V33D-1484 1340h
An Approach to Simulation of the Turbulent Multifield/Multiphase Dynamics of Volcanic Conduits and Plumes
The shape and size of volcanic eruption conduits play a large role in determining the character of eruption phenomena. Conduits for explosive eruptions evolve with time during the eruption greatly affecting mass fluxes and the fate of degassed volatile constituents. Previous modeling has focused on compressible fluid dynamics of gas and solid particle mixtures moving within a cylindrical conduit and expanding into the atmosphere. However, it is clear that evolution of the conduit itself is influenced by the erupting mixture and the flow field of the mixture is controlled in part by the conduit. In order to address this important aspect of eruption dynamics, both compressible fluid flow and solid mechanics must be solved simultaneously. Two different three-dimensional state of the art computer codes have been developed at Los Alamos National Laboratories that are capable of simulating these dynamics together, CFDLib and SAGE. CFDLib is a compilation of well-tested computational fluid dynamics approaches suited for a wide range of fluid and solid dynamics, using well-known Marker-And-Cell (MAC) and Implicit Continuous-fluid Eulerian (ICE) techniques. SAGE employs adaptive grid Eulerian techniques to provide local areas of high resolution for dynamics of complex materials. We are benchmarking and validating these codes for geophysical application as a preliminary step toward modeling the three-dimensional dynamics of volcanic conduits.
V33D-1485 1340h
Experimental Insights into the role of Dynamic Gas-slug Expansion on the Nature of Strombolian Eruptive Activity
The exsolution of gaseous species to form bubbles is common as magma approaches the Earth's surface. Under certain conditions of time-scale and magma viscosity, bubbles can move relative to their host liquid and undergo collision and coalescence processes. These processes are enhanced by conduit inclination, significant wall roughness and major conduit discontinuities. The result is the generation of large gas slugs which form the basis of Strombolian eruptive activity. The behaviour of gas slugs in tubes has been the subject of extensive research, because it is critical to many chemical industries. For insight into volcanological processes, physically small laboratory experiments are scaled by exploring the range of viscous, inertial and surface-tension controlled events. However, these experiments remain unscaled in terms of expansion ratio, (P*), defined here as the ratio of the pressure at which the gas slug was created, (Ps), to the ambient pressure at the top of the magma column, (Pv), i.e., P*=Ps/Pv. In the laboratory case P* is rarely greater than 1.5, some 1 to 3 orders of magnitude less than the value pertinent to volcanic conduits. The consequence of this is that experimental gas slugs will not be subject to the same degree of dynamic super-static pressurization generated by accelerating the magma overburden as the gas slugs giving rise to Strombolian eruptions. In order to increase P* within the laboratory, a denser liquid phase, longer (higher) conduit or reduced 'ambient' pressure (Pv) can be used. We utilise the latter technique, connecting the upper end of the experimental flow tube to a vacuum system. By varying Pv, experiments covering a range of P* values scaled to conduit processes can be carried out. We explore the behaviour of expanding gas slugs, and the pressures and forces generated during rise and bursting. The volcanological implications for eruptive style and volcano-seismic interpretation are discussed.
V33D-1486 1340h
Emplacement Process of the 1991-1995 dome at Unzen Volcano, Japan
Thirteen lava lobes and big endogenous dome had grown during the 1990-1995 eruption of Unzen Volcano. Although most of lava lobes were collapsed and the surfaces were brecciated, their initial structures remain in several locations. The vent site of Lobe 11 was occupied by 4 lava lobes like large petals (petal loves) which grew in November 1993. These petal lobes were extruded from a linear vent (axis of creese structure), turned sideway and flowed to downstream. Spine intruded as the solid state into at the center of the endogenous dome, which had grown for several months in late 1994. Platy lobes were considered to be ramps (imbricated lava plates), which were formed during lateral spreading of lava as the crust of the endogenous dome in 1994. Long axes of phenocrysts in petal lobes are aligned normal to flow direction, and also to the surface of the lobes, except in the distal part and near the vent. They are roughly parallel to the elongation (intrusion direction) of the spine. They in platy lobes are weakly parallel to shear planes. Crystal preferred orientaion reflects deformation processes which were experienced during movment of the lava before complete solidification. Foliation records pure shear caused by injection force of magma and lineation records simple shear by gravity injected as supported experimentally by Buisson and Merle (2002). Foliation observed in spine may record simple shear along the conduit wall. Foliation of platy lobes may record compressive force due to lateral spreading of lobe before ramp structure formed. Water contents are 0.26 wt.% for petal lobe, 0.92 wt.% for platy lobe, and 1.26 wt.% for spine. Assuming that the water contents represent the confining pressures, spine had solidified in levels deeper than other lobes (about 420m below the surface). Therefore, crystal orientation in spine reflects stress patterns operated on moving magma in the upper part of the conduit.
V33D-1487 1340h
The relation between microlite textures and discharge rate during 1990-1995 at Unzen, Japan
The dacite lave eruption during 1990-1995 at Unzen, Japan repeated a series of development of lava domes and the subsequent collapse generating pyroclastic flows. Although, the petrological aspects of Mt. Unzen have been studied well (e.g., Sato, 1996; Nakada and Motomura, 1999), the relationship between the magma discharge rate in dome eruption and the textural facts such as CSD of microlite produced by vesiculation is still poorly understood. Natural volcanic products are supposed to be preserved the signature of the complex processes of coupling between phase changes and magma flow as vesicle and/or microlite textures. The relation between discharge rate of magma and the textural data of microlites should be investigated to understand the decompression history and magma ascent processes in the Unzen eruption as a test field. We carried out the quantitative textural analysis of volcanic products during 1990-1995 eruptions for 42 samples (classified into 13 events). As a result, we found that the trend in the temporal variation of microlite number density and size distribution can be divided into two stages, correlating with the fluctuation of discharge rate. For the 1st stage, microlite number density decrease from early 1991 to late 1992 in positive correlation with discharge rate. On the other hand, from early 1993 to middle of 1995 (2nd stage), microlite number density simply increases regardless of discharge rate. The shapes of crystal size distribution (CSD) curves firstly take power law distribution in 1991 eruption, however, for late 1st stage, CSD curve takes shape like a Poisson-like distribution. For the 2nd stage, the slope of power law type distribution becomes steep. These different trends of nucleation (number density) and growth (crystal size) processes in two stages imply that difference of effective cooling rate (dT/dt) is produced in conduit. Decreasing number density with large crystal size in late 1st stage shows that the crystal growth process is dominant than nucleation. Hence, the exsolution rate of H2O (maximum) becomes low because of lower effective cooling rate. On the other hand, high number density with small crystal size from middle to late 2nd stage indicates that the crystal nucleation dominates over crystallization processes, and the exsolution rate becomes increase. Now three different models are proposed for conduit flow. We will discuss which model proposed for conduit flow is best consistent with these data, together with the data from the conduit drilling samples of USDP4.
V33D-1488 1340h
Cl/OH Partitioning Between Hornblende and Melt and its Implications for the Origin of Oscillatory Zoning of Hornblende Phenocrysts in Unzen Dacite
We carried out high temperature-pressure experiments to determine the chlorine-hydroxyl exchange partition coefficient between hornblende and melt on the 1992 dacite of Unzen volcano. Cl in hornblende and melt are analyzed by electron microprobe, whereas OH in hornblende and melt are calculated assuming anion stoichiometry of hornblende and utilizing the dissociation reaction constant for H2O+O=2(OH) in water-saturated melt, respectively. The partition coefficient strongly depends on the Mg/(Mg+Fe) ratio of hornblende, and is expressed as ln K1 = (Cl/OH)hb/(Cl/OH)melt = 2.37-4.6*[Mg/(Mg+Fe)]hb at 2-3 kb and 800-850°C. It is shown that twofold variation of Cl content in the oscillatory zoned core of hornblende phenocrysts in the 1991-1995 dacite of Unzen volcano is not explained by the dependence of Cl/OH partition coefficient on the Mg/(Mg+Fe)hb, and requires ca. 80% variation of Cl/OH ratio of the coexisting melt. Available experimental data at 200 MPa on Cl/OH fractionation between fluid/melt suggest that ca. 1.2-1.8 wt% degassing of water from the magma explains the required 80% variation of the Cl/OH ratio of the melt. The negative correlation between Al content and Mg/(Mg+Fe) ratio in the oscillatory zoned core of hornblende phenocrysts is consistent with repeated influx and convective degassing of the fluid phase in the magma chamber.
V33D-1489 1340h
Temporal variations in vesicularity of erupted products at Unzen Volcano, Kyushu, Japan
The Unzen Scientific Drilling Project (USDP) commenced in April 1999 with the aim to drill to the volcano's conduit, still hot from the 1990-'95 eruption, in addition to revealing much about Unzen's eruptive history. Recent activity of the volcano has consisted largely of emplacement of viscous domes, which produce block and ash flows. This almost wholly effusive activity is in contrast to the evidently highly volatile nature of the stored magma. An important question that may provide a clue to this paradox is how eruptive behavior has evolved with time. By examining samples from the cores from boreholes drilled into the flank of the volcano, we were able to characterize more than 500 ka of activity. Vesicularity of magma reflects its volatile content and decompression history, and plays a role in producing effusive or explosive behavior. To quantify variations in vesicularity as a function of depth and hence time, we measured bulk density, connected void space, and void-space-to-area ratios for samples from USDP-1 and USDP-2. We found a negative correlation between density and depth, reflecting increasing vesicularity, despite consistent bulk composition. We also found decreasing crystallinity with increasing depth.
V33D-1490 1340h
Magma History of Unzen Volcano, Investigated using Drilling Slime of Conduit Drilling
To clarify the geologic and tectonic growth history, magma evolution and eruption mechanism at Unzen volcano, the scientific drilling project (USDP) started in 1999. Conduit drilling (USDP-4) aiming at penetration into a hot magmatic path was carried out during May-July of this year. Drilling slimes from USDP-4 were investigated for understanding geologic and magmatic historyy during the growth of Unzen volcano. Drilling slime was collected in every 2m depth interval down to 1994m. The sample was washed and sieved into three grain size groups. Based on the shape and color of particles, the original rock was estimated as lava flow, and pyroclastic- and debris-flow. The slime is porphyritic dacite to andesite, and contains phenocrysts of plagioclase, hornblende, biotite, orthopyroxene, clinopyroxene, quartz, and opaque minerals. Plagioclase always includes dusty and clear types. Color, chemistry and magnetic susceptibility were measured. The color of whole sample was expressed in the CIE L*a*b* color space. As a result, color of sample change systematically with depth. L* value express degree of porosity, and a* value and b* value express degree of oxidization or alteration. Chemical composition, for example, SiO2 and K2O/Rb are nearly constant in samples deeper than the 500m depth and variable in the shallower samples. Magnetic susceptibility of the slime is also variable with depth. The value of magnetic susceptibility depends on the quantity of Fe-Ti oxide in the sample and is concordant with total iron content and TiO2 content of the sample. As the rock texture does not largely changed and the magnetic susceptibility is almost identical between different grain size groups, the magnetic susceptibility probably represents the difference of chemical composition and state of oxidation after deposition. Since the Unzen volcanic rocks contain commonly lines of evidence showing magma mixing, it is considered that similar mixing process was operated in the early stage of volcanic history. On the other hand, the mixing process, including chemistries of end members and mixing ratios, was changeable in the late stage.
V33D-1491 1340h
Petrology of Conduit Lava at Unzen Volcano; Result of Unzen Scientific Drilling Project
Lava samples were recovered from the conduit of the 1991-1995 eruption during the scientific drilling project in summer of 2004. Lava from the 1975-1997 m-depth conduit is most fresh among conduit samples and was nested in the hottest zone. The chemical composition including Sr isotopic ratio is very close to the 1991-1995 dome lava. The conduit lava is porphyritic dacite with phenocrysts of plagicolase, bioite and hornblende, though it is altered hydrothermally, such that hornblende was replaced by chlorite, carbonate and rutile, contrasting to freshness of biotite. The groundmass is composed of devitrified glass even in the chilled margins and contains small crystals of pyrite. As 3 wt % of water can be involved in the melt at the depth of 1.3 km where the conduit sample was recovered (about 40 MPa), about half of the initial water (6 wt %) should have been lost until when the magma reached this level. However, bubbles and microlites whose crystallization is induced by degassing are very scarce, compared in the groundmass of the dome lavas. Scarce amounts of microlites and bubbles indicate that degassing-induced crystallization was least in the ascending magma in this level, probably due to slow ascending (low cooling rate), and that bubbles could be escaped smoothly from the fluidal melt. As permeability of the country rocks is very low based on the logging data of the drilling, it is likely that magmatic volatiles should have escaped upward through cracks formed by ascending magma itself. Bubbles in the magma might have migrated into the tip of dyke magma where vapor-dominated tip was pressurized to form cracks. Crystallization of most of microlites commonly found in the dome lava had occurred in the shallower depth due to faster degassing (effective cooling rate).
V33D-1492 1340h
Dispersion of magmatic volatiles through groundwater system at Unzen volcano, SW Japan: a summary
Geochemical studies on groundwater flow system is performed using chemical and isotopic technique at Unzen volcano placed on the Beppu-Shimabara graben in Kyushu Island, SW Japan. More than 400 samples of surface and subsurface waters are collected and analyzed for major chemistry and isotopic compositions, i. e., tritium, D/H, $^{18}$O/$^{16}$O, $^{13}$C/$^{12}$C, $^{3}$He/$^{4}$He etc. The tritium concentration is used for estimate of groundwater retention time, and stable isotopic ratios of water are used for determining recharge area. The carbon and helium isotopic ratios as well as their concentrations are measured in order to know the contribution and flux of magmatic volatiles. The main groundwater flow system around Unzen volcano is developed from the summit area to the east Shimabara city area with a large flow rate of 1.7 m$^{3}$/sec. This flow rate almost corresponds to the recharge rate from an area of newer volcanic body. The hydrogen isotopic ratio of the surface and subsurface waters collected at Shimabara city are in a narrow range from -48 to -45 $\permil$, indicating that groundwaters recharged at the slope are well-mixed in the course of flow. On the other hand, their HCO$_{3}$ concentrations significantly varied suggesting that input of magmatic carbon-bearing species into the groundwater has occurred close to the collection sites. Helium isotopic ratio of the waters also varied which is consistent with the variation of bicarbonate concentration. These magmatic volatiles unlikely come from the summit conduit through aquifers, but ascend vertically through faults. We summarize and map the flux of magmatic volatiles dissolved in groundwater and evaluate two styles of magma degassing 1) degassing from deep reservoirs through fractures or faults and 2) degassing from magma conduits through aquifers, and discuss efficiency of escape of magmatic volatiles through conduit wall during magma ascent.
V33D-1493 1340h
Underground structure around Unzen volcano, Shimabara Peninsula, Kyushu, Japan, inferred from gravity analysis
The Unzen volcanic scientific Drilling Project (USDP) has been performed flank- and conduit drillings around the Unzen volcano in the Shimabara Peninsula, Kyushu, Japan. The results of the flank-drillings revealed the subsurface structure, geological history and physical property around the Unzen volcano in the eastern part of the Shimabara Peninsula. On the contrary, the New Energy Development Organization (NEDO) has conducted various surveys on the geothermal field research project in the western part of the Shimabara Peninsula. Based on the previous study and flank-drillings, the stratigraphy of the Unzen volcano is described as follows: Quaternary Unzen volcanic products, Quaternary-Neogene rocks and Paleogene rocks (in descending order). The density of the Quaternary-Neogene units varies with the facies. The density of the volcanic rocks is higher than that of the pyroclastic or sedimentary rocks. It is difficult to estimate the basement of the Unzen volcanic products from the gravity analysis. The distribution of the basement configure of the Unzen volcanic products is obtained by compiling the borehole data in the Shimabara Peninsula. We attempted to estimate the surface of the Paleogene rocks on the basis of the gravity analysis. The gravity data around Unzen volcano were adopted from the gravity CD-ROM published by the Geological Survey of Japan. We constructed the density model which consisted of the basement (the Paleogene rocks), overlaying layer (the Unzen volcano products and the Neogene - Quaternary deposits) and the high density volcanic rocks revealed by previous geological and geophysical works. The underground structure model of the surface of the Paleogene rocks and the base of the Unzen volcano products was compiled. In the poster, we will present the tectonics around in the Shimabara Peninsula based on constructed underground model and compare the underground structure with the subsurface fault activity estimated by geomorphological analysis.
V33D-1494 1340h
Late Quaternary Activity of active faults in the Unzen Graben, western Kyushu, Japan, based on tectonic geomorphology
The Unzen Graben, which is bounded by normal faults to the north and south, is located on the western end of the Beppu-Shimabara Graben, west Kyushu, Japan. The faults in the Unzen Graben have developed with the growth of the Unzen volcanoes and dislocated volcanic materials such as lavas and pyroclastic deposits. The detailed location and vertical offsets of the active faults have been reported by previous studies, but the timing of faulting was poorly constrained. We estimated the vertical offsets based on topographic profiling, and calculated vertical slip rate based on recently obtained ages of volcanic rocks and sediments. In addition, we excavated two trenches at Chijiwa Town (western part of Unzen Graben) in order to reveal the timing of recent faulting events, and also found several outcrops across the active faults. Based on these surveys, the active faults in the Unzen Graben can be divided into four groups. Two groups of faults, which are located on the northern and southeastern margins of the Unzen Graben, have vertical slip rates as high as 0.13 to 1.0 mm/yr. A group of faults inside the Unzen Graben has slip rate 0.085 to 0.46 mm/yr except for the Akamatsudani fault which has slip rate 0.91 to 1.7 mm/yr. Three groups of faults mentioned above show cumulative vertical offsets and their rates of activity tends to be higher around the edifice of the Fugen volcano. On the other hand, the group outside the Unzen Graben is 0.026 to 0.068 mm/yr.
V33D-1495 1340h
Geochemical study of groundwater in and around the Iwate volcano, Japan: supply of the magmatic components
Thirty-three water samples (spring: 1, groundwater from shallow wells: 25, thermal water from deep wells: 7) were collected at various depth in and around the Iwate volcano located in the NE Japan. Water samples were measured in chemical (major components) and isotopic compositions (hydrogen, oxygen, carbon and helium isotope ratios). Based on chemical and isotopic features of groundwater and thermal water, we discuss the groundwater flow system and supply process of magmatic components. The western part of Iwate is developing an erosional valley and has several volcanic peaks and an active fumarolic area. The eastern part of Iwate forms a comparatively new volcanic body with less-active fumarolic area at a summit. An active fault is located in the southwest of the volcano. In the east flank, a biggest spring called Oide spring is located with a flow rate of 50000 t/d. The chemical and isotopic compositions of Oide spring water show very interesting compared with those of the groundwater around Oide spring. The features of the Oide spring water are, (1) low $\delta$D and $\delta$$^{18}$O ratio, (2) high HCO$_{3}$$^{-}$ content, and (3) high $^{3}$He/$^{4}$He ratio (3.9Ra). The low $\delta$D and $\delta$$^{18}$O ratio indicate that Oide spring water is derived from high elevation whose recharged area fully covered the summit area (Kazahaya et al., 2000). The high HCO$_{3}$$^{-}$ content and high $^{3}$He/$^{4}$He ratio are suggested to result from the magmatic components related to the volcano. For a thermal water of 810m depth located near Oide spring, the features are as follows, (1) much lower $\delta$D and $\delta$$^{18}$O than the Oide spring, (2) Na-Cl, HCO$_{3}$ type water, and (3) low $^{3}$He/$^{4}$He ratio (0.1Ra). The thermal water in the east flank is suggested to have an older age and accumulated radiogenic $^{4}$He. In the southwest flank, some thermal waters are collected from drill holes of about 1000m depth located along a fault. The chemical and isotopic features of these thermal waters are, (1) shifted $\delta$D and $\delta$$^{18}$O, (2) high Cl, HCO$_{3}$, SO$_{4}$ content, and (3) high $^{3}$He/$^{4}$He ratio (2.5-3.3Ra). As contrasted to the water in east flank, the magmatic components are enriched in the thermal waters in the southwest flank, which may be supplied via an active fault. The groundwater system of the Iwate volcano is summarized as follows. (1) East flank: There is a biggest shallow aquifer in which the water recharged from summit area of volcano. The largest amount of magmatic component is supplied to the shallow water system, whereas the thermal waters in this area which form deep aquifer system contained less magmatic components. (2) Southwest flank: The thermal waters containing high concentrations of magmatic component are found along the active fault. The magmatic components are likely supplied via fault to the thermal waters. Based on the above results, further results on the residence time of water and the fluxes of magmatic component into the aquifers will be presented.