S51A-0125 0800h
Focal mechanisms from regional earthquakes in East Africa and refinements in stress patterns
Accurate event locations for eleven events with magnitude greater than 3.5 in Tanzania, East Africa have been obtained using data from the Tanzania Broadband Seismic Experiment. Focal mechanisms for most of these events have been determined by modeling P and SH polarities and amplitude ratios in a grid search method. Epicenters have been constrained by using P and S arrival times. Focal depths have been constrained by waveform modeling of regional and local depth phases. Most of the earthquakes occur in two areas along the western or eastern branch of the East African Rift. The first area is located between southern Lake Tanganyika and Lake Rukwa in the western branch, and the mechanisms in this region are characterized by strike-slip or oblique extension with nodal planes striking E-W to NW-SE. These events confirm that this part of the western rift is under transtension, as compared to the Lake Tanganyika and Rukwa rifts, where the stress field appears to be almost purely extensional. The second area for which we have obtained several new focal mechanisms is near the southern end of the eastern branch where the rift impinges on the eastern margin of the Tanzania craton. The focal mechanisms from this part of the rift have strike slip or normal motions but the orientations of the nodal planes are complex, consistent with the complicated orientation of the block faults. The focal depths for the events in the first area vary from 7 to 36 km, and in the second area, they vary from 11 to 34 km.
S51A-0126 0800h
Focal Mechanisms Of Recent Small To Moderate Sized Earthquakes And Seismogenic Faulting In The Korean Peninsula
We estimate faulting geometry of 88 small to moderate sized earthquakes since 2001 in and around the Korean Peninsula using polarities of first-arrival P-waves recorded by the local seismic network in Korea. Overall pattern of the inverted solutions indicates a combination of thrust and strike-slip faulting characterized by nodal planes striking in the WNW-ESE direction although some biases for small events exist owing to the local heterogeneity of stress field. We analyze the stress fields acting in and around the Korean Peninsula by inverting the compressional and tensional axes of the focal mechanisms. The results show a compressional regime with maximum stress in the NE-SW direction, coinciding with the direction of compression derived from geological information. Some moderate earthquakes show misfits of nodal planes and slip vectors with respect to the stress tensor. In particular, the 29 May 2004, Mw=5.1, offshore Uljin, Korea, earthquake shows the nearly north-trending nodal plane with thrust faulting motion. This indicates that strain orientation of earthquakes occurring in and around the Korean Peninsula varies significantly owing to the orientations of preexisting faults.
S51A-0127 0800h
Focal Mechanisms From Moment Tensor Solutions and First Motion Polarities of Shallow to Deep Local Earthquakes in Eastern Nepal and Southern Tibet
We determined focal mechanisms using waveforms and first motion polarities from local earthquakes recorded during the Himalayan Nepal Tibet Seismic Experiment (HIMNT). The HIMNT experiment included the deployment of 28 broad band seismometers in eastern Nepal and southern Tibet from September 2001 to April 2003. Using a regional moment tensor method (Ammon and Randall, 2001) and first motion polarities for displaying double-couple focal mechanisms (Snokes, 2003), we analyzed the fault plane solutions at three distinct zones of seismicity. Characteristic focal mechanisms in seismically concentrated areas may indicate the presence of fault ramps or a decollement in the Himalayan collision zone. Previous studies of focal mechanisms on the Tibetan Plateau predominantly indicate east-west extension and shallow thrusting at the Himalayan collision zone for shallow to intermediate earthquakes (Ni and Barazangi, 1984; Molnar and Lyon-Caen, 1989; Randall et al., 1995) and east-west extension for intermediate to deep earthquakes (Zhu and Helmberger, 1996; Chen and Yang, 2004). The first zone in southeast Nepal between the Main Boundary and Main Frontal faults consist of earthquakes < Mw 4.0 at depths 40 - 60 km near the epicenter of the 1988 Udaypur earthquake, Mb 6.1, depth 57 km. The second zone north of the Main Central Thrust outcrop in eastern Nepal consists of 14 earthquakes 3.0 - 5.0 Mw at depths < 30 km that indicate north-south strike normal faulting and east-west strike thrust faulting. The third zone is an arc parallel to the Himalayas in southern Tibet and a cluster in northeast Nepal. This zone consists of 45 earthquakes < 4.0 Mw at depths > 50 km. Four earthquakes indicate northwest-southeast compression resulting in northeast strike strike-slip faulting while one earthquake in the northeast cluster indicates east-west compression at a source depth below the crust-mantle boundary. Focal mechanisms from full waveform moment tensor inversions are cross checked with first motion solutions for selected events. Source depths as determined from normalized error of the sum of the squared differences between the data and synthetic seismogram coincide with the source depths determined from the travel time residual inversion.
S51A-0128 0800h
the Cataract Creek, Arizona, 1993 Earthquake Sequence
The 1993 Cataract Creek sequence was the most extensively documented seismic event in Arizona history. The sequence consisted of an mb 4.8 foreshock, followed four days later by the mb5.5 mainshock and numerous aftershocks. Analysis of the foreshock and mainshock indicated that both were normal faulting events with WNW strikes (Lay et.al.,1994). The aftershock analysis supports activity on a NW trending surface for the main shock, parallel to trends of faults of the Cataract Creek system (Sanders, 1997). Examination of the results of both previous studies suggests two further points. The distance between the foreshock and aftershock indicates that they are on separate fault surfaces; and the stress release throughout the sequence was spatially heterogeneous. The foreshock and its aftershocks clustered near the south part of the region of stress release (April 25-29). The April 29 mainshock occurred 28 kilometers to the north of the foreshock followed over the next four days by aftershocks clustered close to the mainshock. The four largest aftershocks from April 25 to May 17 occurred in the southern part of the region of stress release closer to the mb4.8 foreshock. The spatial separation of the mainshock and foreshock and their aftershocks suggests that they may represent two mainshocks of different sequences on separate faults. This would explain the difference in nodal trends of the two shocks.
S51A-0129 0800h
Microseismicity Studies in Northern Baja California, Mexico: the Agua Blanca Fault.
The Agua Blanca fault is major, conspicuous, well-defined, right-lateral strike-slip fault with a minor vertical component that extends from the Punta Banda Peninsula near Ensenada, in northern Baja California, to the San Matias Pass on the Main Gulf Escarpment. Regional catalogs from both the SCSN and RESNOM (ML $>$ 2.1) show an essentially inactive fault; however, some minor seismicity in the Punta Banda Peninsula and a clustered activity in Valle de la Trinidad, near the NW and SE ends of the fault, in addition of some scattered activity in between could be associated with it. It has been suggested that the Agua Blanca fault was originally related with the subduction of the Farallon plate. Besides, the fact that the fault cuts granitic intrusives together with results from trench studies indicate that the fault might be currently active. A detailed microseismicity survey of the Agua Blanca fault was carried out during two months in 2001 with 41 Reftek stations. About 140 very small events (ML $<$ 2.1) were located near the fault; most of them (103, depths mostly between 8 and 13 km, strongly clustered) are associated with Valle de la Trinidad. Less than 20 epicenters with depths mostly between 8 and 16 km, are less than 3 km from the fault surface trace, all of them at its central part. Focal mechanism determinations show mainly strike slip and normal solutions indicating a tensional regime, with a near N-S compressional axis, a result which agrees with those of our previous microseismicity studies in northern Baja California. Apart from earthquakes at the Valle de la Trinidad and a few epicenters located very near to the fault trace, the rest of them is scattered to the south of the Agua Blanca fault and in between this fault and the San Miguel fault system. These results indicate that there is important microseismic activity which is not associated with mapped faults in northern Baja California. In conclusion, our two month-long survey extends the picture of a relatively inactive (ML $>$ $\sim$0.8) Agua Blanca fault given by data from regional seismicity catalogs.
S51A-0130 0800h
Characterization and Relocation of Seismic Clusters in the Area of Bahia de Banderas, Jalisco-Nayarit, Mexico
We analysed the seismic activity that took place the year of 2003 in the area of Bahia de Banderas, between the states of Jalisco and Nayarit, registrated with a local network of 7 stations, which belongs to the Civil Defence of Jalisco and the University of Guadalajara. 400 events have been located, in these earthquakes we identified some series of a similar waveforms. For defining this similarity between seismic events and in order to classify them into clusters, we have applied the cross-correlation method of the P and S arrivals. We found a fourth part of epicentres gathered into 15 clusters of 3-25 events. For some clusters we used relocations relative to a master event. Located south of Bahia de Banderas exist clusters aligned along structures trending N-S in the area of Tuito. This trend agrees with the topographic relief of the area. Other clustes can be related with active tectonic structures at north of Cajon de Peñas dam (Tomatlan). Another cluster was identified at the East, Amatlan de Cañas-Ameca area, and one more in the center of the Bahia de Banderas.
S51A-0131 0800h
Seismic Investigation of Recent Events in the Mendocino Triple Junction Region: Stress Release and Orientation within the Gorda Plate and along the Mendocino Fault
The Mendocino Triple Junction (MTJ) is the intersection of the San Andreas fault zone, Mendocino fault, and Cascadia subduction zone. This region has a history of strong earthquakes of magnitude 6.5 or greater, with much of the seismicity within the Gorda plate. We investigate recent and historic earthquakes in the MTJ region. Data for this study has been collected as digital data downloaded from the North California Earthquake Data Center (NCEDC) for the Berkeley Digital Seismic Network (BDSN) for years 1992 to 2002. We modeled each event for focal mechanism and total moment. Our results have been compiled with larger events with known focal mechanisms and moment from the Harvard Centroid Moment Tensor (CMT) and NCEDC catalogs to reveal the overall state of stress for the region. We divide our events by magnitude and class of focal mechanism (thrust, strike slip, transpressive and normal) and plot the direction of moment as a function of the P-axis. We find that the larger strike-slip events in the Gorda plate align with a P-axis that might be expected for the San Andreas fault, smaller events align with an expected Mendocino fault orientation, and very few events have an orientation related to the Cascadia subduction zone. These results suggest that Gorda plate deformation is not relieving stresses associated with the Cascadia subduction zone, and that the state of stress release for the entire region is dominated by the Mendocino fault.
S51A-0132 0800h
Three-dimensional Geology of the Hayward Fault and its Correlation with Fault Behavior, Northern California
Relationships between fault behavior and geology along the Hayward Fault were investigated using a three-dimensional geologic model of the Hayward fault and vicinity. The three-dimensional model, derived from geologic, geophysical, and seismicity data, allowed the construction of a `geologic map' of east- and west-side surfaces, maps that show the distribution of geologic units on either side of the fault that truncate against the fault surface. These two resulting geologic maps were compared with seismicity and creep along the Hayward Fault using three-dimensional visualization software. The seismic behavior of the Hayward Fault correlates with rock unit contacts along the fault, rather than in rock types across the fault. This suggests that fault activity is, in part, controlled by the physical properties of the rocks that abut the fault and not by properties of the fault zone itself. For example, far fewer earthquakes occur along the northern part of the fault where an intensely sheared Franciscan melange on the west side abuts the fault face, compared to the region to the south where more coherent rocks of other Franciscan terranes or the Coast Range Ophiolite are present. More locally, clusters of earthquakes correlate spatially with some of the contacts between Franciscan terranes as well as mafic rocks of the Coast Range Ophiolite. Steady creep rates along the fault correlate with the lateral extent of the San Leandro gabbro, and changes in creep rate correlate with changes in geology. Although preliminary, the results of comparing fault behavior with the inferred three-dimensional geology adjacent to the Hayward Fault suggest that any attempt to understand the detailed distribution of earthquakes or creep along the fault should include consideration of the rock types that abut the fault surface. Such consideration would benefit greatly from incorporating into the three-dimensional geologic model the physical properties of the rock types along the fault.
S51A-0133 0800h
Heterogeneous Crustal Structure in the Focal Area of the 2000 Western Tottori prefecture Earthquake
On October 6, 2000, a large earthquake (Mw 6.6) occurred in the western part of Tottori prefecture, Honshu, Japan. Although the earthquake is a large intra-island-arc type event, a surface rupture is not clearly observed. The generation mechanism of such an intra-island-arc earthquake has not been clarified yet. The heterogeneous crustal structure is one of the key points to understand these earthquakes. To reveal the heterogeneous crustal structure in the fault area of this earthquake, we conducted earthquake observations in twice. In October 2000, right after this large earthquake, an urgent aftershock observation was carried out (Shibutani et al., submitted to EPS). As a part of this, we deployed a multi-channel seismic (MCS) array along and across the main fault area. The MCS array was operated for 85 hours to obtain quasi-continuous records of aftershocks, and 269 events were recorded by the MCS array. In April 2002, a seismic reflection survey was carried out along the fault (Nishida et al., 2002). At the same time, we conducted a seismic array observation using off-line seismic recorders in the same area to image the crustal structure. The array consisted of 145 seismometers distributed across a southeastern part of the main fault. The array was operated for about one month and 259 events were recorded. The common midpoint (CMP) reflection method is widely used to image the heterogeneous crustal structures. The sources and receivers are located at or near by the surface in the CMP method. Since the natural earthquakes are not located on the surface of the earth, we cannot use the usual CMP method. Therefore, we proposed a new method, the natural earthquake reflection profiling (NERP) method, to image the crust using natural earthquakes. By the NERP method, we calculate a common reflection point for sources in the subsurface and receivers on the surface to transform original data into zero-offset depth section. This method strongly depends on the hypocenter and origin time of earthquakes and the background velocity structure. To estimate these parameters accurate enough, the Joint Hypocenter Determination method was applied. We applied the NERP method to the observed seismic data by assuming both PP and SS reflections. Since both PP and SS profiles image the reflectors at the same depth, we can interpret that the reflectors are not ghost reflection images but real reflectors. The results are follows: Above a depth of 5 km no image is found because of few aftershocks in a shallow depth. Between depths of 5 and 9 km are several reflectors. Between depths of 9 and 14 km, the reflectors are not clearly found. At the depth of greater than 14 km, many clear reflectors are imaged. In the depth range between 5 and 9 km, some less-reflective zones are found. A very small slip occurred on the fault plane during the main shock in these less-reflective zones. Compared with the aftershock distribution and recent swarm-like seismic activities in 1989, 1990 and 1997, no earthquake occurred in these less-reflective zones. The P wave velocity in these less-reflective zones is slightly slower than that in the neighbor area. Those characteristics indicate that brittle fracture does not occur in the less-reflective zone but stably slip deformation is dominant there. It is possible that the stably slip areas are distributed in the upper crust, which may contribute the major characteristics of the 2000 Western Tottori prefecture earthquake.
S51A-0134 0800h
Lower Crust and Upper Mantle Structure in the Region of the 2000 Western Tottori, Japan Earthquake Estimated from Receiver Function Analyses
In this study, we image the P to S converted structure in the lower crust and upper mantle by receiver function (RF) analyses, for the source region of the 2000 Western Tottori, Japan earthquake (Mw 6.6). We use 21 teleseismic events recorded on a dense array of 54 stations that were deployed for aftershock studies following the main shock. The deconvolution technique used RF analyses essentially removes P waves, except for the direct arrival, and leaves conversions and reverberations of P to S types for a plane layered structure. We used data low-passed filtered at about 1 Hz. RF analyses usually use 1-D structures to determine the conversion depths along the assumed ray path of the incident P wave. In this study, however, we could not obtain a clear image by using this method. This is because that the target conversion waves propagate along different ray paths from that of P wave in a complex structure. Therefore, we estimate the arrival direction of the P coda, including the direct P wave, and the P to S converted waves, using a polarization analysis with multi-taper method (Park et al., 1987a). These results show that the arrival directions of P to S converted waves have variances of more than 20 degrees from that of P wave direction, indicating a strong influence from the 3-D velocity structure on the arrival directions of the converted waves. We estimated arrival directions of the converted phases from many teleseismic events over a large azimuthal range of incoming P waves. After we use this information to construct the 3-D surface where the P to S conversions occur near the recording station, we transform the time domain RF into depth domain one. Using this method with data from closely spaced stations, we can produce a clearer image of the upper mantle structure. The results show a surface that is located below the Moho discontinuity at depths of about 40 to 60 km and dipping toward the north, in the vicinity of the rupture area of the 2000 Western Tottori Earthquake. Recently, it has been suggested that there is particular deformation in the lower crust and upper mantle under active faults that cause large inland earthquakes. The structure imaged in this study may be related to these movements and important for understanding the preparatory processes of large crustal earthquakes.
S51A-0135 0800h
Three-dimensional crustal image in the focal region of the 2000 western tottori earthquake using local earthquake and active source data
P and S arrival time data from aftershocks of 2000 western tottori earthquake and P arrival time data from active sources or shots have been used in hybrid travel time tomography for hypocenters, three-dimensional Vp and Vp/Vs image in the focal region.The number of aftershocks used in tomography was 1,091 and the number of shots was several hundred. Tomographic image covered a 32 km x 40 km x 16 km area. The grid spacing was flexibly decided according to the density of data. As the results, velocity anomalies were detected in the focal region. Low velocity anomaly that exists around 2 km along the seismic fault would show the activity of the fault in a past. Another anomalies that exist around hypocenter might correlate with asperity.
S51A-0136 0800h
Dense passive seismic survey of Mt. Fuji, Japan during 2002-2004
Mt. Fuji, the most famous and highest volcano in Japan, has a larger eruption rate than most other island-arc volcanoes by one order of magnitude. Mt. Fuji has erupted mostly basaltic products, although it is an island-arc type volcano. These features may be due to the unique tectonic setting of Mt. Fuji, which is located near three converging plates, including the Philippine Sea (PHS) plate. The volcanic front and plate boundary cross near Mt. Fuji. In addition, the region around Mt. Fuji is a zone of crustal collision where the Izu block collides with the Honshu block. The existence of the PHS plate beneath Mt. Fuji has not been determined. Low frequency (LF) earthquakes intensively occurred in September - December 2000 and April - May 2001 at depths of 10 - 20 km beneath Mt. Fuji. However, no other associated volcanic unrest was observed at that time. A passive seismic survey of Mt. Fuji has been conducted to clarify the source mechanisms of the LF earthquakes and to conduct 3D seismic imaging of the subsurface structure. A dense seismic network was deployed around Mt. Fuji in September - December 2002 by scientists and technicians from 6 national universities of Japan. The network includes 28 temporary and 138 permanent seismic stations. Temporary seismic stations are equipped with three-component seismometers including 6 CMG-3T, 7 Trillium and 15 L4-3D. Half of permanent stations are as a part of the Hi-net of NIED. The seismic stations are deployed 2 - 10 km apart with a gradual increase in station spacing away from Mt. Fuji and also densely distributed in the direction NE-SW across Mt. Fuji. We will show an outline of the passive seismic survey of Mt. Fuji and preliminary results of velocity tomography beneath Mt. Fuji. The between the Izu and Honshu blocks, the PHS plate, and volcanic brief results are as follows: (1) High-frequency seismic activity is clearly seen in several regions, including zones of collision activity of the Izu Peninsula; (2) LF earthquakes are also captured with the network and located at depths of 12 - 16 km beneath Mt. Fuji; (3) A shallow high velocity anomaly is seen at depths of 0 - 5 km beneath Mt. Fuji; (4) A low velocity anomaly is seen at depths of 10 - 20 km beneath Mt. Fuji, corresponding to the LF locations.
S51A-0137 0800h
Spatiotemporal Velocity Changes Around Miyake and Kozu Islands, Central Japan in June,2000 - May,2001
Spatiotemporal velocity changes have been found around Miyake and Kozu Islands, central Japan in June,2000 - May,2001 from seismic tomography method applied to 694,345 arrival times at 122 sites in and around Miyake and Kozu Islands including ocean bottom observations. Due to the high irregularity in the uppermost crust and the difficulties of handling both of the data of seabed and the land simultaneously, station correction is adopted. The arrival time data is divided eleven periods so as to examine the temporal velocity change, taking into account the hypocenter distribution map. We determine Vp,Vs models in each period applying the seismic tomography method. The result indicates that there are mainly two low velocity zones which locate in the west of Miyake Island and the east of Kozu Island and they change temporally their intensity corresponding hypocenter distribution. In the early period ( ~ July 5 ), low velocity zone (LVZ) is limited at Miyake site and that suggests magma is supplied form Miyake Volcano. Next period (July 6 - July 20), LVZ of Miyake site decreases and in reverse that of Kozu site appears with seismic swarm. In the period (July 21 - Aug 14), LVZ of Kozu site is very powerful and expands up to 5km depth with great swarm. This suggests that the new magma intrusion occurs from deeper between Miyake and Kozu Islands. Next, in the period (Aug 15 - Aug 31), LVZ of Miyake site increases with swarm and that indicates the magma flow form Miyake volcano coming again. After that, in the period (Sept.1,2000 - May 6,2001), both LVZs decrease their rates gradually and the seismic activity decrease either. We surmise that those low velocity zones correspond magma intrusion and spatiotemporal changes of magma intrusions cause repeating seismic immigrations between Miyake and Kozu Islands.
S51A-0138 0800h
Structure and Stresses Near Yucca Mountain, Nevada From Tomography and Focal Mechanisms
The local and regional tectonic setting of Yucca Mountain, Nevada, the designated nuclear waste repository, is characterized in geologic, neotectonic, seismological, and geodetic investigations. Although earthquake focal mechanisms have been compiled for M$>\sim$2 events for a number of years, the details and variability of the local stress field as well as the crustal velocity structure in the region is generally poorly understood. In an effort to address this, we have compiled over 1000 focal mechanisms from earthquakes that have occurred around Yucca Mountain since the early 1980's to define the regional stress field in the vicinity of the site. From 1976 to 1992, few earthquakes were observed with magnitudes over M3. However, after the Little Skull Mountain earthquake (06/29/1992 M5.6) and associated aftershocks, several earthquakes above M4 have occurred within 65 km of Yucca Mountain. Nonetheless, the vast majority of earthquakes, especially within 10 km of Yucca Mountain proper, are of M$<$1, where first-motion focal mechanisms are difficult, if not impossible, to develop even in this highly instrumented area. Including small events is crucial to our evaluation of the local stress field in and around the mountain block near the proposed repository site. We have developed a method that determines focal mechanisms and their associated statistical error bounds for events from a user-controlled combination of available first-motions and P, SV, and SH amplitudes. In order to resolve local variations in the stress field as best as possible, we have relocated the events using the hypoDD program. In addition to determining the stress field, we have imaged the 3-D velocity structure within 100 km of the mountain to approximately15 km depth using a non-linear iterative tomographic inversion applying both controlled-source and earthquake travel times. We have collected over 150,000 P and S arrival times from local earthquakes from 1980 through the present and over 4,000 first-arrivals from underground nuclear tests and the 1993 Non-Proliferation Experiment. Controlled-source locations and origin times are precisely known. The tomographic inversion simultaneous solves for optimal earthquake locations (3-D position and origin time) as well as for smooth P and S velocity structures. We compare locations determined in the joint hypocentral inversion result with relocated hypocenters applying standard location routines. Preliminary results demonstrate the commonly observed Great Basin result of a nearly 1-D, well-behaved, structure below about 3 km depth. The upper 3 km indicate a large-scale low-velocity feature that roughly corresponds to a trough (low) in the regional gravity image which includes Yucca Mountain. We will discuss the tectonic implications of the resolved stress field and structure.
S51A-0139 0800h
Precise relocation of earthquakes preceding the March 9, 1998 eruption of the Piton de la Fournaise volcano
The March 9, 1998 eruption of the Piton de la Fournaise volcano (R\'eunion island) was preceded by a seismic crisis unusually long for this volcano. It lasted for more than 35 hours and included about 3100 micro-earthquakes with a maximum magnitude of 2.2. During the crisis, the events migrated slowly and monotonically from a depth of about 5 km below sea level to the surface, leading to the onset of the eruption. The arrival times of 1001 events have been determined using waveform similarity and a set of reference events manually picked. This technique allows to greatly increase the number of located events by including the small magnitude ones and also reduces the scattering related to the inconsistency in the manual phase picking. The program hypo71 has been used with a 1D velocity model to obtain preliminary locations. From about 5 km b.s.l. to sea level, the obtained locations define an almost continuous narrow vertical pass with a horizontal extend of the order of a few hundred meters (about 200-400 meters). Above sea level the earthquakes do not indicate any clear migration. In order to further improve the precision of the locations, and define eventual active structures, we searched for multiplets among the located events. A total of about 770 earthquakes can be grouped into 16 multiplets of more than 5 events centered at different depths along the migration path, some of them including events with reversed focal mechanisms. The application of precise relative relocation techniques using waveform cross-correlation differential times enlightens several streaks or planar features with various orientations. For some of the multiplets, clear upward migration of the earthquakes is observed along the defined features.
S51A-0140 0800h
Double Difference Earthquake Relocation and Tomography at Mount Spurr Volcano, Alaska, 1991 to 2004
Mount Spurr, one of the northernmost active volcanoes monitored by the Alaska Volcano Observatory, is located approximately 120 km west of Anchorage in the Cook Inlet region. Mount Spurr erupted three times in 1992. In the subsequent decade the volcano was relatively quiet. However, seismicity rates began increasing again in early 2004, indicating that the volcano is entering a new stage of unrest. In order to gain a better understanding of the temporal and spatial distribution of seismicity at Mount Spurr, we relocated earthquakes from 1991 through 2004 using the double difference relocation code hypoDD (Waldhauser and Ellsworth, 2000). We also solved for the P-wave velocity structure of Spurr with the same dataset using the adaptive grid double difference tomography algorithm, tomoADD (Zhang and Thurber, 2004), in order to image the internal structure of the volcano and evaluate the applicability of this technique in a volcanic setting. Even though the seismic array at Mount Spurr is relatively sparse (10 short period stations) earthquake relocations result in significantly more precise hypocenters. The hypocenters from the 1992 eruptive sequence define a conduit-like structure extending from the Crater Peak vent (~2 km above sea level) to over 20 km depth, dipping steeply to the southeast. Earthquakes under the summit of Spurr (~3 km above sea level), whose vent has been quiet historically, are more scattered epicentrally and generally restricted to depths of $<$ 5 km below sea level. Using tomoADD we imaged a low velocity body beneath the Crater Peak vent from depths of roughly 1 km above sea level to up to 10 km below sea level. This low velocity body is co-located with the steeply dipping 1992 seismicity lineation defined by relocations and likely reflects the magmatic conduit and/or the presence of a hydrothermal system southeast of Crater Peak. The spatial distribution of low velocity bodies determined using tomoADD is consistent with that determined previously by more traditional finite difference tomography techniques (Power et al., 1994). Both hypoDD and tomoADD appear to be useful methods even in relatively sparsely instrumented volcanic settings.
S51A-0141 0800h
Stress Field Along the Central Denali Fault, Alaska, From Relocated Aftershocks of the 2002 Earthquake Sequence
The Mw 7.9 November 2002 Denali Fault Event was one of the largest intra-continental earthquakes in recent recorded history. More importantly, no such major strike-slip earthquake has been as well-recorded by modern seismic instruments. Augmented by data from temporary stations, we located aftershocks of this event with a high accuracy. By reviewing waveforms and revising previous arrival picks and then applying a double-difference relocation algorithm (Waldhauser and Ellsworth, 2000), we have created the most up-to-date catalog of aftershock locations for the larger aftershocks (M $>$ 3.4). By calculating focal mechanisms, we have shown characteristics of faulting along the ruptured faults. We have identified subsidiary faults just east of the Richardson highway, both north and south of the Denali Fault striking in a similar orientation as the Fault, that exhibit reverse faulting. By implementing the program ZMAP (Wiemer, 2001) we mapped the stress directions across the extent of the rupture. While predominantly a strike-slip rupture, the stress mapping shows some regions of reverse faulting interchanged with regions of strike-slip motion. Furthermore, the orientations of the stress fields exhibit no change across the fault. Where the Talkeetna, Hines Creek, and Denali Faults converge, the largest notable variance in the stress directions occur due to complex interactions with the Hines Creek and Talkeetna Faults as the Denali Fault ruptured.
S51A-0142 0800h
Characterizing Local Seismicity of the Bhutan Himalaya: An Analysis of Earthquakes Recorded With a Temporary Seismic Network During 2002-2003
The University of Texas at El Paso (UTEP) operated a temporary seismic network in the Kingdom of Bhutan between January 2002 and March 2003. This network consisted of five three-component, broadband instruments. Our network, in combination with regional stations, detected approximately 2,100 teleseismic, regional and local events, $\sim$900 of which do not appear in the global NEIC catalog. We reviewed these 900 events and manually timed P- and S-wave arrivals. Using a subset of $\sim$200 events which occurred within Bhutan, a preliminary 1-D local velocity model was developed using the program VELEST. We further evaluated these events by quality weighting each arrival and testing several velocity models for the region. We relocated events with the program HYPOELLIPSE, using a suite of velocity models obtained from other regional studies, as well as our own preliminary 1-D model, to investigate the sensitivity of hypocenter locations and their associated uncertainties to variations in the available models. Using the data output from HYPOELLIPSE, we set specific criteria for developing our highest quality events. These criteria include azimuthal gap, number of stations recording the event, number of phases, and the size of the error ellipses. Based on our criteria, hypocenter locations were obtained and show generally scattered locations; however, they exhibit a concentration in southwestern Bhutan where there appears to be a quasi-linear cluster of events. We evaluate the apparent seismicity concentrations in terms of waveform similarity and the existence of doublet or multiplet / repeating events within our catalogue and present these findings, along with general seismicity parameters for the catalog.
http://www.geo.utep.edu/pub/vgee
S51A-0143 0800h
Renewed Seismic Unrest at Mount Spurr Volcano, Alaska in 2004: Evidence for a Magmatic Intrusion
In early July 2004 the Alaska Volcano Observatory (AVO) detected a pronounced increase in seismic activity beneath the summit of Mount Spurr volcano that continues at present. From 1 July through 31 August 2004, AVO located 1094 Volcano-Tectonic (VT) earthquakes and 177 Long-Period (LP) events within 12 km of the volcano's summit, although many events classified as VT contained mixed frequencies. The largest event has a magnitude of 1.6 and hypocentral depths generally range from 0 to 5 km below sea-level. The cumulative seismic moment for July - August 2004 is 5x10**13 Nm. Focal mechanisms of located events in July and August 2004 are dominated by normal faulting, which is consistent with what has been observed beneath the summit since 1984. This seismicity rate is the highest observed at Mount Spurr since the conclusion of the 1992 eruption sequence. Seismicity in 2004 differs markedly from that observed prior to the eruptions in 1992 in that almost all hypocenters are concentrated beneath the volcano's summit vent and not the historically active Crater Peak vent, site of eruptions in 1953 and 1992. Analysis of AVO earthquake catalogs suggests anomalous seismicity may have begun as early as 20 October 2002 with a prominent swarm of 60 VT earthquakes (Mmax = 2.4) located roughly 2 km west of the volcano's summit. Smaller increases in the shallow seismicity rates were also noted between July and November 2003 and beginning in February 2004. These events ranged in depth between 0 and 4 km below sea-level. A subtle increase of deep LP events was also detected beginning in July 2003 and peaking in June 2004, immediately prior to the onset of strong shallow seismicity. These events concentrate about 4 km to the southeast of Crater Peak, generally range in depth from 20 to 35 km and occur at a rate of 2 to 4 located events per month. Associated with the 2004 seismic activity AVO has also observed anomalous melting and disruption of the summit ice cap that began in late May or early June 2004. On 7 August 2004 an emission rate of 600 metric tons per day of CO2 was measured and H2S was also detected (more than 1 ton per day). The increase in earthquake rates, initiation of both deep and shallow LP events, the melting of the summit ice cap, and the degassing of both CO2 and H2S suggest that new magma intruded beneath the volcano possibly as early as October 2002.
http://www.avo.alaska.edu
S51A-0144 0800h
Free Oscillations of the Earth Observed on the Hydroseismogram Obtained in Closed Boreholes
We have made observations of pore pressure under undrained condition by an airtight borehole penetrating an artesian, or a confined aquifer in the Mozumi observation tunnel excavated at the Kamioka Mine, central Japan. We confirmed that the relation between pore pressure change and stress change is a zero-order system for a wide range of frequency and that stress change, strictly speaking strain change, induced within the rock mass shared by the skeletal framework of rock and pore fluid. The 3 November 2002 $M_{w}$ = 7.9 Denali earthquake (epicentral distance $\Delta$ = $51.3\deg$) provided us to investigate how a closed borehole responds to free oscillations of the Earth. We made a Fourier analysis of the hydroseismogram, or the pore pressure record, produced by the Denali earthquake. We examined (1) whether the closed borehole has sufficient sensitivity to identify free oscillations, and (2) how the closed borehole responds to spheroidal modes and troidal modes. The poroelastic theory predicts that pore pressure should respond only to spheroidal modes since pore pressure change is proportional to volumetric strain change. No pore pressure response is expected from shear strain that is produced by troidal modes. However, it is controversial whether pore pressure responds to shear strain, since phases corresponding S- and Love waves have been usually detected on hydroseismograms. We calculated the spectrum of the 24 hours record from origin time of the event, and the sampling interval was 1 second (86400 points). The spectrum peaks correspond to free oscillations were clearly observed. Identified normal modes below 5 mHz were $_{0}S_{6}$, $_{0}S_{13}$, $_{0}S_{16}$, $_{0}S_{17}$, $_{0}S_{18}$, $_{0}S_{20}$, $_{0}S_{21}$, $_{0}S_{23}$, $_{0}S_{24}$, $_{0}S_{25}$, $_{0}S_{26}$, $_{0}S_{27}$, $_{0}S_{28}$, $_{0}S_{29}$, $_{0}S_{31}$, $_{0}S_{32}$, $_{0}S_{33}$, $_{0}S_{34}$, $_{0}S_{35}$, $_{0}S_{37}$, $_{0}S_{38}$, $_{0}S_{39}$, $_{0}S_{42}$, $_{0}S_{43}$. The peaks in lowest frequency bands ($_{0}S_{2}$ to $_{0}S_{5}$) could not be identified on the hydroseismogram of the Denali earthquake. On the other hand, no spectrum peak corresponding troidal modes was observed. These results confirm that the poroelastic theory correctly predicts the pore pressure response.