeast coast of Izu Peninsula, central Japan, migration rates are on theorder of km per day (Morita et al., 2006). In contrast, in the case of the1989 Mammoth Mountain intrusion, seismicity interpreted to be dueto exsolved gas or fluid movement above the intrusion initiallypropagated at 0.4 km permonth horizontally, increasing later to about1–2 km per month vertically (Hill and Prejean, 2005). There arecounter-examples, however, such as the long-term intrusion-relatedearthquake swarms at Unzen in 1990–91 (Nakada et al., 1999) andSpurr in 1991–92 (Power et al., 2002).
There are two small features in the relocation results that weinterpret to represent faults. Within Cluster A, associated with theinferred magma body, there is a smaller cluster of events thatoccurred exclusively in December 1996 (Cluster D, Figs. 4b and 5e).This cluster appears somewhat scattered in the map and the N–Scross-section. However, in theW–E cross-section, it appears as a thinline of events, dipping about 45° to the east. Most of the events inthis cluster occurred on the same day, with the remainder betweenDecember 20 and 26. We interpret this cluster to be a hidden fault,possibly activated by changing stresses due to pressure in themagmabody or increased pore pressure. Focal mechanisms of these eventsshow one plane parallel to the linear feature (D. Roman, pers. comm.,2008), which is not what would be expected for dike-inducedearthquakes where fault planes are typically oblique to the dike (e.g.,Hill, 1977; Karpin and Thurber, 1987). The parallelism suggests thatgas or fluids are reducing normal stress on a fault and thus triggeringseismicity. The other feature is an isolated, tight cluster of eventslocated about 3 km to the east of the magma body cluster, whichoccurred in January 1997 (Cluster E, Figs. 4a and 5f). This Januarycluster has almost the same strike as Cluster D, suggesting it too is afault. It seems unlikely that the events in Clusters D and E could havebeen directly caused by movement of magma (i.e., dike formation),due to the complete isolation of cluster E and the relatively planarcharacter of Cluster D.
4. Conclusions
This study successfully applied bispectrum-verified cross-correla-tion and DD earthquake relocation methods to relocate over 2000earthquakes occurring between 1994 and 2005 beneath IliamnaVolcano. The relocations provide a clearer picture of the spatial andtemporal patterns of this seismicity, revealing the likely location of amagma chamber beneath themain cluster of seismicity, illuminating asystem of linear clusters extending to the south, and revealing aprobable pair of faults near and to the east of the main cluster. Theseearthquakes are distinct from typical background seismicity at Iliamna,which is primarily concentrated near the summit. We infer that the1996–1997 swarm events may have been caused either by a magmadegassing or fluid exsolution event or by intrusion of magma. We alsotentatively identify the possible location of a magma reservoirunderlying a diffuse zone of earthquakes near Iliamna's summit,based on similar features seen at other Alaskan volcanoes. Our analysisapproach should be useful in studying other volcanic earthquakeswarms to investigate their probable cause.
Acknowledgements
We gratefully acknowledge the staff of the Alaska VolcanoObservatory, especially Jim Dixon, for their long-term monitoringefforts and for providing access to their waveform archive. We thankDiana Roman and Jackie Caplan-Auerbach for reviews of an earlierversion of this manuscript, and two anonymous reviewers for theirconstructive comments. Numerous figures were generated using GMT(Wessel and Smith, 1991). This material was based on research sup-ported by NSF grant EAR-0409291.
331P. Statz-Boyer et al. / Journal of Volcanology and Geothermal Research 184 (2009) 323–332
Appendix A
Fig. A1 shows an example of similar waveforms for earthquakes that occurred in January 1997. They are part of Cluster F identified in Fig. 4b. CCcoefficients between the first trace and the other four traces for the P waves range from 0.94 to 0.98.
Fig. A1. Example of similar waveforms for earthquakes in Cluster F (see Fig. 4b).
References
Brown, J., Prejean, S.G., Zhang, H., Power, J., 2004. Double difference earthquakerelocation and tomography at Mount Spurr volcano, Alaska, 1991 to 2004. EOSTrans. AGU 85 Fall Meet. Suppl., Abstract S51A-0140.
Caplan-Auerbach, J., Petersen, T., 2005. Repeating coupled earthquakes at ShishaldinVolcano, Alaska. J. Volcanol. Geotherm. Res. 145, 151–172.
DeShon, H.R., Thurber, C., Power, J. in press. Earthquakewaveform similarity and evolutionat Augustine volcano, Alaska, from 1993–2006. U. S. Geol. Surv. Prof. Paper.
Dixon, J.P., Stihler, S.D., Power, J.A., Tytgat, G., Estes, S., Moran, S.C., Paskievitch, J.,McNutt, S.R., 2002. Catalog of earthquake hypocenters at Alaskan volcanoes:January 1, 2000 through December 31, 2001. U. S. Geol. Surv. Open-File Report OF02-0342. 56 pp.
Dixon, J.P., Stihler, S.D., Power, J.A., Tytgat, G., Moran, S.C., Sanchez, J., Estes, S., McNutt, S.R.,Paskievitch, J., 2003. Catalog of earthquake hypocenters at Alaskan volcanoes:January 1 through December 31, 2002. U. S. Geol. Surv. Open-File Report OF 03-0267.58 pp.
Dixon, J.P., Stihler, S.D., Power, J.A., Tytgat, G., Moran, S.C., Sanchez, J.J., McNutt, S.R.,Estes, S., Paskievitch, J., 2004. Catalog of earthquake hypocenters at Alaskanvolcanoes: January 1 through December 31, 2003. U. S. Geol. Surv. Open-File ReportOF 2004-1234. 69 pp.
Dixon, J.P., Stihler, S.D., Power, J.A., Tytgat, G., Estes, S., Prejean, S., Sanchez, J.J., Sanches,R., McNutt, S.R., Paskievitch, J., 2005. Catalog of earthquake hypocenters at Alaskanvolcanoes: January 1 through December 31, 2004. U. S. Geol. Surv. Open-File Report2005-1312. 74 pp.
Dixon, J.P., Stihler, S.D., Power, J.A., Tytgat, G., Estes, S., McNutt, S.R., 2006. Catalog ofearthquake hypocenters at Alaskan volcanoes: January 1 through December 31,2005. U. S. Geol. Surv. Open-File Report 2006-1264. 78 pp.
Dodge, D.A., Beroza, G.C., Ellsworth,W.L., 1995. Evolution of the 1992 Landers, Californiaforeshock sequence and its implications for earthquake nucleation. J. Geophys. Res.100, 9865–9880.
Du, W., Thurber, C.H., Eberhart-Phillips, D., 2004a. Earthquake relocation using cross-correlation time delayestimates verifiedwith the bispectrummethod. Bull. Seismol.Soc. Am. 94, 856–866.
Du, W., Thurber, C.H., Reyners, M., Eberhart-Phillips, D., Zhang, H., 2004b. Newconstraints on seismicity in the Wellington region, New Zealand, from relocatedearthquake hypocenters. Geophys. J. Int. 158, 1088–1102.
Got, J.-L., Frechet, J., Klein, F., 1994. Deep fault plane geometry inferred from multipletrelative relocationbeneath the southflankofKilauea. J. Geophys. Res. 99,15,375–15,386.
Hill, D.P., 1977. A model for earthquake swarms. J. Geophys. Res. 82, 1347–1352.Hill, D.P., Prejean, S.G., 2005. Magmatic unrest beneath Mammoth Mountain, California.
A.D., Moran, S.C., McNutt, S.R., Hammond, W.R., 2001. Catalog of earthquakehypocenters at Alaskan volcanoes: January 1, 1994 through December 31,1999. U. S.Geol. Surv. Open-File Report OF 01-0189. 22 pp.
Karpin, T., Thurber, C.H., 1987. Relationship between earthquake swarms and magmatransport: Kilauea volcano, Hawaii. Pure Appl. Geophys. 125, 971–991.
McGimsey, R.G., Neal, C.A., Dixon, J.P., 2007. 2005 Volcanic activity in Alaska andKamchatka: summary of events and response of the Alaska Volcano Observatory.U. S. Geol. Sur. Open-File Report 2007-5269. 93 pp.
Miller, T.P., McGimsey, R.G., Richter, D.H., Riehle, J.R., Nye, C.J., Yount,M.E., Dumoulin, J.A.,1998. Catalog of the historically active volcanoes of Alaska. U.S. Geol. Surv. Open-FileReport OF 98-0582. 104 pp.
Moran, S.C., 2003. Multiple seismogenic processes for high-frequency earthquakes atKatmai National Park, Alaska: evidence from stress tensor inversions of fault-planesolutions. Bull. Seismol. Soc. Am. 93, 94–108.
Morita, Y., Nakao, S., Hayashi, Y., 2006. A quantitative approach to the dike intrusionprocess inferred from a joint analysis of geodetic and seismological data for the1998 earthquake swarm off the east coast of Izu Peninsula, central Japan. J. Geophys.Res. 111, B06208. doi:10.1029/2005JB003860.
Nakada, S., Shimizu, H., Ohta, K., 1999. Overview of the 1990–1995 eruption at UnzenVolcano. J. Volcanol. Geotherm. Res. 89, 1–22.
Poupinet, G., Ellsworth, W.L., Frechet, J., 1984. Monitoring velocity variations in thecrust using earthquake doublets: an application to the Calaveras fault, California.J. Geophys. Res. 89, 5719–5731.
Power, J.A., Jolly, A.D., Nye, C.J., Harbin, M.L., 2002. A conceptual model of the MountSpurr magmatic system from seismic and geochemical observations of the 1992Crater Peak eruption sequence. Bull. Volcanol. 64, 206–218.
Roman, D.C., Power, J.A., Moran, S.C., Cashman, K.V., Doukas, M.P., Neal, C.A., Gerlach,T.M., 2004. Evidence for dike emplacement beneath Iliamna volcano, Alaska in1996. J. Volcanol. Geotherm. Res. 130, 265–284.
Rowe, C.A., Aster, R.C., Borchers, B., Young, C.J., 2002. An automatic, adaptive algorithm forrefining phase picks in large seismic data sets. Bull. Seismol. Soc. Am. 92, 1660–1674.
332 P. Statz-Boyer et al. / Journal of Volcanology and Geothermal Research 184 (2009) 323–332
Rubin, A.M., Pollard, D.D., 1987. Origins of blade-like dikes in volcanic rift zones. In:Decker, R.W., Wright, T.L., Stauffer, P.H. (Eds.), Volcanism in Hawaii, USGS Pro-fessional Paper 1350, pp. 1449–1470.
Rubin, A., Gillard, D., Got, J.-L., 1998. A re-examination of seismicity associated withthe January 1983 dike intrusion at Kilauea volcano, Hawaii. J. Geophys. Res. 103,10,003–10,015.
Waldhauser, F., 2001. hypoDD — a program to compute double-difference hypocenterlocations. U.S. Geological Survey Open-File Report 01-113.
Waldhauser, F., Ellsworth, W.L., 2000. A double-difference earthquake location algo-rithm; method and application to the northern Hayward fault, California. Bull.Seismol. Soc. Am. 90, 1353–1368.
Waythomas, C.F., Miller, T.P., 1999. Preliminary volcano-hazard assessment for Iliamnavolcano, Alaska. U.S. Geol. Surv. Open-File Report 99-373.
Waythomas, C.F., Miller, T.P., Beget, J.E., 2000. Record of late Holocene debris avalanchesand lahars at Iliamna volcano, Alaska. J. Volcanol. Geotherm. Res. 104, 97–130.
Wessel, P., Smith, W.H.F., 1991. Free software helps map and display data. EOS Trans.AGU 72, 441.
Zhang, H., Thurber, C.H., 2003. Double-difference tomography; the method and itsapplication to the Hayward fault. California. Bull. Seismol. Soc. Am. 93, 1875–1889.
