V11B-2018
Explosions, Tephra, and Lava: A Chronology of the 2008 Summit Eruption of Kilauea Volcano, Hawai'i
Beginning in early January 2008, sulfur dioxide emission rates from the summit area of Kilauea increased to 2 to 10 times above background values, prompting partial closure of the summit region by late February. On March 12, 2008, a new gas vent appeared low on the southeast wall of Halema'uma'u Crater. Incandescence was seen at the vent, starting on March 13, and by March 18 the area of incandescence had grown to a diameter of about 30 m. At 0258 on March 19, an explosive eruption occurred, opening and widening the new vent slightly and scattering lithic lapilli and coarse ash over an area of about 50 hectares. The vent continued to degas vigorously after the explosion, with the plume alternating between ash-rich brown and ash-poor white. On April 9, another explosive eruption ejected both lithic and juvenile tephra from the vent, and on April 16 a third explosive eruption, smaller than the previous two, produced pink ash and fine lapilli. During the next 3.5 months, only minor quantities of ash were carried aloft by the vigorously degassing vent, which was typically glowing at night. Then, on August 1, a fourth explosive eruption threw lithic and juvenile debris from the crater, ushering in a period of rapid vent widening. Small collapses occurred nearly every day through August and early September, with larger collapses leading to significant explosive eruptions on August 27 and September 2. By September 5, 2008, the vent had nearly doubled in diameter, improving the view and revealing an active lava lake, about 50 m in diameter, several tens of meters below the rim of the vent. The new vent has provided the first prolonged opportunity since 1967-68 to observe shallow magmatic processes at the summit of Kilauea. The Hawaiian Volcano Observatory, in collaboration with academic institutions and government agencies, is carrying out enhanced monitoring and process oriented research in response to this new eruptive activity.
V11B-2019
Continuous gravity measurements from Kilauea Volcano Hawai'i, 2007-2008
We present more than a year of continuous 1 Hz gravity data collected in an underground vault near the US Geological Survey Hawai'i Volcano Observatory. The data were collected with a LaCoste and Romberg D- meter streaming to a digitizer with a GPS clock. The instrument was installed in February 2007, and has been collecting data continuously since that time, except for periods when the instrument went off scale due to drift or local earthquakes with magnitudes of about M4.0+. This is one of the few installations of a continuous gravity meter at an active volcano and has the potential to provide important new constraints on the physical system of volcanoes. The most significant observation in the time series is a sizable gravity increase during rapid summit inflation at Kilauea in June-July 2007. The inflation was a consequence of refilling of the summit magma reservoir following an intrusion and formation of a new eruptive vent on the volcano's east rift zone during June 17-19, 2007. Uplift should result in a gravity decrease as the instrument moves farther from the center of the Earth, therefore the measured gravity increase during inflation suggests that mass was being added to the summit magma reservoir. Inflation continued until July 21, 2007, when a reorganization of the east rift zone magma plumbing system led to the formation of a new long term eruptive vent and a period of sustained summit deflation. The gravity meter is also an effective low frequency (less than 0.5 Hz) seismometer and we present comparisons of the gravity record with a broadband seismometer. In addition to observing previously identified low frequency (0.004 to 0.5 Hz) events associated with known volcanic activities and sources in the Kilauea summit area, we also recognize previously undetected low frequency events associated with high-frequency earthquakes outside of the Kilauea summit area. We provide a preliminary interpretation of these observations.
V11B-2020
Eruption Products and Collection Methods for the 2008 Summit Eruption of Kilauea Volcano, Hawai'i
After several weeks of elevated sulfur dioxide emissions at Kilauea's summit, a new fumarole appeared low on the southeast wall of Halema'uma'u Crater in mid-March 2008. At 0258 H.s.t. on March 19, an explosive eruption at Halema'uma'u's fumarole created a crater approximately 35 m across. Since that explosive eruption, a gas plume, with variable amounts of ash, has continuously emanated from the vent. Lithic debris was strewn over approximately 50 hectares, but no juvenile material was erupted during the March 19 event. On March 24, however, Pele's hair, Pele's tears, and glass-coated lithics were found on the crater's rim. These samples were the first juvenile material erupted at Kilauea's summit since 1982. After the March 19 explosive eruption, five more significant explosive eruptions have occurred, each of which erupted juvenile and lithic material. Several methods have been used to collect tephra from this eruption. A 10-station network of ash collectors was set up in an array to track changes in fallout from the continuous plume and to calculate a daily accumulation rate. Ash and tephra samples for petrologic and geochemical analysis are collected from wooden boxes placed along the crater's rim. Meter-scale squares on the ground downwind of the vent were cleared of older debris to help recognize new tephra. The samples are classified in 10 categories, based on a variety of characteristics, such as juvenile and lithic component, texture, and appearance. This classification scheme has been devised to help distinguish among the diverse suite of erupted material.
V11B-2021
Rapid Temporal Geochemical and Isotopic Variations of Tephra From the Ongoing 2008 Summit Eruption of Kilauea Volcano, Hawaii
The 2008 summit eruption of Kilauea Volcano is its first summit eruption since 1982 and its first explosive eruption since 1924. Chemical and high-precision Pb isotopic analyses of eight samples of juvenile material from this eruption reveal rapid and systematic changes in magma composition from March 23 to April 14. Compared to the most recent lavas from the ongoing Puu Oo rift zone eruption, the earliest erupted 2008 summit tephra (March 23) has relatively high abundances and ratios of incompatible elements (e.g., K2O, TiO2 and Nb/Y). Most of these chemical parameters are similar to Kilauea summit lavas from the late 1960s to early 1970s. The exception is K2O, which is anomalously high (similar to summit lavas from the early 20th century). Subsequently erupted 2008 summit tephras have abundances and ratios of incompatible elements that are similar to recent Puu Oo lavas. The 206Pb/204Pb ratios of four 2008 summit tephras display a systematic temporal decrease from values intermediate between recent Puu Oo lavas and the September 1982 summit lava to values identical to recent Puu Oo lavas. Taken together, these observations suggest that the early 2008 summit tephra contains a component of magma that had been stored within Kilauea's summit reservoir since the late 20th century. However, this stored magma may have been contaminated within the crust to anomalously raise its K2O content.
V11B-2022
Human Footprints in Relation to the 1790 Eruption of Kilauea
In 1790, a party of warriors and their families was decimated by an explosive eruption of Kilauea; fatality estimates range from about 80 to 5,405. In 1920, thousands of footprints made by barefoot walkers in wet accretionary lapilli ash were found within a few kilometers southwest of Kilauea's summit. In 1921, Jaggar related the footprints to survivors or rescuers of the 1790 eruption, mainly because he assumed that few people visited the supposedly forbidden area except in 1790. Archaeologists from Hawai'i Volcanoes National Park recently questioned whether the footprints were made at that time and by warriors, citing a wide range of directions that people were walking and evidence of extensive human use of the area. Forensic and anthropologic studies indicate that a human foot is about 15 percent of an individual's height. A man's foot may be slightly more that 15 percent, a women's slightly less, but nonetheless the height can be estimated to within a few centimeters. We measured the heel-big toe length of more than 400 footprints and calculated an average height of 1.5 m, including some children only a little more than 1 m tall. Few calculated heights are 1.75 m or more. Early Europeans described Hawaiian warriors as tall, one missionary estimating an average height of 1.78 m. A footprint may be larger than a foot, particularly in slippery, wet ash, so our estimates of heights are probably somewhat too large. The data indicate that most of the footprints were made by women and children, not by men, much less warriors. We traced the footprint-bearing ash into the tephra section on the southwest side of Kilauea's caldera. It occurs high in the section, resting on older explosive deposits. Its surface is indented by small lithic lapilli, which fell into the ash while it was still wet; a few even landed in footprints. The lithic lapilli are at the edge of a thick block and lapilli deposit that fell from a high eruption column; the column reached well into the jet stream, because its fallout was mainly dispersed east-southeastward by westerlies, a wind direction found only at high altitudes in Hawai'i. Surges associated with the high eruption column swept over the southwest and west rims of the caldera. These relations indicate that the accretionary lapilli (footprints) ash was an early stage of a powerful eruption involving both high columns and lithic surges. Hawaiian oral tradition says that the 1790 eruption was large, and Jaggar calculated a column height probably greater than 9 km (30,000 ft) based on observations of a pillar (eruption column) seen over Mauna Loa when viewed from the north. This is about halfway through the jet stream. Our work found two deposits of the late 1700s dispersed east of Kilauea's summit. The younger was probably erupted in 1790. A reconstruction of events in 1790 suggests that the accretionary lapilli ash fell early in the eruption, blown southwestward into areas where family groups, mainly women and children, were chipping glass from old pahoehoe for tools. They probably sought shelter while the ash was falling. but once it stopped, they slogged through the mud, leaving footprints in the 2-cm-thick deposit.. Meanwhile, the warriors and their families, camped at Kilauea's summit (supposedly for 3 days) waiting for the eruption to end, saw the sky clear following the ash eruption and started walking southwestward along the west side of the summit area. Then the most powerful stage of the eruption began, sending surges westward across the path of the doomed group, killing many. Afterwards, any survivors or rescuers who walked on the accretionary lapilli ash, by now dry, left no footprints that are preserved.
V11B-2023
Surface deformation dynamics of Mauna Loa and Kilauea volcanoes, Hawaii, revealed by InSAR time series analysis
We exploit the advanced Interferometric Synthetic Aperture Radar (InSAR) technique referred to as the Small BAseline Subset (SBAS) algorithm to analyze surface deformation of Mauna Loa and Kilauea volcanoes, Hawaii. In particular, we present time series of line-of-sight (LOS) displacements derived from SAR data acquired by the ASAR instrument, on board the ENVISAT satellite, from ascending (mode 2, track 93) and descending (mode 2, track 429) orbits between 2003 and 2008. For each coherent pixel of the radar images we compute time-dependent surface displacements as well as the average LOS deformation rate. We also benefit from the use of the ascending and descending data to discriminate the vertical and east-west components of the displacements. The derived InSAR measurements are also compared to continuous GPS data to asses the quality of the SBAS-InSAR products. Our results quantify, in space and time, the complex deformation of Mauna Loa and Kilauea volcanoes. In addition, the SBAS-InSAR time series can be used to model both magmatic and tectonic processes, leading to a better understanding of volcano and earthquake processes on the Island of Hawaii.
V11B-2024
The Hartousov Mofette Field in the Cheb Basin, Western Eger Rift (Czech Republic/Europe): a Comperative Geoelectric, Sedimentologic and Soil Gas Study of a Magmatic Diffuse Degassing Structure
The mofette field of Hartoušov is located in the Cheb Basin, a shallow Neogene intracontinental basin in Central Europe. The north-eastern part of the Cheb Basin is one of the most seismically active regions of Central Europe. Seismic activity in the Cheb Basin has mainly a swarm-like character. The numerous cold CO2 emanations (>99 Vol.% CO2) at the surface of the basin are supposed to be generally connected to the seismic activity and to stem from the upper mantle. The Hartoušov mofette field has been investigated by combining geophysical measurements (geoelectrical resistivity tomography, self potential) with sedimentological studies (grain size, Corg, mineralogy) and soil gas (CO2 flux and CO2 concentration) data. Key question of the research was to evaluate the structural and sedimentological control at a CO2 degassing location. The investigations reveal a positive correlation between areas of high soil gas (CO2) concentration and flux with geophysical anomalies (negative self potential, positive structures of low electrical resistivity) as well as with specific sediment properties (content of pyrite and organic material, occurrence of dispersed pebbles, uplifted clay layer). These features are thought to be directly or indirectly related to the magmatic caused CO2 flow. Soil gas (CO2) measurements indicate areas of high CO2 content to be marked by anomalous vegetation patterns. These anomalies spread out with a linear trend, suggesting a fault control on gas ascent. Places of highest gas flow form small hummocks, with minor depressions on top. Negative geoelectrical self potentials at such locations were interpreted considering as having been caused by a downward movement of the meteoric water balancing the upward CO2 flux. The top of a pre-Quaternary clay-rich unit with a high content of smectite is highest in the location nearest to the mofette showing the most intense CO2 emanation. Most probably the clays form a domal feature below this mofette, as confirmed by the 3-D geoelectric measurements by low electrical resistivities. The driving force behind the updoming of the clays might be the pressure of uprising CO2. Additionally, the more intense swelling of smectite due to higher rates of fluid flow at these locations might also contribute to this phenomenon. Isolated quartz pebbles dispersed in fine-grained sediments could have been transported upward by gas jets bonded to vents during periods or events of intense gas emanation. The model for the sedimentation at the active mofettes has to consider its bonding to deep-seated faults, the presence of sediment deformation structures due to gas pressure, upward transport of sediment particles by gas jets and reducing conditions caused by the magmatic CO2 flux.
V11B-2025
Soil Properties Affecting the Reductive Capacity of Volcanic Ash Soils in Korea
Volcanic ash soils or Andisols have distinct chemical and mineralogical properties. The unique chemical properties of Andisols are due to their Al-rich elemental composition, the highly reactive nature of their colloidal fractions, and their large surface area. The soils that developed from volcanic ash on Jeju Island, Korea, were classified as typical Andisols. The soils had an acidic pH, high water content, high organic matter, and clay-silty texture. The crystalline minerals in the samples were mainly ferromagnesian minerals, such as olivine and pyroxene, and iron oxides, such as magnetite and hematite derived from basaltic materials. A large amount of gibbsite was found in the subsurface horizon as a secondary product of the migration of excess Al. In addition, we found that considerable amounts of poorly ordered minerals like allophane and ferrihydrite were present in the Jeju soils. The SiO2 contents were lower than those of other soil orders, while the Al2O3 and Fe2O3 contents were higher. These results reflect some of the important chemical properties of Andisols. The chromium (VI/III) redox couple was used in the reductive capacity measurement. The mean reductive capacity of the Jeju soils was 6.53 mg/L reduced Cr(VI), which is 5.1 times higher than that of non-volcanic ash soils from inland Korea. The reductive capacity of the inland soils was correlated with the total carbon content. Such a high capacity for the reduction of soluble Cr(VI) must also be due to the relatively high carbon contents of the Jeju soils. Nevertheless, despite having 20 times higher total carbon contents, there was no correlation between the reductive capacity of the Jeju soils and the carbon content. These results imply that the reductive capacity of Jeju soils is not only controlled by the carbon content, but is also affected by other soil properties. Correlations of the reductive capacity with major elements showed that Al and Fe were closely connected to the reductive capacity of Jeju soils. In addition to the carbon content, other factors controlling the reductive capacity of Jeju soils may have a close relationship with the properties of andic soils, which contain considerable amounts of allophane and ferrihydrite. Andic soils offer distinct advantages regarding a high reductive capacity for Cr(VI), such as a very large surface area and unique adsorption characteristics. The adsorption of anions by Andisols results from chemisorption on ¡®active' Al and Fe occurring in various forms, such as allophane, organometallic complexes, and ferrihydrite. In addition, since Andisols have a predominantly positive surface charge at soil pH, the soils can retain mobile anions, such as phosphate, nitrate, and chromate. The positive net charge is derived from the large amount of organic C and allophane. Therefore, the high reactive capacity of the Jeju soils suggests that the combined effects of a large number of electron donors, such as organic matter and Fe(II), and the adsorption characteristics of Andisols are complex and should be considered simultaneously when estimating the reductive capacity of andic soils.
V11B-2026
40Ar/39Ar ages of the older eruptive units of Somma-Vesuvius volcano, Italy
40Ar/ 39Ar ages have been measured on the older major explosive eruptions of Somma-Vesuvius
volcano in Italy. These eruptions all have pumice fall, and pyroclastic surge and flow deposits. The eruptive
history of Somma-Vesuvius volcano has previously been based on uncalibrated 14C ages, mostly on
carbon from paleosols, reported by Delibrias and others (1979) and Sigurdsson and others (1985). These
assigned ages, plus measured 14C ages, calibrated 14C ages (denoted as 14C*) and
40Ar/ 39Ar ages (all in years) are:
Mercato Tuff ~8500 years
14C=8263±29, 14C*=9250±49, 40Ar/
39Ar=9155±461, 9541±460
Verde Tuff ~15000 years
14C=14420±130, 14C*=17200±380, 40Ar/
39Ar=18456±302
Pomici di Base ~17000 years
14C=17229±398, 14C*=20360±139, 40Ar/
39Ar=21759±306, 21568±328
The assigned ages are from Delibrias (1979). The 14C ages (Aleisso and others, 1971, 1973, 1974,
1978) use 1σ counting errors reported by laboratories. The uncertainty in calibrated 14C ages
(above) are these reported uncertainties plus 1σ uncertainties of the calibration curves in IntCal04
(Reimer and others, 2004). The uncertainties in 40Ar/ 39Ar ages are 1σ errors. Pomici di
Base and Mercato samples were analyzed twice. The weighted mean of plateau ages for Pomici di Base fall
unit is 21,670 ± 224 years, and the weighted mean of isochron ages is 21,313 ± 408 years. The
weighted mean of plateau ages for Mercato Tuff is 9348 ± 326 years, and the weighted mean of
isochron ages is 9577 ± 332 years. The period of time from 0-12.4 ka used in calibrating 14C ages
is based on dendrochronologically-dated tree ring samples. The calibration older than 12.4 ka is based on
marine samples, primarily corals, and the marine calibrations are in dispute. The calibrated 14C ages
above are based on IntCal04. Chui and others (2007) have presented another calibration based on fossil
corals. In the younger part of their record, less than 30,000 years, the age difference averages less than
100 years.
Aleisso and others, 1971, 1973, 1974, and 1978, University of Rome Carbon-14 Dates IX: Radiocarbon, v.
13, 15, 16, and 20.
Chiu, T.C., Fairbanks, R.G., Cal, L., and Mortlock, R.A., 2007, Analysis of the atmospheric 14C record
spanning the past 50,000 years derived from high-precision 230Th/234U/238U,
231Pa/235U, and 14C dates on corals: Quaternary Science Reviews, 26:18-36.
Delibrias, G., DiPaola, G.M., Rosi, M., and Santacroce, R., 1979, La Storia Eruttiva Del Complesso Volcanico
Somma Vesuvio Ricostruita Dalle Successioni Piroclastiche Del Monte Somma; Rendiconti Societa Italiana di
Mineralogia e Petrologia, 35:411-438.
Reimer, P.J.and others, 2004, IntCal04 Terrestrial Radiocarbon Age Calibration, 0-26 Kyr BP: Radiocarbon,
46:1029-1058.
Sigurdsson, H., Carey, S., Cornell, W., and Pesdatore, T., 1985, The Eruption of Vesuvius in A.D. 79, 1985:
National Geographic Research, 1:332-387.
V11B-2027
Real-time radon monitoring at Stromboli volcano: influence of environmental parameters on 222Rn degassing
Two real-time stations for radon monitoring are currently operative at Stromboli volcano. The 222Rn electronic dosimeters are interfaced with an electronic board connected to a radiomodem for wireless data transfer (through a directional antenna) to a receiving station at the volcano observatory (COA). Radon activity data and enviromental parameters (soil temperature and atmospheric pressure) are sampled every 15 minutes and are instantaneously elaborated and transferred via web so that they can be checked in remote. Collected time series show that there is an overall inverse correlation between radon emissions and seasonal temperature variations. Signal processing analysis show that radon emissions in sectors of diffuse degassing are modulated by tidal forces as well. In addition, radon activities recorded at the summit station, located along the summit fracture zone where the gas flux is concentrated, are positively correlated with changes in atmospheric pressure and confirm the occurrence of the 'atmospheric stack effect'. It is not excluded that this process may play an active role in modulating Stromboli explosivity.
V11B-2028
MODIS thermal anomalies during strombolian activity at Stromboli
We investigated the spectral radiance of Stromboli volcano through a time-series of more than 1200 nighttimes and daytimes MODIS granules. MODIS images allowed us to analyse the infrared radiation of Stromboli volcano during the period spanning between 1 January 2008 and 4 September 2008. MODIS data has been resampled within a spatial mask (15 km X 15 km) including the volcano. Then we applied a cloud mask algorithm and the NTI to the all pixels in order to detect the pixel containing the thermal anomalies. According to this principle, the MODVOLC algorithm detects an alert whenever a pixel has an NTI higher than a fixed threshold (-0.8); such thresholds were settled in order to avoid false alarms on a global scale. At Stromboli volcano a clear seasonal pattern is evident for the measured specific NTI trend and many thermal anomalies may be detected when this parameter is lower than -0.8, as well. Therefore we processed all data by means of an automatic routine without setting any fixed threshold. Since during typical strombolian activity, thermal anomalies are located within the crater area, we monitored the pixel with highest NTI within the spatial mask. We suggest that the thermal energy measured for these peaks can be better related to the amount of lava erupted, thus providing a tools to refine the estimate of the effusion rates during the typical strombolian activity.
V11B-2029
Inflation deflation cycles related to explosive activity at Stromboli volcano
Strombolian activity is mainly driven by gas dynamics and it is then very sensitive to the changes in gas flux
regime in the magmatic feeding systems. Gas dynamics is the main responsible for a large variety of physical
phenomena recorded at Stromboli, before, during and after each explosive event. We show how high
sampling of high resolution tilt signals reveals new insight of this volcanic activity.
Two Pinnacle 5000T electrolytic borehole tiltmeters have been installed in two 5 m deep boreholes on
Stromboli volcano (Italy), at a distance of <1 km from the active vents. Tiltmeters were sampled at a
frequency of 1 Hz and this sampling rate allows to record a large range of deformations, from magma
intrusion to mild strombolian activity. The instrument has a nominal resolution of ±1 nrad with a quite high
signal-to-noise ratio, which allows to detect earth tides, thermal ground deformations and to discriminate
single explosive events with an amplitude of only few nano-radians.
We present here data from 2005 to 2007 when the activity at Stromboli spanned from mild to vigorous and
was characterized by the effusive eruption of February-March 2007, culminating with the paroxysmal
explosion of March 15th. Ground deformation related to normal explosive activity (Strombolian), are recorded
as distinct repetitive waveforms, associated with the activity at different craters. Tilt data collected during this
periods show that at a shorter time-scale each single explosion is controlled by a continuous process of gas
charge and discharge within the conduit, which is detected as continuous ground inflation/deflation cycles.
During the major explosion of March 15th tiltmeters have detected a large inflation process almost 4-5
minutes before the visual onset of the explosion. Waveforms and deformation vectors indicate that the major
explosion was characterized by an non-isotropic source, well represented by a shallow dyke conduit system.
V11B-2030
Newly discovered submarine flank eruption at Stromboli volcano (Aeolian Islands, Italy)
The NW submarine portion of Stromboli volcano has been investigated by deep-towed sidescan sonar, bathymetric surveys, video camera runs and dredging during two research cruises in 2002 and 2004. The surveys resulted in the identification of an extensive pillow lava field (106-107 m3)at about 2300 m of water depth and 9 km from the shoreline of Stromboli Island. Marine geology data coupled with volcanological observation and volatile content in the glass, demonstrate that this lava flow-field is in place and is related to a recent submarine effusive eruption originating from a nearby eruptive fissure. Anyway bulk rock and glass composition of the lava differs significantly from present day summit products and from those erupted during the historical flank eruptions. Pillow lavas compositions show a high K2O content and a general geochemical pattern more similar to the less evolved K-alkaline products erupted between 13 and 6 ka during the Neostromboli period. However, the silica content is outside of the known compositional range of Neostromboli, hence the pillow lava could represent a not yet documented stage of the structural evolution of the volcano. The compositional and volcanological evolution within the subaerial Neostromboli products is relatively well constrained, including the early stages of the period; on the other hand, the passage between Neostromboli period and early stages of development (6-2 ka) of the Recent Stromboli period (after the Neostromboli collapse) are still poorly known. The unique composition of the pillow lavas is interpreted as the result of the interaction between magmas feeding the new cycle and batches of previous Neostromboli magmas, still present after the emptying of the Neostromboli magma chamber. In this perspective, the investigated flank eruption could be the 'missing link' of early Recent Stromboli evolution by representing the beginning of this new intrusive cycle. In addition to its structural and volcanological significance this eruption opens new scenarios in Stromboli volcanic hazard: while flank eruptions on the subaerial part of the volcano could have an obvious direct impact on inhabited areas, submarine eruptions have the potential to trigger slope instability. Processes of flank inflation, opening of the eruptive fissures and gravitational loading due to the lava flows piling-up, are all able to induce sediment destabilization and possible flank failure.
V11B-2031
3D Finite Element Meshing of Stromboli and Mount Etna
The development of monitoring networks both at Mount Etna and Stromboli provided a fairly detailed database of geodetic and seismological observations during the unrest and eruptive/explosive phases of the last few years. These data reveal a tight interaction between magmatic and seismic activities. Their interpretation requires a new generation of numerical models of the volcanic edifices, based on the finite element method and characterized by realistic topography, anelastic rheology, internal discontinuities and lateral variations of mechanical parameters. We focus on the problem to make a flawless spatial discretization, which is an essential step for an accurate finite element simulation. If 3D unstructured tetrahedral meshes can be achieved quite easily with commercial or non-commercial algorithms, the creation of 3D non-structured hexahedral meshes is still recognized as a challenging issue. For complex models, as in the case of realistic geological volumes, generating a hexahedral mesh with the available meshing algorithms can require weeks or even months. Nevertheless, all-hexahedral meshes are still eagerly requested and, in some cases, preferred to all-tetrahedral ones, mainly because of the superior numerical accuracy and stability, but also for the lower computational cost. Taking advantage of CUBIT (www.cubit.sandia.gov) and of the expertise acquired in the meshing process for seismological problem, we present the mesh of the two most active italian volcanoes: Stromboli and Mount Etna. The grids are based upon updated and detailed digital elevation models with a resolution of tenth of meters in the zones where the most significant deformations are observed and include the major structural discontinuities. An unstructured scheme is implemented in order to obtain a lower resolution away from the volcano summit.
V11B-2032
A Stochastic Model For Long-Term Eruption Forecasting At Open Conduit Volcanoes: Examples From Etna And Kilauea
Volcanic eruptions are complex processes, governed by a large number of degrees of freedom of comparable weight, often non-linearly related. Because of this, they are intrinsically unpredictable from a deterministic point of view. However, volcanic eruptions can be described as stochastic processes. In this work, we address the long-term eruption forecasting at two very active open conduit volcanoes, i.e. Etna and Kilauea, by means of a Bayesian Hierarchical Model (BHM) for past eruptive activity. In particular, we set up a BHM based on the generalized time-predictable relationship between erupted volumes and inter- event times. In the model, we properly take into account measurement and model uncertainties. We use part of the available catalogues for model set up, and the remaining part for model check. Since we want to test whether our BHM is able to capture the relationship between erupted volumes and inter-event times, we test the BHM performance against the null hypothesis of Poissonian occurrence. The results indicate that BHM performs better than Poisson model, thus it can potentially be used for long-term eruption forecasting at these two open conduit volcanoes, in a Probabilistic Volcanic Hazard Assessment perspective.
V11B-2033
1992-2000 InSAR time series analysis at Hekla volcano, Iceland: Inflation periods and the 2000 eruption co-eruptive changes.
The surface deformation signals of deep magma accumulation (> 10 km) can be difficult to constrain as the signals cover a large area (more the 30 km in diameter). Gradual deformation rates (< 10 mm/yr) over such large area is difficult to detect with conventional InSAR. New improved analyses of InSAR images, looking at series of multiple images focusing on pixels with low-phase variance (Persistent Scatterers, PS) open new possibilities. We have used these techniques to study Hekla Volcano in South Iceland. Hekla volcano is one the most active volcanoes in Iceland with 18 summit eruptions during the last 1100 years. It's most recent eruption occurred in February 26 to March 8, 2000 when about 0.11 km3 erupted to the surface. From earlier geodetic measurements, a deep magma source and a shallow dike have been inferred, connected by a narrow feeder channel. However, the deformation field observed in the vicinity of the volcano, before eruption, is complex and the characteristics of suggested deformation sources are uncertain. We use data, from ERS1-2 satellites, acquired from 1992 until 2000 to study crustal deformation in relation to the 2000 eruption and the inter-eruptive period since an earlier eruption in 1991. Prior to the eruption in the 1992-2000 period We find a broad area around the volcano (radius about 20 km), with satellite LOS line of sight shortening of up to 5 mm/yr. This is interpreted as inflation due to pressure increase in a deep seated magma chamber. In the closer vicinity of the volcano, within ~6 km radius from it, an average velocity of 7-8 mm/yr, (LOS) is detected, correlating partly with recent lava flows. After the 2000 eruption the inflation bulge appears to deflate. A local deformation signal is also observed next to the eruptive fissure. The scale of this deformation over limited area suggests the feeder dike of the 2000 eruptive fissure originates at shallow depth. Preliminary results, from Envisat time series from 2003 until 2008, show increased inflation rates around Hekla starting in mid year 2007.
V11B-2034
The Prehistoric Development of Hekla Volcano and its Modern Eruptive Style
Hekla is located in southern Iceland, at the intersection of the Eastern Volcanic Zone and the South Iceland Seismic Zone. It has erupted 18 times historically, mostly from a system of fissures. This study focuses on the prehistoric volcanic deposits that lie to the northeast of Hekla's summit carapace. A large fissure (~2 km long) and several smaller ones erupted most of the rocks in the area. The large fissure is the oldest feature in the area, although hyaloclastite fragments are found within the ejecta, which suggests that Hekla is constructed on a series of sub-glacial ridges. Whole rock major and trace element analyses indicate that the eruptive products on the NE ridge span a wider compositional range than those produced during Hekla's historic activity, extending to both more mafic and more felsic compositions (49 to 63 wt.% SiO2). The large fissure and a cone dissected by it erupted the most mafic compositions, but these basalts fall along the fractionation trends defined by historic Hekla lavas and tephra. The more felsic rocks were collected from the lower walls of the craters and fissures and form Hekla's infrastructure. The siliceous rocks define fractionation trends that deviate from the dominant Hekla trend. These results call into question models that invoke melting of the crust to produce Hekla's siliceous rocks. Furthermore, the progressively narrower compositional range of erupted material at Hekla may reflect the development of a long-lived magma chamber beneath the volcano that filters and partially homogenizes melts prior to eruption. A similar phenomenon has been proposed to explain the wide variety of compositions observed at the Pleistocene sub-glacial ridges that surround and form the base of the Hekla structure.
V11B-2035
Comparison of Historic and Prehistoric Tephra Deposits at Hekla Volcano, Iceland
Hekla is a volcano located in southern Iceland near the intersection of the South Iceland Seismic Zone and
the Eastern Volcanic Zone. Hekla has erupted 17 times since its first historic eruption in 1104 AD. Few
prehistoric lavas are exposed on the volcano, thus the prehistoric eruptive history of Hekla is mostly recorded
by tephra deposits. The prehistoric Plinian deposits at Hekla are much larger and more siliceous than the
historic effusive eruptions. Nevertheless, the historic and prehistoric eruptive products generally define
similar geochemical trends in major and trace element compositions, though some trace element contents do
vary. The H3 tephra deposit is ~3000 years old and has the largest volume of erupted material of any
known Hekla eruption. Typical of other Hekla eruptions, the most evolved eruptive products from the H3
eruption are at the bottom of the deposit, and they become progressively less siliceous upward, from at least
70% to ~63% SiO2. Incompatible trace elements in the most siliceous pumices are enriched by
~125% over the more mafic ones. Some of the pumices from the H3 eruption preserve swirled textures,
indicating incomplete mixing.
A distinctive 0.75 m thick tephra layer immediately overlies the H3 deposit at
the pumice quarry north of Hekla. The bottom half of the deposit is light in color with silica contents up to
64%. The top of the deposit is dark with silica contents as low as 56%. Unlike all the historical Hekla tephras
and H3, this deposit is strongly bimodal, changing abruptly from felsic in the bottom half of the section to
more mafic compositions in the top half; all known Hekla tephras are gradationally zoned. The tephra are
compositionally more akin to Hekla than known samples from Torfajokull or Katla.
The eruptive style at Hekla
has changed since the last Plinian eruption in 1104 AD. Historic eruptions are diminished in volume,
explosivity, and silica content of the most evolved tephra, all of which appear to correlate with length of the
repose period. These changes may reflect a shift in the fractionation processes in the magma chamber, or in
the magma supply itself. Another possibility is that the magma chamber now intercepts all the magmas,
creating a shadow zone, whereas in the past, the chambers were ephemeral.
V11B-2036
Ongoing inflation and magma accumulation of Grimsvotn subglacial volcano, Iceland
The sub-glacial Grimsvotn volcano, one of Iceland's most active volcanoes, erupted in 1983, 1998 and 2004. Since 1998, annual GPS measurements have been conducted at the only available nunatak at the volcano, located on the rim of its caldera. A clear pattern of deformation is observed that can be attributed to magma inflow and outflow, uplift due to glacial thinning, and plate movements. The different effects can be separated by considering the horizontal and vertical displacements and their temporal evolution. Horizontal displacement away from the caldera and uplift is caused by inflow of magma into a shallow magma chamber, and rapid deflation in reverse direction occurs during eruptions. This "saw-tooth" pattern is superimposed on uplift due to glacio-isostacy, and plate movements. The magma chamber is suggested to be at shallow depth in the centre of the caldera complex. Deformation of magmatic origin in 1998-2004 indicates four phases; contraction in connection with the eruption, followed by a high rate of inflation in a recharge phase with the duration of 1.5 -2 years. A following intermediate phase had a lower rate of magma accumulation. The deformation rate increased 1-1.5 years prior to the 2004 eruption with associated increase in seismic activity. Eruptions in 1998 and 2004 appear to have occurred at similar pressure in the shallow magma chamber. Continuing inflation of the volcano suggests an eventual eruption within a few years. Seismicity is expected to increase months or a year before the eruption, and then finally an intense earthquake swarm associated with the formation of a feeder dyke is anticipated to give a short-term warning of the next eruption.
V11B-2037
New Isopach Maps of Holocene Rhyolitic Tephras at Medicine Lake Volcano, Northern California
Medicine Lake Volcano, located in the southern Cascades ~55 km east-northeast of Mount Shasta, is a large rear-arc, shield-shaped volcano with an eruptive history spanning nearly 500 ka (Donnelly-Nolan, et al., 2008). The most recent eruptions at Medicine Lake Volcano are the late Holocene explosive to effusive events at Glass Mountain (~950 yr) and Little Glass Mountain (~1000 yr), which began as sub- Plinian to Plinian eruptions of rhyolite pumice from fissure vents (Heiken, 1978), and culminated in the rhyolite-dacite of Glass Mountain and the rhyolite of Little Glass Mountain. Vents for these eruptions are located 15 km apart on opposite sides of the summit caldera of Medicine Lake Volcano. Glass Mountain erupted from a 5-km-long fissure on the east side and Little Glass Mountain from an 8-km-long fissure on the west side of the volcano. New isopach maps of tephra deposits from these eruptions are based on more extensive fieldwork and on a different interpretation of Little Glass Mountain tephras than previous work. The maps show a strong northeast-southwest trend of the Little Glass Mountain tephras as previously shown by Fisher (1964) and by Heiken (1978), extending in the direction of Mount Shasta, where D. Miller found individual lapilli from the Little Glass Mountain eruption. Tephras from Glass Mountain are not deposited along a single strong trend, but rather are found in lobes extending from the fissure vents to the west, north, and northeast. More than 2 m of Little Glass Mountain tephra was deposited proximal to its vents, although the deposit thins quickly to less than 50 cm within 2 km perpendicular to trend and within 7 km along trend. The maximum observed thickness of Glass Mountain tephra is between 7-8 m proximal to its vents, thinning to less than 50 cm within 7 km distance from the vents. As rhyolite eruptions in the Cascade Range are quite rare, mapping the thickness, extent, and character of these tephra deposits to better comprehend the dynamics of these eruptions is important for understanding future events and the hazards they may pose, as well as the possibly analogous eruption of Chaiten Volcano in Chile, the only modern rhyolitic eruption of similar scale.
V11B-2038
Environmental and Compositional Controls on the Texture and Composition of Palagonitized Hyaloclastites
Palagonitization is a widespread geochemical process in which sideromelane is converted to hydrous alteration products (palagonite and a variety of authigenic minerals) in subaqueous environments that produce volcanic glass. Recognizing that an 'aging' process may influence the textural and compositional properties of palagonite over time, we have been investigating the direct effects of specific environmental conditions (e.g. glass and fluid composition, fluid/rock ratio) on palagonite formation. For this study, hyaloclastite samples inferred to have palagonitized in relatively low fluid/rock ("closed") environments (submarine volcano flanks, submarine volcaniclastic basins) as well as from relatively high fluid/rock ("open") environments (tuff cones, subaqueous and subglacial eruption sites) were subjected to petrographic, electron microprobe, and LA-ICP-MS analysis. Palagonite rind textures are broadly consistent at each sampled site, but there is a wide range of textural variation between the sites. Rind thickness in all our samples varies from 0.05 to 1.0 mm, and the thicker rinds are generally from localities where it is believed that the aging of palagonite glass has occurred over a relatively long time. Authigenic zeolites are much more common in samples palagonitized in lower fluid/rock environments. Inferred water content of palagonite in our samples varies from 17 to 37 weight percent and appears to be highest from tuff cones, intermediate from subaqueous eruption sites, and lowest from submarine volcaniclastic basins. In submarine volcaniclastic sandstones sampled offshore of Kilauea and Mauna Loa volcanoes, palagonite rind compositions from adjacent grains of alkalic and tholeiitic sideromelane are distinguishable and reflect the original glass composition. Palagonite REE patterns parallel those of adjacent sideromelane, but the REE concentrations are higher in the palagonite and increase systematically towards the outermost palagonitized rim. Gradients in the Sr-isotopic composition of sideromelane palagonitized in submarine environments follow a mixing relationship between seawater and sideromelane initial values, and imply that palagonitization is a process that proceeds inwards with time. These observations suggest that the initial glass and fluid composition, as well as the fluid/rock ratio of the palagonitization environment, influence the textural and compositional properties of palagonite, and that palagonites formed in relatively low vs. high fluid/rock environments have distinguishable characteristics.
V11B-2039
SIMULATING THE SPATIAL, SPECTRAL, AND RADIOMETRIC RESPONSE OF SOME SATELLITE REMOTE SENSING INSTRUMENTS COMMONLY USED TO STUDY ACTIVE LAVAS
Satellite remote sensing provides a means to quantify active volcanism at regular, repeated temporal intervals, at a range of spatial scales, and using electromagnetic radiation at a range of visible and infrared wavelengths. Using data acquired by low spatial resolution, but high temporal resolution environmental monitoring satellites (e.g. AVHRR, MODIS, GOES) and high spatial, but low temporal resolution, sensors (e.g. Landsat Thematic Mapper, Terra ASTER), volcanologists have developed techniques for determining lava surface temperature, lava cooling rates, and lava effusion rate, amongst other things. As with all measurement techniques the imaging process acts as a filter, the exact nature of the filtering being determined by the characteristics of each individual sensor (i.e. instantaneous field of view of the imager; quantization of the data; spectral responsivity of the detectors). Volcanologists currently use many remote sensing instruments to quantify volcanic activity, and the fidelity of the imaging process (i.e. how accurately scene content is recorded in the image data), will vary from instrument to instrument. In this presentation, we simulate the response of some commonly used satellite remote sensing instruments to real lava flows. We do this by convolving the spectral radiance emitted from a series of benchmark lava flow targets with the point spread functions (electronic, optical, detector, image motion) of the instruments in question. Our benchmark data are generated by using stochastic analysis of high spatial resolution (i.e. mm scale) thermal video camera images of real lava flows. By varying the spatial and thermal characteristics of a set of simulated target scenes we demonstrate how well commonly used remote sensing instruments can be expected to record spatial and temporal variations in the properties of an active lava flow field. Our results have implications for the use of low spatial resolution remote sensing data to document effusive volcanic eruptions on Earth, and on the Jovian moon, Io.
V11B-2040
Very Long Period Oscillations Observed at Iwojima, Japan
Iwojima, located at 1250 km to the south of Tokyo, Japan, is a volcanic island belonging to the Izu-Ogasawara arc. The island has a volcanic cone Suribachiyama at the southwest and a dome-like mountain Motoyama at the northeast, which is the central part of the caldera with the diameter of about 10 km. Motoyama repeats large-scale uplift every several years, and has been uplifting with the rate of more than 50 cm/year since August 2006. During the large uplift period, we observed two types of very long period seismic signals in the island. The one is a very long period earthquake (VLP earthquake), which has a short period signal at the initial part and a damped long period oscillation lasting 10 minutes. The other is a very long period tremor (VLP tremor), which lasts a few days. These seismic signals have the same nine dominant spectral peaks between 6.9s and 29.8s, and exhibit the same oscillation directions at each seismic station. The agreement of the dominant spectral peaks shows that they are caused by eigen oscillation of the same resonator. VLP earthquakes are probably originated from the damped oscillation of the resonator excited by relatively large shallow short period earthquakes. The hypocenters of the short period seismic signals at the initial phase are distributed at the shallow part of Iwojima(<2km) as well as other short period earthquakes beneath the island, but their magnitude of M~3 is larger than the other earthquakes(M<2.4). Excitation source of the resonator for VLP tremor is unknown, but we suppose continuous pressure source such as volcanic fluid flow possibly excites long lasting oscillation of the resonator. We propose that the resonator is a magma chamber beneath Motoyama on the basis of the following reasons. i) The oscillation directions are similar to the directions of displacement vectors of GPS at the seismic stations during the large uplift. The similarity suggests that an inflation source oscillates in the radial direction. ii) The nine eigen periods of the oscillation can be explained by oscillation of a low velocity fluid sphere in an infinite elastic medium (Sakuraba et al., 2002). The density and P wave velocity ratios between the sphere and the infinite medium are 1 and 6.5, respectively. iii) The diameter of the sphere should be larger than several km to explain the very long period. iv) The sphere is located at the center of Motoyama (the center of the caldera) about 6 km deep. The estimated resonator is probably a main magma chamber which caused the roof collapse during the formation of Iwojima caldera. Acknowledgements: We thank Ministry of Defense for providing us with seismometer and tiltmeter data.
V11B-2041
Ground deformation of Suwanose-jima volcano as viewed from ALOS/PALSAR InSAR
Suwanose-jima is one of the most active volcano of Japan with quasi-continuous unrest since 1957. It is monitored with broadband seismometers and tiltmeters but the difficulty in ground access to the island prohibits us to construct large-enough network to understand the magma plumbing system solely from the ground-based monitoring network. Groud deformation observed by the InSAR analysis thus has a potential to gain more insights into our understanding of the magma plumbing system of this volcano. We analyzed 13 images (6 ascending and 7 descending images, respectively) taken between March, 2007, and February, 2008. Despite 70 explosive eruptions in 2007, no deformation was detected for interferograms from all possible pairs except for small (several centimeters of line-of-sight changes) deformaiton in an interferogram between August 17, 2007, and January 2, 2008, which could represent deformation due to explosive eruptions in December, 2007. Explosive eruptions without significant deformation can be interpreted as eruptions without feeding magma from depth or magma propagation through the conduit without deformation.
V11B-2042
Effects of Host Rock Stratigraphy on the Formation of Ring-Faults and the Initiation of Caldera Collapse
Most collapse calderas can be attributed to subsidence of the magma chamber roof along bounding sub- vertical normal faults (ring-faults) after a decompression of the magma chamber, following eruption. It has previously been shown that for ring-faults to initiate, and thus facilitate collapse, the stress field both at surface and around the magma chamber must satisfy specific critical conditions. Here, we present the results of new numerical models that use a Finite Element Method to investigate the effects of variable stratigraphy (lithology/thickness/order of strata) on local stress field distribution. Results are then compared with existing criteria for ring-fault initiation. Different subsurface scenarios were simulated by varying the stiffness (Young's modulus) of seven thin crustal layers placed above the magma chamber, and the host rock in which the chamber is seated. We consider the magma chamber to be subjected to a magmatic under-pressure of -15MPa, imposed at the chamber walls, in order to simulate magma withdrawal. Results indicate that for a given geometrical set-up, the magnitude and position of maximum tensional stress at the Earth's surface are influenced by the occurrence and relative distribution of mechanically stiff or soft lithologies above the magma chamber. For example, tensional stress at surface may be reduced by the presence of stiff layers (e.g. lavas), or increased by soft layers (e.g. pyroclastic units) compared to generic simulations using a homogeneous background medium. The observations suggest that the mechanical properties of crustal stratigraphy are therefore a further variable in the rare achievement of the stress field conditions required for ring-fault formation, and may be influential in generating or inhibiting caldera collapse. Therefore an understanding of pre-caldera stratigraphy could provide important insight into the likelihood of future caldera collapse events.
V11B-2043
Insights on Post-Emplacement Deformation of the Bandelier Tuff, Rio Grande Rift, New Mexico
Many studies of the Tshirege Member of the Bandelier Tuff (1.25 Ma), erupted from the Valles caldera in New Mexico, have attributed deformation of the cooling units almost exclusively to tectonic structures of the Rio Grande rift. We present data collected from a field site, a large excavation at Los Alamos National Laboratory, which indicates the faults, fractures, and fissures within the Bandelier Tuff at this location developed through degassing and cooling of the tuff. The tuff exhibits fissures that are commonly funnel- shaped and upward-flaring, varying in width from centimeters to meters, some with fines-depleted textures characteristic of fossil fumaroles. The unit overlying the funnel-shaped fissures shows intense fracturing in all cases, and units over the fissures also demonstrate evidence of fumarolic alteration through the presence of clays and the reduction of cristobalite and feldspars in the fractured zone. These fossil fumaroles are the focus of most observed faulting with vertical displacements up to 36 cm. Comprehensive analysis of fractures indicates orientations do not match a regional structural grain. All data indicate that the lower unit began actively degassing following emplacement and continued fumarolic activity during the early stages of cooling. Subsequent deposition of younger pyroclastic units sealed the underlying fumarolic pipes. Localized overpressuring, and perhaps some small-scale phreatic explosive activity, occurred, causing partial opening of the fumarolic pipes and disrupting the pipe interiors and the overlying units. Field relations require that all of the units had to be sufficiently cooled by this time to deform in a brittle manner. Thus, the sites of fumarolic activity provided weaker spots in the cooling tuff that accommodated differential settling of the deposits in the form of small-displacement faults. Earliest-formed fractures typically associate with gas-escape structures such as fumaroles, which occur with observable alteration and mineralization. Degassing of ash-flow tuffs may be short lived, but studies of similar thick (100 m) pyroclastic deposits in Nevada show that tuff compaction and welding cease, in most cases, within about 10 years of emplacement. Based on these previous studies, cooling of the upper unit at our field site to temperatures allowing brittle failure may have only taken weeks to months as the unit is only 5-10 m thick. Thus, the fissures, fumaroles, faults and fractures observed in the excavation all likely formed in the first 10 years, and possibly in only a few months, after deposition of the Tshirege Member of the Bandelier Tuff while it degassed, cooled, and compacted.
V11B-2044
Sem Analysis of particles from the 28, 000 B.P El Zaguan debris avalanche deposit, Nevado de Toluca volcano, Central Mexico: evidences of flow behavior during emplacement
The Zaguan deposit originated at 28, 000 yr. B.P from the flank collapse of the Nevado de Toluca volcano, a dacitic stratovolcano of the Transmexican Volcanic Belt. A Scanning Electron Microprobe analysis (SEM) was made to some clasts of this deposit to observe microtextures produced during transport and emplacement of the debris avalanche flow. Particles from 2, 0 and -2 Φ granulometric classes were randomly selected and their surface textures were described. The textures observed were divided in two groups, collision and shear structures indicating different clast interaction. Shear textures were observed predominantly on the basal part of the deposit and consisted of parallel ridges, parallel grooves, scratches and lips. Collision textures were mainly present in the upper part of the deposit and consisted of fractures, percussion marks, and broken or grinded crystals. These characteristics, coupled with field observation, like the presence of clast dikes and deformed lacustrine megaclasts, indicate that the basal part of the debris avalanche was moving in a partially liquefied state, were particles were not able to move freely because of the confinement exerted by the upper part of the flow, so shear stresses dominated. On the contrary, the particles in the upper part were able to move freely so the principal mechanism of interaction between particles was collision. These microscopic textures are in agreement with previously described behavior of emplacement of debris avalanches of volcanic origin, that suggest a stratified flow dominated by different transport and depositional mechanism depending on flow depth and possible fluid content at their base.
V11B-2045
Structure and change of Piton de la Fournaise volcano inferred from gravity surveys (Reunion Island, Indian Ocean)
A new gravity map of Piton de la Fournaise volcano has been established using new on land and offshore measurements. The data coverage allows for the differentiation of shallow and deeper structures. 3D and 2 3/4 D models have been calculated. Short wavelength positive anomalies depict the presence of piles of thick dense lava flows filling volcano-tectonic depressions. The lateral extension and the depth of paleo- depressions associated with the collapses of the Plaine des Sables-Fond de la Rivière de l'Est and of the Enclos Fouqué are thus estimated. The negative short wavelength of the Central Cone suggests it has been built by thin, highly vesisculated and fractured lava flows. Low density hydrothermally altered rock beneath the summit can also contribute to the gravity low as well as a column of fractured rocks between the surface collapse and a magma reservoir. Negative short to medium wavelength anomalies have been observed in the Rivière des Remparts-Rivière Langevin zone and above the offshore continuation of the NE and SE rift zones. We speculate that the former zone is underlain by breccias related to erosion or mass- wasting events. The offshore continuation of the rift zones is most likely built by hyaloclastites and pillow lavas. Two main deeper dense structures exist: the Grand Brûlé complex and a complex beneath the Plaine des Sables and part of the Enclos. From a deep drill-hole it has been established that the Grand Brûlé complex is a hypovolcanic complex of intrusions and cumulates. We show that this structure is disconnected from the present day Piton de la Fournaise volcano. Its interpretation as the hypovolcanic complex of the ancient concealed Les Alizés volcano remains valid. The similar nature of the other dense complex is inferred by analogy with comparable anomalies in this geological context and by the presence of frequent gabbro and peridotite xenoliths in eruptive products in this area. We suggest that this complex has developed during the Ancient Shield period of Piton de la Fournaise. The fact that this structure is apparently not offset by the fault that limits the Enclos to the west suggests that the latter is a listric fault. Observed apparent contradictions between the seismic tomographies and the gravity pattern could be resolved considering the relative sensitivity of each method. Strong shallow gravity signals significantly hide moderate signals from deeper structures. Conversely, with a moderate station and signal coverage, the seismic tomographies fail to define precisely the subsurface structures. The use of both methods increases the accuracy in the determination of the internal structure of volcanoes. Pre and post April 2007 eruptive and volcano-tectonic crisis data show a significant gravity change in the central area. The observed mass deficit can be explained by creation of the new Dolomieu crater only or by the sum of the effects of the crater formation and of the fracturing of a column of rocks between the surface and the drained magma reservoir.
V11B-2046
Submarine structure of Reunion Island (Indian Ocean) inferred from gravity
La Reunion is a large (diameter: 220 km; height: 7 km), mostly immerged (97%) oceanic volcanic system. New land and marine gravity data are used to study the structure of its submarine part. The gravity models are interpreted jointly with the published geology interpretations and compared with magnetic models. This allows us to derive a new model of the shallow and internal structure of the submarine flanks. Recent cruises have collected high quality gravity, magnetic and multi-beam swath bathymetry data over the submarine flanks of La Réunion and the surrounding oceanic plate. A new Bouguer anomaly map has been computed for a reduction density of 2.67.103 kg m-3. A magnetic anomalies map covering the same area has been also built. Studies based on bathymetric and acoustic data have previously shown the presence of different types of submarine features: a coastal shelf, huge bulges built by debris avalanches and sediment deposits, erosion canyons, volcanic constructions near the coast, isolated seamounts offshore, and elongate volcanic ridges on the Mascarene plate. On the new Bouguer anomaly map, all these features are associated with negative anomalies. They have been modeled using 2 3/4 D modeling techniques. The short wavelength anomalies over the coastal shelf area can be explained by piles of low density layers. This suggests that they are mostly built by hyaloclastites which are generally characterized by lower densities than lava flows. The voluminous debris avalanche deposits which formed the huge Submarine Bulges to the east, north, west, and south of the island have also been modeled as low density formations. Each bulge is modeled with an overall density less than 2.67.103 kg m-3, in order to account for its long wavelength anomaly. Some shorter wavelength features are superimposed on these long wavelength negative anomalies. They probably represent heterogeneities within the bulges. Some shallow ones can be associated with observed surface geological features such as secondary landslide deposits or sedimentary areas. The interpretation of the negative short wavelength gravity anomalies associated with the seamounts and the volcanic ridges is less simple. Available data indicate that they are more likely built by lava flows than by low density pyroclastites or breccias. The observed negative anomalies can be accounted for if we consider an isostatic compensation of the surface constructions. Finally, a joint analysis of the gravity and magnetic anomalies allows us to refine the structural models of particular areas. Reversely magnetized formations near the seashore can be associated with volcanic constructions. Conversely, the deeper submarine flanks are poorly magnetized. This is coherent with the interpretation of the bulges as piles of mass-wasting deposits. 3D models have been then calculated in order to provide a first insight of the internal structure of the Reunion complex submarine part.
V11B-2047
Insights into the dynamic processes of the 2007 Stromboli eruption
The unrest of Stromboli volcano leading to the February 27-April 2, 2007 eruptive period and the March 15 paroxysm is constrained by combining broadband seismic data and 1-Hz GPS (High Rate GPS – hereinafter HRGPS) measurements. During the preeruptive stage, the simultaneous examination of seismic and HRGPS data, together with weather parameters, suggests the possible influence of external perturbations on the magmatic system, which evolved toward a critical state after January 2007. At the beginning of the eruption striking variations of the waveform and spectral content of very long period (VLP) events were observed. Moreover, some days after the onset of the eruption, a sudden change of the seismic and eruptive behaviour was recognized, while ground deformation began to show a deflation. The March 15 paroxysm was preceded, some days before, by a peak of the HRGPS spectral power densities (a small inflation) and by the variation of location and features of long period (LP) events together with the occurrence of a few VT earthquakes located at depths down to 3.5 km b.s.l.. These findings constrain, for the first time at Stromboli volcano, the deep origin of a fast rising magma batch, rich in gas, that led to a strong explosive event, and highlight a strict relationship between VLP seismicity and the eruptive activity.