V21D-0627
Initiation of Pacific Cenozoic Seamount Chains at Older Fracture Zones
Most of the Cenozoic seamount chains on the Pacific plate are parallel to the Hawaiian chain and most start on the young edge of a fracture zone. Example include the Erben, Moonless and "West Moonless" chains on the Murray FZ, the Crosstrend on the clarion FZ, the Crossgrain Chain on the Clipperton FZ and the Tuamotus and Society chains on the Marquesas FZ. The Hawaiian chain itself is broken and offset at the Murray FZ. Exceptions include the Marquesas and Emperor chains. Many of the chains do not extend past the next FZ to the south. Older FZ-bounded compartments, not older than more than a few million years relative to the adjacent compartment, are relatively free of seamount chains as compared to adjacent younger compartments, even close to the sites of initiation. The site of initiation is commonly at a small change in orientation or style of the FZ at a point or along a segment of the FZ. After initiation, the chain extends itself to the southeast, in a direction roughly opposite to Pacific plate motion. A proposed initiation mechanism for some chains is a new stress on a "flaw" along the FZ, oriented so as to produce a local tensional strain that opens local fissures that allow magmas to separate and rise from the asthenosphere to erupt as a volcano on the thinner lithosphere within or on the young side of the FZ. The mechanism need call only on upper mantle sources for the volcanism and requires no deep plumes.
V21D-0628
Flexural Isostasy Of Dok Seamounts At The Northeastern Part Of The Ulleung Basin In The East Sea(The Sea Of Japan)
The isostasy compensation and the loading process of Dok seamounts located at the northeastern part of the Ulleung Basin in the East Sea(the Sea of Japan) were studied by using gravity data and bathymetry data. Comparison between observed gravity anomalies and calculated gravity anomalies from isostasy models suggests compensation degrees and elastic thicknesses associated with Dok seamounts. Dok seamounts are composed of three seamounts which have guyot summits. There are the highest and the lowest free-air gravity anomaly peak in the 1st Dok seamount and the 3rd Dok seamount, respectively. In spite of similar size, the difference between free-air gravity anomaly peaks of two seamounts is about 50 mGal, suggesting different compensation degrees. The magnetic anomaly amplitude of the 3rd Dok seamount is much lower than that of the 1st Dok seamount. Comparison between the calculated and the observed gravity anomalies of Dok seamounts suggests that the flexure model is the most applicable model. Assuming the effective elastic thickness along seamounts to be 5 km for the 1st Dok seamount, 3 km for the 2nd Dok seamount, and 2 km for the 3rd Dok seamount, the modeling result shows that the flexural rigidity of lithosphere of the 1st Dok seamount is stronger than the 3rd Dok seamount, indicating that the age of lithosphere at the time of loading of the 3rd Dok seamount is younger than that of the 1st Dok seamount. Effective elastic thicknesses of Dok seamounts range from 2 to 5 km, which is similar to the case of seamounts near ridges in the Pacific. This suggests that Dok seamounts erupted on oceanic crust near the spreading axis of the Ulleung Basin. Effective elastic thicknesses of Dok seamounts approximate $200°C - $400°C isotherms in the cooling plate model. Above results and the age dating of the Dok island over sea level suggest that the 3rd Dok seamount was formed first and followed by the 1st Dok seamount. The low magnetic anomaly in the 3rd Dok seamount is supposed due to processes of metamorphism or weathering during long term after the volcanism of the seamount and it also shows good coherence with these results.
V21D-0629
Dilatancy and Failure in Basalt From Mt Etna Under Triaxial Compression
The recent history of Mt Etna volcano was marked by several flank eruptions from fractures that opened and feeded lava flow towards the eastern flank of the volcano. In Mt Etna as in most volcanic systems pervasive fracturing of rocks whether it is the lava dome or the surrounding rocks is a dominant feature. Understanding how the strength of volcanic rock varies with stress state, pore fluid content and pressure, damage (content and anisotropy) is fundamental to understanding the dynamics of volcanic systems and in particular modeling the progressive transport of magma towards Earth's surface that leads to eruptions. In this study, we investigated the micromechanics of failure in Mt Etna's basalt. Our block of basalt had a nominal connected porosity (measured by water saturation) of 5% and was composed mainly of pyroxene, olivine and feldspar. Microstructural observations of the intact material revealed the presence of thin cracks (probably formed during the rapid cooling of the lava) and quasi-spherical voids formed during degassing. Some of those cavities appear isolated suggesting that the total porosity of this rock could be significantly higher than the connected porosity. Under hydrostatic conditions however, significant compaction was observed even after closure of the cracks up to 450 MPa of effective pressure. We performed around 20 conventional triaxial experiments on water saturated samples in drained conditions at confining pressures between 10 and 150 MPa and with 10 MPa of pore pressure. Dilatancy and brittle faulting were observed in all samples. Below 150 MPa of effective pressure, a single shear band oriented at 30 degree cut through the samples. At 150 MPa of effective pressure, a somewhat different failure mode involving conjugate shear bands was observed. Up to 50 MPa of effective pressure, Young's modulus increased linearly with pressure and dilatancy was not accompanied by any increase in acoustic emission (AE) activity before the macroscopic failure of the sample. Beyond 50 MPa of effective pressure, the failure envelope became nonlinear and AE activity increased significantly after the onset of dilatancy. Several experiments were stopped at different stages of the deformation after dilatancy at 10 and 50 MPa of effective pressures. Petrophysical thin sections of the deformed samples were prepared. Microstuctural observations revealed the micromechanism leading to dilatancy before and after closure of pre-existing cracks in Mt Etna basalt.
V21D-0630
Evaluation of Microcracks orientation at Stromboli volcano using a Magnetic Ferrofluid and the Method of Anisotropy of Magnetic Susceptibility
Most crustal rocks are anisotropic. In volcanic areas, anisotropy primarily results due to preferred directions of microcracks as magma cools. This effect is, in turn, enhanced due to local stress fields during deposition. The combined effects of these processes may thus give rise to a complex anisotropic fabric. Such fabrics can play crucial roles when enhancing the formation of slip surfaces which can lead to sector collapses of volcanic edifices, as is the case of Stromboli volcano (Italy) which experienced 4 sector collapses in the past 13ka. However, the rapid analysis of anisotropic microcrack fabrics (in terms of magnitude and principal direction) remains non-trivial. Current methods range from time consuming microcrack analysis of thin sections to the preparation of oriented cores for elastic-wave velocity measurement. To further our understanding of how microcrack fabrics influence the bulk properties of volcanic basalt, we employ a novel method which rapidly evaluates the 3-D microcrack orientation using technique of Anisotropy of Magnetic Susceptibility (AMS). First, we determine the rock matrix AMS (mAMS) using standard methods (via a Agico KLY-4 Kappabridge). Samples are then saturated with a magnetic ferrofluid, filling the microcrack network with a magnetically susceptible suspension of microscopic (10nm) magnetite particles. The AMS is then re-measured, with the matrix susceptibility values subtracted from these readings to yield the average 3-D pore space shape, size and orientation (pAMS). We describe the use of this method using basalt from Stromboli and comparing to a granite (Takidani) from the Japanese Alps in order to verify the technique and to investigate the relationship between the basalt microcrack geometry and field scale observation. For Takidani granite we find the structural anisotropy formed by the void space, as measured by pAMS, is well described by elastic wave velocity measurement; exhibiting anisotropy values of 19.1% and 7.6% for P-waves and S-waves respectively. Stromboli basalt possesses a weaker anisotropy of 4.7% and 3.0% (P-wave and S-wave velocity). We relate our pore space AMS measurements to the layering observed in Stromboli basalt on the flanks of the volcanic edifice; and infer that the microcrack network is both formed by this deposition and active tectonics as well as providing a key control on its physical properties. Such data has crucial significance upon the accurate assessment of flank stability, with consequences to hazard assessment for the surrounding area.
V21D-0631
Relationships between tectonism, volcano-tectonism and volcanism: the Ischia island (Italy) case.
The resurgent calderas of Ischia, Campi Flegrei and Pantelleria are characterized by differentially displaced blocks, and distribution of later eruption vents in a well defined sector of the resurgent area. These features suggest a simple shearing block resurgence mechanism. Moreover, the studies carried out on Ischia and Campi Flegrei evidenced a very complex structural pattern due to deformation related to the local stress regime induced by magmatism and volcanism and also to reactivation of regional structures. In order to better define the relationships among tectonic, volcano-tectonic and caldera resurgence mechanism, a structural study has been carried out at Ischia, where the Mt. Epomeo has been uplifted of about 900 m in the past 30 ka. The measures taken on 1,400 planar surfaces (faults, joints and fracture cleavages) show that the resurgent area is composed of differentially displaced blocks whose uplifting is maximum for the Mt. Epomeo and decreases southeastward. The resurgent area has a poligonal shape resulting from the reactivation of regional faults and by the activation of faults directly related to volcano-tectonism. The limit of the resurgent area is not defined towards the north, as beach deposits displaced at variable elevation by E-W and NW-SE trending faults, are exposed along the coastline. The western sector is bordered by inward-dipping, high-angle reverse faults, whose directions vary from N40E to NS and N50W from NW to SW of the block, testifying a compressional stress regime active in this area. These features are cut by late outward-dipping normal faults due to gravitational readjustment of the slopes. Vertical faults border the block at NE ad SW with right transtensive and left transpressive movements, respectively. The area located to the east of the most uplifted block, characterized by a tensile stress regime, has been deformed by N-S, N40-70E and N15W trending normal faults, with maximum elongation direction along N50W. The results of our study and the volcanological data of the past 3 ka, suggest that the eastern part of the resurgent block is the area with highest probability of vent opening in case of renewal of volcanism. Occurrence of landslides just before and after eruptions, suggest that resurgence occurs through discontinuous vertical movements which likely trigger the volcanic activity.
V21D-0632
Constraints on the Timing and Geometry of Kula-Farallon Ridge Subduction and Implications for the Eocene Magmatic and Tectonic History of the Pacific Northwest
Dates on basalts in the WA-OR Coast Ranges and on scattered centers of adakite and near-trench magmatism in WA and Vancouver Island constrain the timing and geometry of subduction of the Kula-Farallon Ridge (KFR) beneath western North America between ~60 and 35 Ma. These data indicate that: (1) Coast Range basalts from widely separated areas show extensive age overlap with no apparent geographic trend over time, (2) 43 Ma adakites in WA are coeval with the Tillamook volcanics in Oregon, and (3) between 48 and 35 Ma the focus of adakite / near trench magmatism migrated northward from the southern Olympic Peninsula to central Vancouver Island. Points (1) and (2) require left-stepping offsets of the KFR, which would have led to multiple ridge-trench intersections and the development of a series of fraternal slab windows (Thorkelson, 1996). The exact geometry of this fraternal window system (FWS) cannot be reconstructed as there are no constraints on the geometry of subducted ridge offsets, but it would have been larger in area than a window resulting from a single ridge intersection (e.g., Breitsbacher et al., 2004). Preliminary calculations of the maximum extent of the KFR FWS indicate it could have extended from SE Oregon to southern BC and as far inland as Montana. Development of this FWS provides an explanation for widespread Eocene volcanism of the Challis event. The arc geochemical signature of many Challis rocks may reflect: (1) the presence of the subducting Farallon plate under some portions of the region, and/or (2) contamination of the lithosphere during earlier subduction episodes (Hooper et al., 1995). Subsequent establishment of the modern Cascade arc at ~39 Ma corresponds in space and time with the reappearance of a subducting slab beneath OR/WA as the KFR migrated northward. This model for the transition from Challis to Cascade magmatism eliminates the need for a steepening or foundering of the subducting Farallon slab, accounts for the width and orientation of the Challis arc, and explains the lack of an east-to-west age progression in Challis age volcanism. Other Eocene phenomena in the Pacific NW that may have resulted from this FWS include bimodal volcanism in WA, high-temperature metamorphism in the North Cascades, S- and A-type plutonism in WA and BC, and development of sedimentary basins.
V21D-0633
Structural Morphology and Backarc Tectonic Imprint on Submarine Volcano Construction and Destruction, Central Kermadec arc ($35° - $30°S), SW Pacific.
Recent investigations aimed at illuminating the interplay between magmatic and tectonic processes have focused largely on continental rifts and oceanic ridges. In comparison, such work for active arc systems, where rifting is proximal, is less advanced, certainly for the case of submarine arcs. We present a detailed structural and volcanic morphological analysis of high-resolution multibeam data of central Kermadec arc volcanoes that clearly demonstrate the control of backarc tectonism on the evolution of the arc front. Of the 11 areas investigated, Macauley, Kuiwai and Brothers illustrate the range in edifice types (caldera, stratovolcano, satellite cones) and the influence of oblique backarc rifting on edifice construction and destruction. Fabric-parallel volcanic lineaments (striking $030°-$060°) include volcanic ridges ± cones and aligned cones ± faults. These features predominate outside areas of significant edifice construction, though they may also occur on the flanks of major volcanoes (e.g. Kuiwai). We infer control by dike intrusion parallel to the structural fabric of the rift. In contrast, major edifices are not clearly related to fabric-parallel structures, though they are in places intersected by them. Lineaments associated with major edifices have variable orientations, oblique and orthogonal to the predominant rift fabric. Magmatic intrusion associated with major volcanic centres may perturb the regional stress field, providing a plausible explanation for this structural variability. We suggest that the presence of major arc edifices provides, through thermal weakening, efficient transfer zones for the accommodation of extensional strain within the mechanically segmented upper crust. However, the lack of discrete transfer faults together with the prevalence of dike-controlled features indicates that this is achieved through magmatic rather than purely mechanical processes. Volcanic ridges, aligned cones and lineaments parallel to the rift fabric bound regions of sector collapse (vertical drops > 1500m); edifice destruction is in part controlled by faulting and dike intrusion.
V21D-0634
Ar/Ar and U/Pb Ages and Geochemistry of the Benton Range Dike Swarm, SE California: New Evidence for an Independence Poly-phased Dike Swarm
The Independence dike swarm (IDS) is a locally profuse, mostly NNW striking and ~700 km-long dike swarm occurring throughout southeastern California and possibly extending into northern Mexico. Dike compositions range from mafic to silicic (though strongly bimodal) and span the composition range of the coeval Sierran calc-alkaline arc plutons. Recent geochronological and structural investigations had cast some doubt on the accurate definition of the Independence dike swarm as the swarm more likely represents a poly-phase dike assemblage including at least two generation of dikes (i.e. 90 and 150 Ma; Chen and Moore, 1979; Coleman et al., 2000). To date, most of the geochronological and geochemical investigations available are strongly localized and part of the swarm lacks basic data. Here, we present new 40\Ar39Ar (n=4) and U/Pb (n=1) ages and detailed major, trace and REE geochemical data on mafic (E-W to N-S) and more abundant silicic (NW-SE to N-S) dikes from the Benton Range dike swarm (BRDS), north-easternmost IDS. Two of the silicic dikes yielded concordant biotite 40\Ar39Ar mini-plateau and weighted mean ages of 153 ± 2 and 152 ± 3 Ma (2 sigma), in agreement with previous K-Ar biotite ages of 150-155 Ma (Renne et al., 1987) and similar to the "accepted" age for the Independence swarm (~150 Ma). These biotite ages, however, may record a cooling age that significantly post-dates dike intrusion. One silicic dike yields a significantly older, preliminary $^{206}$\Pb$^{238}$U zircon age of 164.6 ± 0.8 Ma from single-crystal analyses. Two E-W striking mafic dikes which cross-cut NNW-striking silicic dikes yield hornblende plateau and mini-plateau ages of 171 ± 2 and 166 ± 2 Ma. BRDS chemical compositions are typical of mafic (SiO2 = 47-57 wt%; La/Yb$_{n}$ = 3-14) and granitic (SiO2 =67-77 wt%; La/Yb$_{n}$ = 5-31) arc magmas (e.g. Nb anomaly) and typify the known end-member compositions of the IDS. These results, together with compilated published and unpublished geochronological and geochemical data and field relationships suggest that: (1) The poly-phased Independence dike complex was apparently intruded in several major, magmatic episodes (>170-140 Ma; 115-120 Ma and 100-85 Ma). The first and last diking time periods are correlated with the two major pulses of arc-related igneous activity throughout California (e.g. Ducea, M., 2001) while the second period occurred at the onset of the second major pulse. The dikes were intruded during a time span coinciding with major changes in the orientation of plate convergence (e.g. J2 cusp at ~150 Ma; May et al., 1989). Therefore, the orientation of the dikes is unlikely to be entirely controlled by stress induced by the converging plate. Rather, the data suggest that the dikes are preferentially associated with weakened lithospheric pathways (imposed by the orientation of the free margin?) that have controlled dike orientations over >90 Ma (e.g. Coleman et al., 2000). (2) Geochemical data are consistent with a lithospheric mantle origin for the mafic magma. Isotopic data (in progress) are required to test whether silicic dikes involve a pre-Mesozoic crustal component or are evolved directly from a mantle melt. As yet, no correlation is observed between the age and the composition of the dikes. Chen & Moore, Geology 7, 1979 Coleman et al., GSA Bulletin 112, 2000 Ducea et al., GSA today, nov. 2001 May et al., Tectonics 8, 1989 Renne et al., Geology, 1987
V21D-0635
Volcanic eruptions following mega-earthquakes of the 20th century and implications for the Sumatra-Andaman volcanoes
Two volcanic eruptions in the Sumatra-Andaman volcanic line following the 2004-2005 M9.3 and M8.7 earthquakes add evidence to the hypothesis that large earthquakes trigger volcanic eruptions. The eruptions appear to be encouraged by elastic stress transferred to the volcanic systems, yet the type of stress changes that encourage eruptions as well as the mechanisms of eruption triggering are not understood. Here we analyze eruptions in volcanic lines associated with the subduction zones of four mega-earthquakes (Kamchatka 1952, Chile 1960, Alaska 1964, Sumatra-Andaman 2004). We show that during these earthquakes numerous volcanoes experienced extensional strain. These volcanoes display also an abrupt increase of the eruption rate, most notably for the Mw 9.5 1960 Chile earthquake the largest-ever recorded earthquake. This study suggests that earthquake-induced extensional strain encourages eruptions. Possible mechanisms for eruption triggering include unclamping of the rock in the vicinity of the volcanic system facilitating dike propagation, bubble growth in magma chambers leading to the build-Up of magma overpressure, and decompressional melting. The results of other mega-earthquakes imply an likely increase of eruptions at the Sumatra-Andaman volcanic line.
V21D-0636
Geodetic constraints on the magma chamber of the Hekla volcano, Iceland
Hekla is one of the most frequently erupting volcanoes in Iceland, with at least 18 eruptions in the past 1000 years. During the last 40 years it has erupted about every 10 years. The most recent eruption occurred on February 26 - March 8, 2000. The initial sub-Plinian phase lasted less than one hour, later the eruption became effusive. The crustal deformation network around Hekla consists of strainmeters, campaign and continuous GPS, optical tilt stations and InSAR. Strain and tilt at one station were measured immediately before the 2000 eruption, but the GPS network was surveyed last in 1996. After the eruption the GPS network has been surveyed annually. The far strain stations and one tilt station are used to model the magma chamber, located at 11-km depth under the summit of Hekla. The volume of the dike is estimated from calculated erupted volumes and the ratio of volume that left the magma chamber at different times. The eruptive fissure is 6.6 km long on the surface and with a width of 0.8 m. The dike model is constrained by a strainmeter at 15 km distance and by deformation observed with the GPS network, and is estimated to be about 1 km in height, and fed by a narrow conduit. Diking associated with the 2000 eruption of Hekla appears to be limited to very shallow depth, suggesting that the dike is mostly within the topographic edifice of Hekla (about 1.5 km high). From this it seems that Hekla is presently behaving as a stratovolcano, rather than a typical Icelandic rift-zone volcano.
V21D-0637
Dike Propagation Driven by Volcano Collapse: A General Model Tested at Stromboli, Italy
Analogue experiments investigate how flank collapse affects dike propagation within volcanoes. Water (magma analogue) is injected within a gelatin cone (volcano analogue) with a lateral collapse. The injections form dikes that, away from the collapse, become radial. The dikes propagating nearby the collapse focus towards the collapse sides, becoming subparallel to them, because of the stress reorientation due to the unbuttressing. Only dikes formed along the collapse axis (passing through the mean points of the collapse, map view) propagate radially within the collapse. This general model is applied at Stromboli. The Stromboli dikes cluster along and parallel to the collapse sides, similar to the experiments. Nevertheless, the current location of the conduit, along the collapse axis, leads to dike propagation within the collapse infill, rather than at its sides. This may have occurred during the 2002-03 eruption, when diking in the collapse triggered a landslide and tsunami within the collapse.
V21D-0638
Tectonics of the Hengill Volcano, Southwest Iceland
The Hengill Volcano is the center of a 5-10 km wide and 60-70 km long volcanic system in the West Volcanic Zone of Iceland. The volcano is mostly composed of hyaloclastites (basaltic breccias), but contains also pillow lavas, pillow breccias, as well as andesitic and acid rocks, mainly dacite and rhyolite. The volcano formed in the past 0.8 Ma, is the site of powerful geothermal fields, many of which are related to large normal faults that dissect the volcano and generate a major graben through its top region (at 800 m.a.s.l.). This graben continues north along the volcanic system and is referred to as the Thingvellir Graben in the Holocene lava flows at Lake Thingvallavatn. Hengill has no collapse caldera. Drill-hole data indicate that inclined, basaltic sheets make up 50% of the rock at 1.5 km depth and nearly 100% below 2 km depth. Hengill was subject to intense seismic activity in 1994-2000 when more than 80,000 earthquakes (reaching up to M5) were recorded in the volcano and its surrounding. Most earthquakes occurred on NNE-trending dextral and ENE-trending sinistral faults, but some on NE-trending normal faults. Coinciding with the seismic activity, there was crustal doming, with a maximum uplift rate of about 2 cm/a. Most of the seismicity and deformation came to an end following two M6.6 earthquakes in South Iceland in June 2000. To understand better the deformation, seismicity, and geothermal activity in Hengill, we made a detailed tectonic study of the volcano in 2005. The data obtained include measurements of 2044 joints and minor faults (many conjugate), 970 mineral veins, 60 large faults, and 7 dykes. The veins have a general circular strike distribution, but a noticeable peak in the NE direction. Some 35 large normal faults dissect an E-W profile across Hengill and its graben: they trend mostly NNE and range in apparent displacement from about 10 m to 160 m. The minor faults have a general circular strike distribution, but peak at two main trends: NNE and ENE. Displacements are normally difficult to determine, because of lack of marker horizons. Nevertheless, these minor faults clearly correspond to main trends of the dextral and sinistral faults generating the intense seismic activity in 1994-2000.
V21D-0639
Conceptual and Numerical Models of Ring-Fault Formation in Composite Volcanoes
Some ring faults are dip-slip; others are partly faults (shear fractures) and partly ring dikes (extension fractures). Even if stresses tend to concentrate at discontinuities such as ring faults, most unrest periods in existing calderas do not result in caldera slip. Also, caldera formation is a rare event in the lifetime of any composite volcano. Here we present new conceptual and numerical models of caldera formation, including nested calderas, in volcanoes with shallow spherical or sill-like (oblate ellipsoidal) magma chambers. In all the models, the host rock (including the volcano) above the chamber is composed of 30 comparatively thin layers with stiffnesses (Young's moduli) alternating between 1 GPa to 100 GPa. The chamber itself is located in a single, thick layer. The crustal segment hosting the chamber is either 20 km or 40 km wide but has a constant thickness of 20 km. The loading conditions considered are: (1) a crustal segment subject to 5 MPa tension; (2) crustal segment subject to excess magmatic pressure of 10 MPa at the bottom (doming of the volcanic field containing the chamber); and (3) a combination of tension and doming. In all models, the magma-chamber top is at 3 km depth; the diameter of a sill-like chamber is 8 km (its thickness is 2 km), that of a spherical chamber is 4 km. We conclude as follows: (1) Excess pressure in a chamber normally results in dike injection rather than caldera formation. (2) For doming or tension, a spherical magma chamber favors dike injection except when the chamber is in a very soft (10 GPa) layer, or one with recent dike injections: then the stress field favors ring-fault formation. (3) For a sill-like chamber in a 20-km wide volcanic field, a ring-fault can be generated from either tension or tension and doming; for a 40-km wide field, doming alone is sufficient to generate a ring fault. We conclude that stress fields in composite volcanoes with sill-like chambers subject to tension, doming, or both are likely to generate ring-faults.
V21D-0640
Infrastructure of the Geitafell Volcano, Southeast Iceland
The Geitafell Volcano is an extinct 5-6 Ma Tertiary composite (central) volcano with a collapse caldera, 4 km in diameter. Parts of the volcano are eroded to depths of about 2 km beneath its original top. The erosion provides an exceptionally well-exposed infrastructure of a typical composite volcano in Iceland. In the core of the volcano is a gabbro pluton which forms the uppermost part of its extinct shallow magma chamber. The pluton makes a sharp contact with, and acted as a source of, a swarm of (mostly basaltic) inclined sheets. Next to the contact with the gabbro the inclined sheets form 80-100% of the rock, but the intensity falls off rapidly with distance from the contact. To understand better the infrastructure and tectonic evolution of the Geitafell Volcano, we measured more than 500 inclined sheets and dikes, 68 (mostly normal) faults, more than 400 mineral veins, and nearly 1100 joints and minor faults. The inclined sheets dip from horizontal to vertical; they have a general circular strike distribution but with one minor E-W peak and a major NE peak, which coincides with the general trend of the volcanic zone within which the volcano developed. The strike of joints and minor faults has a more even distribution than the strike of sheets. Nevertheless, there are two minor peaks of N-S and NE-trending joints, and a major peak of E-W trending joints. For the mineral veins, the strike distribution is also generally circular with two peaks: E-W and N-S. Most joints are cooling (columnar) joints and occur in the outermost part of the gabbro pluton. Many joints were subsequently used by geothermal water, which is one reason why the attitudes of the joins and mineral veins are so similar. Some joints were used as pathways by the latest dykes and sheets to be injected from the chamber. These late-formed sheets passed through a cooled but still-hot envelope of the magma chamber on their way out to the main sheet swarm. The sheets that dissect the gabbro envelope, however, are few in comparison with those that constitute the high-intensity swarm at the margin of the gabbro pluton.
V21D-0641
Morphological Study of Jaraguay and San Borja Volcanic Fields, Baja California, Mexico.
Volcanism younger than 12.5 Ma has occurred mainly as monogenetic volcanic fields along the Baja California Peninsula, but until now very little attention has been given to the morphological description of this type of volcanism. In this study we present the preliminary results of the first stage of elaboration of a Geographical Information System (GIS) of the northernmost volcanic fields of the Peninsula; Jaraguay and San Borja, which are among the less studied fields in the region. The present status of the GIS includes the main morphological characteristics and localization of over 350 eruptive centers identified in both volcanic fields. Our data show that over 90% of the eruptive centers are cinder cones, whereas the rest of volcanic structures include some stratovolcanos, shield volcanoes and calderas. Detailed analysis of digital elevation models and 14 m resolution Landsat TM images show a remarkable diversity of the size of reported eruptive centers: the average height lies around 720 m with peaks that reaches 1,412 m asl. Preliminary graphic analyses show local concentrations of the biggest volcanic structures in some areas in both fields. This spatial distribution is more evident at the San Borja volcanic field where the biggest volcanic centers are systematically located in its south and north-western boundaries. A similar concentration of bigger volcanic structures is found at the western edge of Jaraguay volcanic field where cinder cones are largely confined to its eastern most edge. The observed morphological changes of volcanic structures occur in both cases within a distance of less than 100 km. We interpret such variations as the result of heterogeneities of the low velocity zone below this particular area of the Baja California Peninsula, although the final evaluation of the distribution of the regional stress field and its relationship with the actual spatial distribution of eruptive centers must wait until more information becomes available. In any case, the results obtained so far can be used to guide future surveys aiming to elucidate the tectonic association of post-subduction volcanism in the Baja California Peninsula.
V21D-0642
Evolution of magma plumbing systems in the late Cenozoic NE Honshu arc, Japan
NE Honshu, Japan is an island arc associated with the back-arc Japan Sea basin. Recent advances in seismic tomography enable visualization of the crust-mantle structures beneath the NE Honshu arc. Combined with geology and petrology, the seismic images have become valuable tools for revealing the mantle to crustal structures of the arc. The structure of the present mantle to crust inferred from geological and petrological data correlated well with the 3D seismic structures. Intra-crustal thermal structures would have been affected by intensive magma intrusions to form large magma storages beneath the late Miocene to Pliocene calderas in the NE Honshu arc. The remnants of these thermal disturbances are still detectable by seismology. Geological and petrological studies of the late Cenozoic volcanic rocks from the NE Honshu arc have revealed secular variations in the magmatism. These variations are closely related with tectonic evolution of the arc. The island-arc stage in the NE Honshu arc can be subdivided into four substages. These are submarine volcanism (13.5-8Ma), late Miocene caldera-forming (8-5.3Ma), Pliocene caldera-forming (5.3-1.7Ma), and Quaternary stratovolcano-forming (1.7-0Ma) substages. Changes in mode of eruption reflect the intra-crustal stress regime controlled mainly by the plate motion. The reflection experiments (Sato et al., 2002) of the distribution area of the late Cenozoic calderas were confirm the existence of hot, solidified remnant of laccolithic magma reservoir with fluid-saturated top beneath the late Miocene caldera at depth of 3-5 km. The seismic tomography also show that the existence of hot, solidified magma reservoirs beneath the each late Miocene and Pliocene calderas at depths of 3-7 km and 10-15 km, respectively. The site of caldera-forming reservoirs migrated to the back-arc side and depths of the magma reservoirs increased with time from the late Miocene to the Quaternary. And, the magma plumbing systems changed from laccolith associated with piston-cylinder type caldera to vertically elongated stock associated with funnel type caldera. Many Quaternary stratovolcanoes located on the walls of these preexisting calderas.
V21D-0643
Drilling a Volcano: Scientific Experiment at Alban Hills, Italy
Only a few deep boreholes have been drilled for scientific purposes on active volcanoes in the whole world. Indeed, data collected from deep wells are fundamental to better model geophysical processes. Within the Italian research project INGV-DPC-V3.1 (funded by the Italian Civil Protection Department), we planned to drill a 400m hole with the main goal to define the orientation and magnitude of present stress field in the shallow crust in the Alban Hills. The Alban Hills are considered a quiescent volcanic district, belonging to the Quaternary volcanic belt of the Tyrrhenian coast. They are located in a densely populated area close to Rome, then an eruption would be a real risk, also considering the type of their past activity. Alban Hills have been fully studied by means of surface or very shallow observations and indirect methods: now we are going to start the first scientific program to investigate them directly at depth. We will perform some hydrofracturing tests at different depth in the drilling located in a key area, to compute, for the first time beneath a volcano, the absolute values of stress principal axes and reconstruct the stress path along depth. Analysis on core samples will allow to better understand the geomechanical characters of volcanic rocks and their underlying sedimentary basement. Coupling these studies with structural, geochronological and palaeomagnetic investigations will constrain the recent volcano-tectonic processes. The comparison of new data with the available stratigraphic logs will give insights on the occurrence of tectonic movements. Analysis of the anisotropy of the magnetic susceptibility could provide information on the Middle Pleistocene strain to be compared with present-day data. These results will be integrated with new geodetic and seismological data obtained by other research units and will be used for physical and numerical modeling to understand the behavior of the whole volcanic complex. This experiment represent a first step to test methods and instruments for future applications also in other sites. At the end of the project the borehole will be used to host instruments, e.g. a seismometer that will provide good quality seismic data in such a noisy area where earthquakes are characterized by swarms of small magnitude and shallow depth.
V21D-0644
Influence Of Phase Transition And Interaction Between Hot Spots On Convection In The Upper Mantle
The mathematical modeling is widely used for study of convection in the upper mantle (with different approaches). The models of flow dynamic in the upper mantle includes, usually, one phase transition at the depth of 480 km and only hot spot. This paper presents simulation for model of the upper mantle with several hot spots in close arrangement and deals with the effect of their interaction on flow dynamics in the mantle beneath the oceanic and continental crust. Our computations demonstrate a drastic change in the flow mode in the mantle for the case of two interacting hot spots. Further increment of hot spots does not change the mantle flow pattern. The stress distributions in the upper lithosphere above the zones of anomalous mantle were the object for simulation; these distributions produce the hot shearing zones. We believe that interaction of several hot spots might be significant for development of intraplate continental basic melts, including the case of flood-basalt formations (LIP). For this situation, as simulation evidences, we have to take into account all phase transitions in the upper mantle and lithosphere; these transitions are crucial for lithosphere structure. This work was supported by the RFBR (Grant No. 04-05-64107), the Presidium of SB RAS (Grant 106), the President's Grants (NSh-2118.2003.5, NSh-1573.2003.5).
V21D-0645
Emplacement of Saucer-Shaped Sills and Long Dykes: Constrains from Detailed Field Work and AMS Analyses in the Karoo Basin, South Africa.
The plumbing system of large igneous provinces (LIPs) consists largely of sheet-like intrusions that are vertical (dykes) and sub-horizontal (sills). Detailed seismic mapping in the NE Atlantic and field work in the Karoo basin, South Africa, show that the magma tends to develop into interconnected 3D networks of saucer-shaped sills in undeformed basin provinces. Minor dykes, up to 1 m thick and 200 m long, may originate from the sills whereas major dykes are not involved in the formation of the saucer-shaped sill complexes. We have recently initiated a major integrated field, laboratory and theoretical study of magma emplacement processes in sedimentary basins. The remarkably well-exposed Golden Valley Sill Complex and two long (> 100 km) dyke systems in the central Karoo basin have been selected for detailed field, geochemistry, and anisotropy of magnetic susceptibility (AMS) studies. The Golden Valley Sill Complex contains four individual saucers, each having a horizontal inner sill, transgressive inclined sheets and sub-horizontal outer sills. Macroscopic magma flow indicators include ropy flow structures and tubes. The upper sill surfaces commonly exhibit large-scale undulations. A total of 115 localities have been analyzed for low field AMS to define the ellipsoid of anisotropy. The AMS and field data are interpreted in terms of syn-emplacement (magma flow directions) and post-emplacement (thermal contraction, shearing) processes. The flow indicators support an emplacement model where the saucers are fed from the center and transgress around the edges of the sub-horizontal inner sill. Localized shearing of the magnetic fabric and the undulations suggest a late phase of inward magma flow during cooling and contraction of the saucer. A total of 24 localities from two sets of dykes have been analyzed for AMS. The first set consists of NW-SE trending dykes, between 20 to 100 km long and 15-20 m wide, forming prominent topographic highs. The second set, the ''Gap Dykes'', consists of two parallel E-W trending dykes, each more than 300 km long and 100 m wide, having negative or no topographic expressions. Magma flow indicators are rare, and consist primarily of the upward bending of sedimentary strata in contact with the dykes suggesting an upward magma flow. AMS measurements on both dyke systems reveal vertical magnetic lineations, suggesting that the magma flowed vertically upward during emplacement of the dykes.
V21D-0646
Vocanic Deformations During Repose Interval Revealed by GPS Measurements, Batur Volcano, Indonesia
Batur volcano is located north west of Bali Island in Indonesia.This volcano has two calderas with more than 10 km in diameter. Recent eruptions with lava flow occurred in 1963 and 1974. No effusion of lava has been observed since 1990, although steam explosions occurred August 1994, November 1997 and June 1998.This suggests that magmatic activity of this volcano would keep its high level since 1994.GPS observation network of this volcano has been kept by Institute Technology of Bandung (ITB) and Volcano Survey of Indonesia (VSI).The network was consisted of 10 observation points at the beginning in 1999, and now it becomes 23 observation points. We have made GPS campaign observations about five times from 2003 to 2005. Each campaign observation consisted of a couple of days of measurements for each observation point. In order to keep the quality of data as high as possible, observations have been made at least 12 hours of continuous data for each point. In this report, results of two campaign observations (December 2004 and July 2005) are used for the analysis. The data thus obtained are fitted to the Mogi source (i.e. a point source model) to locate the depth and amount of volume changes for 7 months. Location of the Mogi source was obtained about 4km southeast of the summit of central cone, and 3km depth with deflation volume change of 1.3_~$10^6 m^3$ for about 7months. For the period from 1999 to 2004, estimated volume change suggests a continuous deflation throughout this period, although the reliability of data was not so high. Continuous deflations might be likely after the last effusive eruption in 1974, would suggest that shallow part of magma beneath the central cone would probably be drained down to further deep, or shrinkage of magma associated with the cooling or solidification. Further data are obviously needed to discriminate the mechanisms of the deformation process during the repose period in this volcano.
V21D-0647
Comparing the Morphology of Mud Versus Magmatic Volcanoes
Morphological parameters of offshore (Gulf of Cadiz, Mediterranean Sea, and Black Sea) and onshore (Azerbeidjan) mud volcanoes are compared with data from magmatic volcanoes. The offshore data series are compiled from multibeam bathymetry measurements (Gulf of Cadiz and Black Sea), and data from literature (Mediterrean Sea). The onshore data series (Azerbeidjan) is derived from Landsat 7 ETM+ and SRTM-DEM data. The studied mud volcanoes range between 25 m and 285 m high and are up to 5.8 km in diameter. The main morphological aspects of the mud volcanoes, both onshore and offshore are, from the margin towards the centre, a subsidence rim or moat, the mud volcano slope, in some cases a deep crater, and a recent central mud dome at the top. The slope varies between concave, linear or convex and is characterised by radial outward sediment flow deposits or by a concentric pattern of terraces and steps. The sediment flow deposits can be divided into elongate outflows that accumulate at the base of the slope and short bulky outflow deposits that freeze on the steep slope. Morphological parameters distinguish two end-members: 'effusive' low rounded mud volcanoes and 'explosive' steep sloped mud volcanoes. Comparing with morphological data from magmatic volcanoes shows great similarity despite the large differences in size. It is demonstrated that the shape of a volcano is independent from its size but seems to be controlled by the flow behaviour of the erupted material. This study is the first to generate a representative data series about mud volcano morphology and to discuss its variability in terms of eruption mechanism. Comparison with magmatic mud volcanoes demonstrates that mud volcanoes are a good small scale analog to study big-scale volcanic processes.
V21D-0648
A GIS-based Spatial Analysis of Volcanoes in the Central Andes: Insights Into Factors Controlling Volcano Spacing.
Volcano spacing has received little attention since the mid-70's when studies undertaken by Vogt (1974; EPSL) and then Marsh (1979; J Geol) suggested a regular spacing of volcanoes in arcs that ranged from 50 to 75 km for different arcs. The spacing was thought to be influenced by the thickness of the lithosphere or gravitational (Rayleigh-Taylor) instabilities related to source layer thickness and viscosity respectively. We have revisited these ideas through a detailed study of volcano distribution in the Central Volcanic Zone (CVZ) of the Andes where volcano spacing was thought to be around 70 km. The CVZ was selected as it is the type example of continental arc volcanism, built on an extremely thick crust of up to 70 km. The availability of a comprehensive dataset describing the relative age, location, and geomorphic characteristics of each volcano (Volcanoes of the Central Andes, de Silva and Francis, 1990, Springer Verlag) made this a compelling case study. The ready availability of ARC GIS Geographic Information Systems software and the geospatial analysis tools therein, allowed a comprehensive spatial analysis of the volcanoes to be conducted. Of the 1,118 volcanoes of ages from 23Ma to active in the CVZ, we focused on the 106 active and potentially active large composite volcanoes that define the modern arc. These volcanoes are related in time and thus to a consistent set of tectonic factors. The frequency distribution of inter-volcano distances shows a peak frequency in the 10 - 30 km range (71%) with subordinate between 40-80 km (19%) and 80 - 120 km (10%). The characteristic spacing is thus much smaller than the characteristic spacing of 70 km found previously and is consistent with Baker (1974; EPSL). The primary cause appears to be clustering of volcanoes into groups. The density of volcanoes is variable along the arc with regularly spaced clusters of two to three volcanoes in northern and southern parts of the arc (13 $r^(0)$S to 19 $r^(0)$S and 24 $r^(0)$ to 27 $r^(0)$S). The central part of the arc between 22 $r^(0)$ and 24 $r^(0)$S shows the highest density of volcanoes. An interpolated continuous raster grid of volcano base elevation revealed a strong positive correlation with the density of volcanoes. Interestingly, the spacing of the peaks in density is fairly regular at about 75 to 110 km. We interpret the different orders of spacing as reflecting control by different levels within the plumbing system for the arc. The spacing of density peaks (clusters) is likely the primary spacing of diapirs rising from the source layer into the lithosphere. The dominant 10 to 30 km spacing of individual volcanoes probably reflects the influence of intra-crustal regions superimposed on the primary spacing. These regions may be mid-crustal, low-velocity, high-conductivity zones that have been imaged geophysically and/or the influence of older upper crustal batholiths that may have been emplaced into the thickest crust between 22 $r^(0)$ and 24 $r^(0)$S.
V21D-0649
A Closer Look at Recent Deep Mauna Loa Seismicity
In 2002, Mauna Loa Volcano showed signs of reawakening, some 18 years since its last eruption in 1984. First, in April, a brief flurry of microearthquakes occurred at cataloged depths from 25 to 55 km beneath Mauna Loa's summit caldera. Then in May 2002, after the microearthquake swarm had ended, geodetic monitors across Mauna Loa's summit caldera registered a change, from line-length shortening to extension, interpreted as reinflation of a magma body approximately 4 km beneath the volcano's summit. Accordingly, the Hawaiian Volcano Observatory issued advisories related to Mauna Loa's stirring. In July 2004, HVO began to record deep long-period (LP) earthquakes beneath Mauna Loa. Historically, interpretations of such seismicity patterns have associated LP source volumes with magma chambers and magma pathways. Over a few weeks, this seismicity dramatically jumped to levels of several dozen per day. Between the months of July and December 2004, nearly 2000 Mauna Loa LPs were located between roughly 25 km and greater than 60 km depths by HVO seismic analysts. In late December, these earthquakes rather abruptly ceased, and their levels have remained low ever since. We seek a more detailed understanding of how these earthquakes may factor into Mauna Loa's eruptive framework. Given that their first arrivals are typically emergent, hypocentral estimates using only P-wave first-arrival times of LP earthquakes are often marginally constrained. With such hypocentral estimates, it is difficult to establish clear relationships among the earthquake locations themselves, or between the earthquakes and other processes like crustal extension or magma accumulation or withdrawl. Building on earlier applications to deep earthquakes in Hawaii and LP earthquakes beneath Kilauea, we are reexamining this unprecedented Mauna Loa deep seismicity with waveform correlation and precise earthquake relocation techniques. Work to date reveals that, although the waveform correlation coefficients are low, a significant subset of the deep Mauna Loa LPs can be relocated to improve our understanding of the remarkable 2004 swarm. We are currently seeking stronger resolution to determine whether the waveform data are consistent with the vertically extended, conduit-like source distributions suggested by the catalog locations or, alternatively, whether the events are consistent with one or more narrowly extended point sources.
V21D-0650
Tectonic Controls on Magmatism at Taal Volcano, Philippines
Space geodetic methods combined with imaging are used to inspect active tectonics and magmatism within the Macolod Corridor and the Taal Caldera, a geologically complex area in Southwestern Luzon. This area is characterized by extensive volcanism and widespread faulting. Radar and multispectral imageries subjected through analytical shading and image filtering techniques and combined with digital terrain models are used to analyze fault orientations and detailed geomorphic features of the area. Campaign geodetic observations (1996-2002) from GPS stations within Luzon are used in combination with remote sensing and earthquake slip vectors to derive the kinematics of Luzon. We use an elastic block modeling approach, which characterizes crustal deformation as a result of rotation of discrete elastic microplates around Euler poles. The resulting best-fit model indicates that the Luzon area is composed of six microplates. The vicinity Macolod area in SW Luzon is best represented with three mobile microplates. Active tectonics of the Philippine mobile belt is dominated by eastward subduction along the Manila Trench (~20-100 mm y-1), westward subduction along the Philippine Trench (~29-34 mm y-1), and sinistral strike-slip faulting along the Philippine Fault (~10-40 mm y-1). The velocity field indicates localized transpression along the N-S trending Marikina Fault (~10-12 mm y-1), and transtensional motion along the NE-SW trending Macolod Corridor fault zone (~5-10 mm y-1). Observations from the continuous single- and dual-frequency GPS stations of the Taal Volcano network from 1998-2005 indicate a sequence of inflationary and deflationary events, which include several periods of rapid volcanic inflation (~120 mm uplift from February to November 2000) and rapid deflation (~33 mm subsidence from June to December 1999). The most recent episode of inflation extended from June 2004 to March 2005 indicated ~73 mm y-1 extension across the volcanic edifice, with about 50 mm uplift with respect to the caldera wall. A recent deflationary pattern starting April 2005 is also detected. Models for the earlier inflation and deflation events indicate that a Mogi point source 4-5 km deep centered at the Volcano Island describes the deformation effectively. The inflationary trends are interpreted to be episodes of magma intrusion to a shallow reservoir beneath Volcano Island, which is significantly affected by regional tectonism.
V21D-0651
Seismicity in Andaman - Nicobar - Java - Sumatra Region and its Bearing on the Volcanism in the Region, With Special Reference to the Barren Island.
Barren Island volcano in the Andaman Sea (Indian Ocean) is the lone active volcano in the Indian Subcontinent. The island showed renewed activity (commenced from May 28, 2005) after the great earthquake of Sumatra (December 26, 2004) along with increased mud volcanism in Bartang (south of Barren Island) and first ever reported mud volcanism on Narcondum (north of Barren Island) in the Andaman-Nicobar Archipelago. These islands lie on a volcanic arc that extends from the extinct volcanoes like Mt. Popa, Mt. Wuntho of Myanmar in the north to the active volcanoes of Sumatra and Java in the south. Regional tectonism of this region is largely driven by the subduction of the Indo-Australian plate beneath the Asian (Burmese) plate. Regional seismicity pattern reflects different tectonic regimes, namely, thrust dominated subduction front, strike-slip faulting (west Andaman fault) and the extensional processes in the Andaman spreading center. Earthquakes of magnitude more than 4.5 on Richter Scale are quite frequent in the region and are related to the subduction-related processes. Continuous seismic activities in the Andaman-Nicobar-Java-Sumatra region cannot be dealt with separately as evident from the increased volcanic activities following the great earthquake of Sumatra. More recently increased seismic activity in the vicinity of the dormant volcano of Mt. Toba is very much likely to culminate in a catastrophic eruption of this volcano in near future.
V21D-0652
Numerical inversion of 1993-97 deformation data at Mount Etna (Italy)
Since 1993 geodetic data obtained by different techniques (GPS, EDM, SAR, leveling) detected a consistent inflation of the Mount Etna volcano. The inflation, culminated with the 1998-2001 strong explosive activity from summit craters and recent 2001 and 2002 flank eruptions, may be interpreted by magma ascent and re-filling of the volcanic plumbing system and reservoirs. Our purpose is to model the 1993-97 GPS and EDM data by pressurized sources simulating the magma reservoir using a 3D Finite Element modeling coupled to a Neighbourhood Algorithm inversion. The power of this technique, if compared with analytical inversions, is that sources can be placed in complex media (heterogeneous, with topography, inelastic, etc.) so that the inversion result is not influenced by the usual approximations of elastic, homogeneous half-space. We model the inflation by a deep volcanic source, considered as point source. The source intensity is given by its stress tensor, identifying an ellipsoidal pressure source. The FE model of Mt. Etna is characterized by a regular mesh below the volcanic edifice, and by arbitrarily distorted brick elements elsewhere. The potential point sources are contained in a cubic volume located below the summit craters, between 4 km and 8 km b.s.l.. For each potential source we compute the deformation and elongation at GPS and EDM observation points. The solution is a weighted combination of displacements (one for each stress component) computed for each of the potential sources. The best source parameters search is performed by the Neighbourhood Algorithm inversion technique. Good results from synthetic tests and previous inversions with only GPS data confirm the robustness of the method. We consider four classes of models characterized by: i) homogeneous medium and flat free surface; ii) homogeneous medium with topography of Mt. Etna; iii) heterogeneous medium with flat free surface; iv) heterogeneous medium with topography. Results from data inversion show evidence for a vertically elongated body. This result is a robust feature of the inversions, since it is common to models with/without topography and with/without rigidity contrasts. Differences in the location and dimension of the ellipsoid are due to the presence of the rigidity contrasts and topography. Solutions are compared also with those derived by inversions of analytical forward models in the homogeneous medium with flat free surface, showing good agreement.
V21D-0653
Investigating Mount Cameroon$^{'}$s Magma Chamber System
Mount Cameroon, located in the middle of the Cameroon Volcanic Line (CVL) on the passive western margin of Cameroon, is one of Africa$^{'}$s most active volcanoes with seven eruptions in the past century. Many population centers are in lava flow hazard zones, making monitoring imperative. Despite its high activity level, monitoring is limited to a small network of seismometers, and only one instrument was functional immediately before the 1999 eruption. Seismic activity rose above background level only a few days prior to the eruption (Suh, et al.,2003, Bull. Volcanol., 65:267-281). In order to improve spatial coverage and potential warning time, we are designing a ground deformation monitoring network. Using existing seismic, petrologic, and historical eruption data (Ambeh,1989, PhD thesis, Leeds U., UK; Suh et al. 2003) along with corresponding models of subsurface magma storage, we apply simple elastic half-space models (Mogi, 1958; Pinel and Jaupart, 2000) to explore the nature of the volcano$^{'}$s plumbing system. We first assumed a closed-system and a 12km deep chamber (Ambeh, 1989), with sufficient overpressure to initiate the 1999 eruption. Given the total erupted volume of 8 x 108 m3 and the required pressure, the chamber must be ~1.2 km in diameter. Overpressure in this chamber immediately prior to eruption causes maximum surface tilt of ~0.3 microradians at 6km from the center. This small amount of deformation is not detectable by surface mounted tiltmeters, however it is detectable by borehole strainmeters. Following conduit initiation, at an average extrusion rate of 40m3 s$^{-1}$ recorded for the 1999 eruption (Suh et al., 2003) and assuming a 10m radius conduit, magma would erupt roughly one full day after conduit initiation. However, if the system fills a shallow chamber at ~3km depth prior to eruption, as is the case in Hawaiian volcanoes (Dvorak & Okamura, 1987), then maximum surface tilt of ~17 microradians occurs at 1.5km from the center, making eruption prediction with tiltmeters feasible. At a 40m3 s$^{-1}$ chamber filling rate, the shallow chamber would take nearly a month to reach an eruptible pressure, providing two weeks of warning with tiltmeters.
V21D-0654
Uplift in the earthquake swarm area of Ontake Volcano, central Japan detected by precise leveling in 2002-2005
Earthquake swarm is repeated on the eastern submountain region of Ontake Volcano, Central Japan for 25 years since late 1970s. Ontake Volcano erupted at the summit in 1979, earthquake M6.8 occurred in 1984. However maximum magnitude of the earthquake swarm is limited to M4. We start the precise leveling to detect the vertical movements in the swarm area in 1999 and repeated the leveling every year. As a result of the precise leveling, uplift of few mm is observed. We extend the leveling route to cross the swarm in 2002 and 2004. From the leveling in 2005, we detect the uplift of 5 mm extending circle of 5 km in the just most active segment area of the swarm. Up lift amounts to 10 mm for three years since 2002. It is significant uplift exceeding of the survey error of +- 2.5 mm. One spherical pressure source is estimated at a depth of 3-4 km from the ground surface (2-3 km below sea level) using Mogi solution. Low risibility zone is estimated from AMT observation (Kasaya et al., 2004) and active gas blow of CO2 is observed in the area (Azuma et al., 2004). These observation and uplift ground deformation from our leveling suggest a hydrothermal injection. Vertical deformation is one of important monitor of the earthquake swarm activity.
V21D-0655
Crustal Control on Crystallization Depths? Preliminary Evidence from Mt. Cleveland, Chuginadak Island, Eastern Aleutian arc
Mt. Cleveland is one of the most active volcanoes of the eastern Aleutian arc with eruptions in 2005 and 2001 and a total of 23 known eruptions in approximately the last hundred years. Several features, perhaps indicating a complex crustal geometry make Mt. Cleveland unusual. First, the Islands of Four Mountains (Herbert, Carlisle, Cleveland, Tana, Kagamil) define an area of unusually closely-spaced Quaternary stratovolcanoes relative to the central and western Aleutians. Also, the Islands of Four Moutains are closer to the trench compared to the rest of the arc. Finally, the earthquake record shows a linear cluster of earthquakes leading from Cleveland toward the trench (map view); as yet, no interpretation exists to explain how the defined linear feature may relate to a crustal structure or the downgoing slab. The crust beneath the volcano is approximately 30 km thick and is transitioning from oceanic to continental affinity. We use clinopyroxene thermobarometry to investigate the depths at which Mt. Cleveland magmas tend to stall prior to eruption to examine whether basement structure may influence the ascent of the magmas. The peak consists largely of lavas and interbedded debris flows although several cinder cones and andesitic domes have formed within five kilometers of the summit. The lavas are primarily calc-alkaline andesites and most contain two pyroxenes, plagioclase and occasionally olivine or titanomagnetite. Two of the cinder cones, in contrast, are olivine-plagioclase-clinopyroxene basaltic andesites. Preliminary data show that a Holocene flow (1994?) from Mt. Cleveland crystallized clinopyroxene at approximately 1035°C and 13+/-2 kb pressure, which corresponds nearly with the base of the crust. In contrast, the 2001 tephra essentially crystallized at 1 atm, consistent with rapid ascent through the crust. The basaltic andesite of the youngest cinder cone crystallized clinopyroxene at approximately 1300°C and 3 kb pressure, indicating some residence time in the crust.
V21D-0656
Ground deformation from dike development
We investigate the relationship between ground deformation from dike emplacement and geometries using analytical models. We used the free surface half space model of Coulomb 2.5. We varied four basic dike parameters: dip, depth of the top, depth of the bottom, and thickness of dilation. We present cross-sections to illustrate temporal vertical deformation. In vertical dike (dip 90o) models, vertical ground deformation increases while the dike develops upward, and the deformation pattern is symmetric about the dike, forming two peaks at both sides of the dike where the lateral extent of surface vertical deformation increases as the dike develops. For a non-vertical dike, vertical deformation is asymmetric about the dike and extends further laterally when compared to the vertical case. Deformation amplitude is greater as well. Surface expression is negligible when the depth of dike bottom is > 6km. Vertical deformation reaches 0 and is downward at the location where the dike will open on the surface. For models with dikes dipping > $45°, along a perpendicular bisector starting from center of the dike vertical deformation decreases on the dipping side of the dike to negative values and then back to positive values, forming a second peak. In models with dikes with dip less than $45°, vertical deformation is negative after the initial decrease near the dike location. Finally, vertical deformation increases when the thickness of dike dilation increases. These models can help guide the initial interpretation of leveling observations made in volcanic areas.
V21D-0657
Volcanoes Behave as Composite Materials: Implications for Modeling Magma Chambers, Dikes, and Surface Deformation
By definition, composite volcanoes are composed of numerous alternating material units or layers such as lavas, sediments, and pyroclastics. Commonly, these layers have widely different mechanical properties. In particular, some lava flows and welded pyroclastic flows may be stiff (with a high Young's modulus), whereas others, such as non-welded pyroclastic units and sediments, may be soft (with a low Young's modulus). As a consequence, even if the loading (tectonic stress, magmatic pressure, or displacement) is uniform, the stresses within the composite volcano will vary widely. In this sense, the behavior of composite volcanoes is similar to that of general composite materials. The deformation of the surface of a volcano during an unrest period results from stresses generated by processes and parameters such as fluid pressure in a geothermal field or a magma chamber, a regional tectonic event, and a dike injection. Here we present new numerical models on mechanics of magma chambers and dikes, and the associated surface deformation of composite volcanoes. The models show that the surface deformation during magma-chamber inflation and deflation depends much on the chamber geometry, the loading conditions, and the mechanical properties of the rock units that constitute the volcano. The models also indicate that the surface deformation induced by a propagating dike depends much on the mechanical properties of the layers between the dike tip and the surface. In particular, the numerical results show that soft layers and weak contacts between layers may suppress the dike-induced tensile stresses and the associated surface deformation. Many dikes may therefore become injected and arrested at shallow depths in a volcano while giving rise to little or no surface deformation. Traditional analytical surface-deformation models such as a point source (Mogi model) for a magma-chamber pressure change and a dislocation for a dike normally assume the volcano to behave as a homogeneous, isotropic half space. The present numerical results, combined with field studies, indicate that such analytical models may yield results that have little similarity with the actual structure being modeled.
V21D-0658
Do Flexural Stresses Explain the Mantle Fault Zone Beneath Kilauea Volcano?
Recent relocation and focal mechanism analyses of deep earthquakes beneath Kilauea volcano, Hawaii indicates that seismicity is concentrated on a horizontal fault zone at a depth of 30 km, with seaward slip on a low-angle plane. We discuss whether the observed localization of the earthquakes can be explained solely by stresses induced by flexure of the Pacific plate beneath the Hawaiian load, or if other perturbations of the stress field are required. We find that flexural stresses are consistent with the observed fault plane orientation, and the direction and rate of slip. However, the model has four shortcomings: (1) the fault zone is displaced 15-20 km to the NW of the region of predicted maximum shear stress. (2) The maximum shear stress in the vicinity of the fault zone seems too low to overcome Coulomb friction (by about a factor of 2, assuming hydrostatic pore pressure). (3) The fault zone is much more localized laterally than is the region of large flexural stresses and stressing rates. (4) The fault zone is more localized vertically than might be inferred from the calculation as well. Simple and plausible extensions of the plate flexure model by accounting for spatial variations in the location of pore fluids, and/or the possible existence of a passive shear-stress free magma transport sytstem can overcome most of these shortcomings. Several magma pipes would be necessary to explain the observed earthquake locations, and simple thermal arguments indicate that such pipes could be conduits for porous flow if they are a few km in radius.