V53B-1543
A Study of the 2450 BP Pululagua Plinian Eruption (Ecuador): Implications for Models of Tephra Dispersal
The 2450 BP Plinian eruption of Pululagua is known to have occurred in no-wind condition, resulting in circular-shape isopach and isopleth maps. Previous studies showed that the Plinian fallout is composed of 3 main units: the basal grey ash (BGA), the basal Plinian fallout (BF) and the white ash (WA). This study focuses on the BF unit, (consisting of three main layers: BF-I, BF-II and BF-III) and the co-Plinian unit (WA). The circularity displayed by Pululagua fallout was verified before sampling 23 locations along a south-east axis. Granulometry and componentry analysis have been performed on the collected samples throughout the usual sieving procedure. The fraction < 1.0$\Phi$ of each sample was also analysed using the Malvern PharmaVision 830 optical device. Results include: (1) grain-size distribution; (2) morphological parameters of each single particle, such as: (i) mean diameter; (ii) area; (iii) volume; (iv) roundness; (v) convexity and (vi) mean intensity; and (3) single image for each individual particle. Shape parameters such as roundness, convexity and mean intensity can be used as fingerprints of the sample material and could be a powerful tool for a fast componentry analysis of the volcanic ash. The absence of wind provides an ideal opportunity to calibrate models for tephra dispersal. Data from unit BD-II were therefore used to calibrate the model TEPHRA. An inversion technique also helped us determine crucial eruptive features, such as column height and volume. In addition, different techniques typically used to determine the distribution of maximum clasts were tested in the field. Resulting discrepancies significantly affect the derivation of important eruptive features (e.g. column height and wind speed).
V53B-1544
Complex Eruption Dynamics in a Hawaiian Explosive Eruption: Kilauea Iki 1959
The 1959 eruption of Kilauea Iki volcano is an outstanding example of Hawaiian volcanism in which high fountaining of ejecta reached elevations as great as 580m. A total of 17 fire fountaining episodes over 37 days built a steep sided cone and deposited a tephra blanket extending over 15 km down wind from the cone. The unusual geometry of the 1959 vent, located on the side of a steep-walled pre-existing collapse crater, lead to ponding and the formation of a lava lake. The eruption shows a complex interplay between lava lake and cone. Lava ponded in the lake strongly influencing the close of explosive episodes and cone failures fed back to change the position of the vent and the geometry of the lake. The durations and intensities of each of the 17 phases of explosive activity vary markedly over two orders of magnitude. We interpret this diversity in the context of contrasting inputs from (1) the melt recycled from the newly formed lava lake and (2) freshly arrived and volatile-rich melt ascending through the shallow summit reservoir. The resulting ejecta includes two types of clasts reflecting these two melt types: (1) achneliths, which are black fluidal scoria that contains a wide range of bubble sizes and (2) ragged, highly micro-vesicular golden pumice/reticulite.
V53B-1545
Comparing Vesicularity in Pyroclastic Deposits
The investigation of processes associated with explosive volcanism requires the knowledge of the physico-chemical properties of the rock material involved. Parameters such as the rock's density and vesicularity highly influence the rheological properties as well as the fragmentation behavior of magma. Because observations inside the volcanic conduit are not possible, information on the spatial and temporal variability of the ascending magma's vesicularity is more easily obtained by measuring a statistically reliable amount of representative erupted material. In order to correlate density/porosity distributions of eruptive products with specific volcanic settings and eruption characteristics, data from six field campaigns in the circum-Pacific area were analyzed and compared. Densities from ~3000 pyroclastic- deposit samples collected on St. Augustine (Alaska), Bezymianny (Kamchatka), Unzen (Japan), Colima (Mexico), and Merapi, Krakatau and Kelut (Indonesia) were mostly measured directly in the field. This field-based density measurement method is based on the Archimedean principle whereby a sample's mass is measured in air and under water (Kueppers et al., 2005). Previous results indicated that the density peaks of St. Augustine and Bezymianny were at 1.5 and 1.6 g/cm3, respectively, whereas the most abundant densities at Mt. Unzen and Merapi were significantly higher at 2.2 and 2.3 g/cm3. However, samples collected from the 1883 Krakatau pyroclastic deposits during our most recent campaign follow a lower density distribution with a density peak of ~0.6 g/cm3. This suggests that the erupted material had a higher gas content. Results will be supported by detailed laboratory-based density and porosity measurements, and investigations of their influence on rock fragmentation and viscosity. This comprehensive work is part of the BMBF project SUNDAARC, which aims to quantify the potential risk of selected highly-explosive volcanoes by combining field and laboratory investigations.
V53B-1546
A Textural Exploration of the Physical Attributes of Pumice Clasts From the 23,000 Y.B.P. Eruption of Popocatepetl Volcano in Central Mexico
The eruption of Popocatepetl Volcano 23,000 Y.B.P. produced the youngest in a series of debris avalanches. This 9 km3 debris avalanche was followed by a large blast deposit and then the White Pumice fall, on which we focus. We are investigating whether the large decompression that resulted from the debris avalanche may have caused the vesiculation process to change, thereby altering the physical properties of the deposit. The White Pumice is the thickest fall deposit at Popocatepetl and has an estimated minimum volume of 1-2 km3 DRE. Its relative reverse grading in lithic sizes suggests mass flux increased throughout the eruption, assuming no significant change in winds during deposition. While the mass flux appears to increase, the average vesicularity of pumice clasts decreases through the deposit. Bulk vesicularity of the pumices ranges from 55.3% to 88.3%, with a mean value of 75.4% and a mode of 77%, based on density measurements of 900 pumice clasts. We are also characterizing the textural attributes of pumice from the deposit, which we are qualitatively analyzing through ESEM pictures of the surface of low, medium, and high vesicularity pumice clasts. Preliminary results suggest that individual pumice clasts throughout the vesicularity range share the same textural qualities in different proportions. A variable degree of vesiculation on the surface of individual pumices occurs in sections that transition into one another. Some areas are smooth with little to no vesiculation and an abundance of cracks. In low vesicularity sections of individual pumice, there are predominately rounded to slightly elongate vesicles with irregular bubble shapes. These areas often have a rough surface texture due to pieces of glass welded to the smooth, glassy bubble walls. In more vesicular regions of the same pumice, the rough texture due to welding is less common and the vesicles become more elongate with smoother bubble walls and a more regular shape. Areas with tube vesicles are the most vesicular, and have strings of smooth glass representing the diminishing bubble walls. Pumice with low bulk vesicularity consist of mostly low to moderately vesicular sections with few highly vesicular areas and a larger proportion of smooth areas. Highly vesicular pumice consist of mainly of highly vesicular and tubular vesicle sections, and a low percentage of low to moderately vesicular sections and smooth areas. The medium vesicularity pumice tend to incorporate characteristics of both extremes. Our results preliminarily show that the vesicularity is similar to that of other eruptions that were not preceded by a large debris avalanche, suggesting that the sudden decompression may not affect the process of vesiculation. Future use of X-Ray Computed Tomography will allow us to clarify bubble interactions both internally and on the surface of the clasts, thereby characterizing the differing spatial vesicularity.
V53B-1547
Textural Characterization Of Scoria From Low Energy Strombolian Events At Villarrica Volcano: Preliminary Data From November 2004
During November 2004 a multidisciplinary experiment was deployed for a 10-days period at Villarrica Volcano (Chile). Activity since the 1984-85 eruption has consisted of continuous, passive degassing from a summit lava lake/vent and subordinate strombolian activity. During the experiment, a small lava lake was active within a 70-m-deep summit pit. Within the lake regular bubble bursting ejected fragments that rose and fell within the crater itself, together with splashing of lava onto the inner pit walls. Rarer, more energetic events emplaced scoria around the crater rim extending just a few 10's of meters from the pit rim. The limited dispersion of the products, confirm the low energy character of even these events. This report presents preliminary analyses of the textural features of the juvenile material ejected by the low energy strombolian events that occurred between November 8 and 17. We collected two sample suites of 100-to-300 clasts lying on fresh snow at the eastern and western sides of the crater area and within 50 m of the pit edge. Products included decimeter-sized bombs, scoria lapilli, centimeter-sized reticulites and coarse ash comprising fresh glassy shards. Bombs were rounded as well as elongated, and included spindle bombs and flattened spatter forms. Apart from the golden reticulites, all products were black and iridescent, with metallic surface lusters. We have completed density measurements on pyroclasts ranging from 2 to 7 cm in diameter. The juvenile population of the first (8 November) sample was dominated by scoria (65%) with vesicularities of 50-80%, with the remaining 35% comprising pumice/reticulite with vesicularities of 79-94%. The second (17 November) sample had a lower percentage of reticulite (less than 15% of the total sample) with vesicularities of 89-96%. The scoria in this sample had vesicularities of 41-81%. In general reticulite is characterized by rare large vesicles (up to 5 mm diameter) in a closely packed network of mostly spherical medium-to-small bubbles, with scarce phenocrysts (olivine and plagioclase) and clear microlite-free glass. In contrast the scoria fragments contained abundant phenocrysts and microphenocrysts (olivine and plagioclase) which increase in abundance with increasing density. Low and medium density scoria were typically characterized by a single large (cm-sized) rounded vesicle at their core, surrounded by abundant small vesicles. The high-density scoriae show a decrease in the abundance of small vesicles.
V53B-1548
Estimate Of Deposition Temperature Of The PDC Deposits Of The AD 472 Pollena Eruption Of Somma-Vesuvius, Italy.
The AD 472 Pollena eruption of Somma Vesuvius, Italy, was a complex subplinian event which generated several pyroclastic density currents (PDCs). PDC deposits were emplaced by dilute-to-highly concentrated stratified currents both during the magmatic and phreatomagmatic phases of the eruption. PDCs are the most devastating of all volcanic phenomena, due to their high velocity, temperature, ash concentration and missels content. A key piece of information in understanding PDC dynamics and their hazard is given by the temperatures at which they are emplaced (Tdep). Because the Pollena eruption is considered to have a size and style similar to that of the Maximum Expected Event defined for Somma-Vesuvius, knowledge of the PDC deposits and their temperatures is very important in terms of hazard evaluation and physical modelling. In this research we completed a thermal analysis of the magnetic remanence carried by lithic clasts embedded within the deposits to infer the Tdep of the PDC deposits allowing us to develop a detailed picture of the depositional history of the entire pyroclastic sequence of Pollena eruption. Sampling of the lithic clasts embedded within the deposits was done both vertically through the stratigraphic sequence, as well as within single flow units along transects extending radially from the vent. Clasts, ranging in volume from 2 to 100 cm3, were collected from twelve sites. The clasts were subjected to stepwise thermal demagnetization at temperature steps of 20-40 °C. This procedure allowed to separate two remanence components in each individual clast: the primary, high blocking-temperature component and the secondary, low blocking temperature component acquired because of reheating during the PDC emplacement. The Tdep at a particular site was then estimated from the overlapping of the reheating temperature range for all invididual lithic clasts collected at that site. Preliminary results show that the Tdep values vary between 260 and 320 °C in most PDC units and reach in one unit the highest value of 340 °C.
V53B-1549
The 340-240 ka Ignimbrite Flare-Up Event in Taupo Volcanic Zone, New Zealand
The onset of continental rifting often sees outbursts of intensive silicic volcanism - ignimbrite flare-Up events - with complex relationships between extensional, magmatic and volcanic processes. The central Taupo Volcanic Zone (TVZ) in New Zealand is a young vigorously active silicic focus in a rifting arc setting. Much of the present-day character of central TVZ was established in a remarkable flare-Up event from ~340 to ~240 ka. At this time ≥ 7 ignimbrite-forming and numerous smaller eruptions occurred, totaling at least 1700 and probably 3000 km3 of magma, and forming calderas that pepper a 90 × 40 km area. The first and largest event(s) generated the Whakamaru-group ignimbrites and their coeruptive fall deposits, and these were followed by a modest-sized ignimbrite (Chimp) prior to 280 ka. Two ignimbrites were emplaced west (Pokai) and east (Matahina) of TVZ at ~280 ka, and three ignimbrites were erupted around ~240 ka: of these one (Kaingaroa) is distributed east and two (Mamaku and Ohakuri) west of TVZ. These last two ignimbrites closed the flare-Up event. However, the parental eruptions occurred from vents centered 35 km apart, but formed a complex overlapping eruption sequence. Early wet and dry pumice fall deposits from a southerly ('Ohakuri'?) source underlie, and are coeval with, flow units of the Mamaku ignimbrite from a northerly source now expressed as Rotorua caldera. A fine-grained coignimbrite ash bed was then deposited on both fall and flow deposits. This ash bed was then only minimally disturbed by erosion before eruption of the Ohakuri ignimbrite from the southerly source. Such close temporal linkages between major eruptions imply that faulting readjustment was critical in triggering neighboring events. Future work is aimed at reconstructing the magma bodies for the flare-Up ignimbrites to see to what extent their crystallization histories overlapped in time, and what degree of commonality in chemical and isotopic characteristics there was between roughly to closely coeval but spatially separated large magma bodies.
V53B-1550
Stratigraphy of Late Pleistocene-Holocene pyroclastic deposits of Tacana Volcano, Mexico-Guatemala
Tacana volcano (4,060 masl), the highest peak of the Tacana Volcanic Complex, is an acitve volcano located on the Mexico-Guatemala border. Tacana resumed phreatic activity in 1950 and again in 1986. After this last event, the volcano became the locus of attention of authorities and local scientists began to study the complex. Tacana's stratigraphic record has been studied using radiocarbon dating and these indicate that the volcano has been very active in the past producing at least 12 explosive eruptions during the last 40 ka years as follow: a) Four partial dome destruction events with the generation of block-and-ash flow deposits at 40, 28, <26, and 16 ka. b) Four small-volume phreatomagmatic events that emplaced dilute density currents at 10.6, 7.5, 6, and 2.5 ka. c) Four eruptions that emplaced pumice-rich fall deposits, three of them widely dispersed towards the NE flank of the volcano in Guatemala and dated at ~32, <24 and <14 ka, and finally a 0.8 ka fall deposit restricted to the crater vicinity that might represent the youngest magmatic eruption of the volcano. Although refining of these stratigraphic sequence is still underway, the eruptive chronology of Tacana volcano cleary indicates that explosive eruptions producing plinian fall and pyroclastic density currents have taken place every 1 to 8 ka years. This record constrasts with the small phreatic eruptions that occur 1-2 per century. So, this indicates that Tacana volcano is more active than previously considered and these results must be considered for future researches on hazards maps and mitigation.
V53B-1551
Holocene Eruptions of Machin Volcano: Stratigraphy and Eruptive Dynamics
Cerro Machin (2,750 m a.s.l.) is a Quaternary dacitic volcano located in the Central Cordillera of the Andes in Colombia. It is a 2.4-km wide tuff ring that hosts three domes with fumarolic activity. The volcano's origin is linked to the development of a pull apart basin bounded by the Machin and Cajamarca faults. Detailed stratigraphy and radiocarbon dating indicate that six eruptions have taken place during the Holocene. Each eruptive event established a Plinian column that deposited pumice fall deposits followed by the collapse of the column ensuing pumiceous pyroclastic flows, surges and secondary lahars. These eruptive events are dubbed and dated (average age) as the Espartillal (5000 yr.), P0 (4600 yr.), P1 (3600 yr.), Guaico (2600 yr.), P2 (1200 yr.), and Anillo (900 yr.). The pyroclastic fall deposits have dispersal axes to the NW, they are exposed up to 60 km from the crater and cover a total approximated area of 2000 km2 and have a total volume of 4.9 km3. Calculated column heights oscillate between 19 and 32 km. The collapse of these columns generated pyroclastic flow deposits that traveled up to 15 km around the volcano infilling the Coello valley. Some pyroclastic flows were able to surmmount 200-m high topographic barriers. Remobilization by water of this unconsolidated material generated lahars that followed the Coello and Magdalena Rivers and traveled up to 115 km from the volcano covering and area of 1000 km2. The deposits of all these eruptions have affected large areas where the modern important cities of Colombia are now located, as well as routes of great economic importance. Although the volcano is in a quiescent state it must be closely monitored since it has produced cataclysmic eruptions ca. every 1000 yrs.
V53B-1552
Volcanism at Hualca Hualca Volcano, Southern Peru
Nevado Hualca Hualca (6025m), in southern Peru, is the northernmost edifice in a north-south trending chain of 3 volcanoes that includes Ampato and the active Sabancaya stratovolcano. The oldest in the chain and considered extinct, virtually no research exists about the history of this large volcano. The summit of the volcano shows deep incision by glaciation, which from aerial photographs appears unaffected by later volcanism. Its north slope, however, possesses numerous volcanic domes, extensive lava flows with distinct levees and transverse ridges, and pyroclastic flow deposits. Deposits on the northwestern slope of Hualca Hualca include breadcrust-rich block-and-ash flows (BAF), several dacite lava flows including one with an identifiable source dome about 15km from the summit, and a sequence of small pyroclastic flow deposits with minor associated tephra. Analyses of these deposits show a restricted range of compositions (63-68 wt% SiO2). The PF sequence has an upward decrease in SiO2 and basaltic andesite (56 wt% SiO2) inclusions occur in the uppermost PFs. Principal phenocrysts include plagioclase, biotite, hornblende, clinopyroxene, orthopyroxene, Fe-Ti oxides, and sphene. Fine grained, angular to sub-rounded magmatic enclaves occur within the breadcrust-rich BAF deposits and the youngest lava flow. They are characterized by randomly oriented acicular hornblende, lack of chilled margins, and a few voids indicative of a quench texture. Plagioclase crystals with "dusty" rims or cores present in most of the deposits suggest resorption caused by magma recharge. These features imply a stratified magma chamber subject to magma recharge events and mingling to produce the quench texture enclaves. Chemical analyses indicate that the volcanic products result from magma mixing processes; the basaltic andesite inclusions may represent the mafic end-member of the mixing process. The physical characteristics of the deposits and chemical analyses were compared with data from the 1990-98 eruptive episode of Sabancaya volcano. Quench-texture enclaves and dusty-rimmed plagioclase exist in practically all of the Sabancaya deposits. The Sabancaya chemical analyses plot in line with those from the Hualca Hualca deposits; the Hualca Hualca samples are more evolved in almost every case except for the basaltic-andesite inclusions. This indicates a common formational history for the products of these two volcanoes and suggests a longer crustal storage time for the more evolved Hualca Hualca volcanics.
V53B-1553
Spatiotemporal evolution of pyroclastic density currents and eruption dynamics during the 6.5 ka caldera-forming eruption, Kikai volcano, Japan
The 6.5 ka caldera-forming eruption from the Kikai volcano devastatingly damaged to prehistoric human activities of southern Kyushu, Japan. Stratigraphy, component, and lithology of the pyroclastic deposits were analyzed, with respect to a precise temporal framework to determine how large pyroclastic density currents can evolve in time and space and how large explosive eruptions evolve during caldera collapse. Stratigraphical sections are characterized by plinian pumice fall deposits (Unit A), intraplinian flow deposits (Unit B), voluminous ignimbrite (Unit C), and co-ignimbrite ashfall deposits (Unit D). A plinian stage is subdivided into a first small phase and a second large one. A column height in the second phase was estimated at 40-43 km, and collapse of the column produced Unit B, which consists of multiple thin subunits with stratified or cross-stratified facies in various degree of welding. Each thin subunit is a few to a few tens of centimeters in thickness, and is composed of a basal lithic-rich layer and an upper welded pumice-rich layer. It is just one flow unit of a small-scale pyroclastic density current. Lithic-rich layers are composed of mostly altered lithic, bearing obsidian clasts with chilled cracks and submarine boulders, and are fine-depleted. These evidences indicate that, during the plinian column collapse, high temperature dilute currents were generated repeatedly from phreatomagmatic explosions and segregations of dense pyroclasts in a turbulent condition resulted in producing lithic-rich layers. Unit C is subdivided into three units (C1-C3). Unit C1 shows non-welded stratified facies, which consists of lithic and crystals, minorly including quenched juvenile materials. Unit C2 shows welded stratified facies, which consists of lithic-rich layers and pumice-rich layers. These two subunits only occur in topographic depressions and can gradually change into Unit C3, which is thickest and poor-sorted with non-welded massive facies and sometimes includes fragments of welded tuff of Unit B. These facts indicate that a climactic voluminous current was deposited after a welding of Unit B at some locations, and it has had a turbulent basal part producing Unit C1 and C2 and a main sustained body producing Unit C3. Erupted magma has gradually changed from rhyolite into andesite-bearing one and pyroclasts in the current should have aggraded progressively because andesitic juvenile clasts are only included in the Unit C3. Main caldera subsidence may have started before a Unit C3 deposition, from a finding in the form of a fault overlaid by Unit C3 on the caldera rim.
V53B-1554
The explosive activity of the Colima volcano in 2005
The Colima volcano, Mexico, showed a new cycle of explosive activity beginning in March and April the 2005. This increased gradually and in May it appeared an explosive event which generated piroclastic flows on all flanks of the volcano. On May 23 a new dome was created, and it was observed from the Volcanological Observatory of the Nevado de Colima. Hours later its dome was destroyed by a strong explosion, which formed a column 3 km hight and piroclastic flows that reached a distance of 5 km on the ravines of the South sector. On May 30 the most intense explosion from 1999 occurred when the plume reached heights over 3.500 m above the crater, and piroclastic flows. In the month of June it generated four explosive events of characteristics similar to those of May. These constant explosions caused constant morphological changes on the top, being the most significant the collapse of the North and South walls of the crater in the first week of June, and the creation of a new crater in July. This activity was similar to the one shown in 1902-1903 and reported by Severo Diaz (1906), but without reaching the maximum levels of activity reported for 1903, where it had levels of three to five maximum explosive events per day. The explosions deposited great amount of nonconsolidated materials, like ash, lithics and rocks on the flanks of the volcano, which with the present rainy season have generated lahares, two in the month of June, ten in July and eight in August (RESCO reports). These have flowed in small flows on the ravines of La Lumbre, Montegrande, San Antonio and La Arena. None of them have caused damages until August, 2005.
V53B-1555
Eruption sequence of the Suwanoharu tuff ring in the Pliocene Oyano formation, western Kyushu, Japan
Volcanism in the Ariake sea region, west Kyushu, Japan, where is the northern extension of the Okinawa trough, have been active during the last 5 million years, from Pliocene to present Unzen volcano. The region has been subsided with development of the tectonic graben. Pliocene Oyano formation is predominated in shallow marine pyroclastic deposits including some base surge deposits. A series of stratified pyroclastic deposits, which were originated by phreatomagmatic eruption cycles is exposed on a sea cliff and is considered to be a remnant of tuff ring, which is called Suwanoharu tuff ring. Suwanoharu tuff ring is expected to preserve the complete eruption sequence. The sequence is basically alternation of massive white tuff layers and laminated gray coarse to lapilli tuff layers. We can identify seven eruption cycles bounded by the massive white air fall pyroclastic deposits (unit 1- 7). The gray pyroclastic deposits consist of massive dark gray medium tuff, crudely laminated medium to coarse-grained tuff, inversely graded lapilli tuff, cross-bedded lapilli tuff, and lapillistone and are considered to be a base surge deposit. Ascending to the eruption cycles, the thickness of massive white tuff layers is decreasing, but the gray layers are increasing. Representative samples were collected from 42 horizons of the tuff ring sequence and analyzed the bulk rock chemical compositions using XRF and the mineral assemblages using XRD. The bulk rock chemical compositions of the gray layers changes from andesite at the lower horizon to basalt at the upper horizon. On the other hand, massive white tuff layers have almost identical and are andesite. Because the compositions of basal layers are very similar to the silt of the basement rock, the beds in the unit 1 may have been deposited fragmented basement rocks by the first phreatomagmatic eruption. At the upper layers of the first eruption cycle (unit 1), chlorine contents are up to the maximum level (ca. 4%) and the level is decreasing exponentially according to the later eruption cycles. Mafic minerals, such as hornblende and hypersthene, are scarcely observed in massive white tuff layers, but relative abundance of kaolin is increasing with progressing the eruption cycle from unit 1 to unit 4. As rapid increasing of the chlorine contents observed at the basal part of the sequence, interaction between magma and seawater generated the early phreatomagmatic eruption. Because the basement rocks were indurated silt, the sequence lacks basal breccia, which is commonly observed in a tuff ring sequence. Increasing of the kaolin abundance implies the expansion of the vent area at the boundary between magma chamber and wall rocks. The widened vent may promote the relative abundance of the essential material in the pyroclastic deposits due to the direct connection of the magma chamber and surface. As a result, the eruption style changed from phreatomagmatic to strombolian type at the final eruption cycle.
V53B-1556
Volcan Reventador's Unusual Umbrella
In the past two decades, field observations of the deposits of volcanoes have been supplemented by systemmatic, and sometimes, opportunistic photographic documentation. Two photographs of the umbrella of the December 3, 2002 eruption of Volcan Reventador, Ecuador, reveal a prominently scalloped umbrella that is unlike any umbrella previously documented on a volcanic column. The material in the umbrella was being swept off a descending pyroclastic flow, and was, therefore, a co-ignimbrite cloud. We propose that the scallops are the result of a turbulent Rayleigh-Taylor (RT) instability with no precedents in volcanology. We ascribe the rare loss of buoyancy that drives this instability to the fact that the Reventador column fed on a cool co-ignimbrite cloud. On the basis of the observed wavelength of the scallops, we estimate a value for the eddy viscosity of the umbrella of $4000 ~m^2/s$. This value is consistent with a previously obtained lower bound ($200 ~m^2/s$, K. Wohletz, priv. comm., 2005). We do not know the fate of the material in the umbrella subsequent to the photos. The analysis suggests that the umbrella was negatively buoyant. Field work on the co-ignimbrite deposits might reveal whether or not the material reimpacted, and if so, where and whether or not this material was involved in the hazardous flows that affected the main oil pipeline across Ecuador.
V53B-1557
Morne aux Diables. a potentially active volcano in northern Dominica, Lesser Antilles
The island of Dominica, which is located near the center of the Lesser Antilles island arc, comprises at least 8 potentially active volcanoes. One of these is Morne aux Diables, an isolated composite cone situated at the extreme northern end of the island. Age dating suggests that the main cone building activity occurred between 1.5 and 1.0 million years ago. Exposed on the volcano's flanks however are a number of unconsolidated valley-fill block and ash flow deposits suggesting more recent activity. One of these deposits, on the north-east flank of the volcano, has been recently dated at > 46,000 years B.P. Other evidence of potential activity from this center includes the presence of warm (27°C), acidic (pH 1.6), sulfate-rich springs on the summit of the volcano, hot springs on the coast, and the occurrence in 2002 and 2003 of shallow earthquake swarms partially located beneath the volcano. Morne aux Diables is dominantly composed of deposits of block and ash flows and associated domes from Pelean-style activity, however, semi-vesicular andesite block and ash flows and surges (Asama-style activity) and pumiceous lapilli falls (Plinian-style activity) are locally abundant. The Pelean domes are located both in the summit region and along the southern flanks of the volcano. Petrologically, the volcano is composed of a monotonous series of porphyritic andesites and dacites containing phenocrysts of plagioclase+augiteñhypersthene with very sparse crystals of hornblende and quartz. Petrological models suggest the Morne aux Diables andesites and dacites can be produced by fractional crystallization of basaltic magma (such as those erupted from centers such as Morne Anglais and Morne Plat Pays in the south). Minor variations within this suite of andesites and dacites can be related to upper crustal fractionation of phenocryst phases.
V53B-1558
Stratigraphic and Granulometric Studies of three Pyroclastic Sequences from Soufriere Volcano, St. Vincent, Lesser Antilles.
Soufriere volcano on the island of St. Vincent is a complex strato-volcano that has been active since at least the Late Pleistocene. Stratigraphic and granulometric studies have been carried out on three pyroclastic sequences from this volcano: 1) Deposits from Late Pleistocene eruptions. These deposits, which range in thickness from 50m in the north east to 2 m in the south and cover about 55% of the island, form a complex stratigraphy produced by a variety of eruptive styles including Plinian, Vulcanian and Strombolian; 2) Deposits from the 1902 eruption. These were studied in a series of newly excavated pits near the Wallibou river on the west coast. The stratigraphy of these deposits indicate an initial sequence of surges and flows overlying a well developed paleosol in the deposits from the 1812 eruption, followed by many thin fall beds. Both fall and flow deposits are composed of a combination of vesicular juvenile clasts, lithic fragments (altered and unaltered), juvenile crystals , accretionary lapilli and accreted grains. Although the abundance of accretionary lapilli, and accreted grains in most deposits indicates a significant hydrologic component throughout the eruption, the variations in their amounts between the different beds suggest that the water-magma contact fluctuated significantly during the eruption; 3) Deposits from the 1979 eruption. These are dominantly fall deposits with abundant accretionary lapilli, plus minor pyroclastic flows and surges. The fall deposits, which can be subdivided into a lower blue-grey (explosive events of April 13 and 14) and an upper yellowish brown unit (explosive events April 17-26), generally show a bimodal grain-size distribution with the secondary coarse mode usually produced by the abundant accretionary lapilli. The earlier eruptions (April 13-22) all had a general E-W distribution, in contrast the maximum distribution for the April 26 eruption was NW-SE. The presence of abundant uncharred wood and cauliflower-surfaced bombs in the pyroclastic flows, and the abundance of accretionary lapilli together with the fine-grained nature of the fall deposits all suggest that the eruptions were phreatomagmatic in origin. The apparent absence of juvenile fragments in the lower pyroclastic flows suggests that the initial eruptions represented the explosive shattering of the 1971 dome that existed in the crater prior to the 1979 eruption.
V53B-1559
Correlation of Tephra Deposits at Newberry Volcano, Oregon: Refining Estimates of Plinian Eruptions
Preliminary field data suggest that some of the 40-50 Pleistocene tephras identified by Kuehn (2002) at Newberry Volcano, central Oregon, may represent different lobes of tephra generated by compositionally zoned, caldera-forming eruptions, rather than discrete eruptive events. Newberry volcano is a large, bimodal shield volcano. The physical, mineral and chemical properties of each tephra deposit were recorded in over 30 test pits on the eastern and northeastern flank of the volcano. The individual components of each tephra deposit were described in 10 cm increments for each pit. The 99-5 tephra, previously thought to represent a discrete eruptive event, is proposed to be part of the Evans Well tephra. Stratigraphic relations from Kuehn dictate that the Evans Well tephra was emplaced during a large plinian eruption between 150,000 and 55,000 yr BP. The 99-5 and Evans Well tephras are composed of coarse, white/light-gray, subangular, fibrous to frothy pumice with rounded vesicles. Pumice diameters range from 2-8 cm and the coarsest pumices are pink or have a pink interior. Both tephras contain mafic and banded pumice throughout, with an increase in concentration at the top of the deposit. Red, black (usually bomb fragments) and purple volcanic lithic fragments are located throughout each of the deposits, with concentrations ranging from 15% (99-5) - 25% (Evans Well). The two tephras nowhere overlap despite their proximity. Other tephra deposits in similar stratigraphic positions at Newberry lack concentrations of mafic and/or banded pumice at the top of the deposit, contain pumices with different physical properties (size, color, vesicularity, inclusions), contain different accidental and juvenile material and have different chemical signatures. Preliminary major, minor, and trace element chemistry analyses show that tephra 99-5 and Evans Well are chemically similar and both tephras show chemical variations with depth. The majority of major and minor elements analyzed in tephra 99-5 are within an acceptable range of values for the corresponding elements in the Evans Well tephra. For example, the SiO2 variations in 99-5 range from 60.1-70.5% and the Evans Well values of 61.9-68.6% fall within that range making the correlation of these two tephras chemically valid. The Evans Well tephra is stratigraphically located between the Pumice Flat eruption, interpreted to have a 40 km plinian column and the caldera-forming Paulina Creek eruption. The stratigraphic position of Evans Well places it in a very active eruptive time period for Newberry Volcano (150,000-55,000 yr BP). Deposits of interpreted Evans Well tephra are found within a large area on the volcano's eastern flanks and in most locations are 2 m thick, similar to deposits from the Pumice Flat eruption. Evans Well, combined with tephra 99-5 may represent a large plinian eruption similar to that of Pumice Flat.
V53B-1560
Magma plumbing system and eruption mechanism in recent eruptions of Haruna volcano, central Japan
In last tens of thousands years, volcanic activity in Haruna produced at least five lava domes along with pyroclastics deposits (Oshima, 1986). For future eruptions, author revealed eruption trigger and magma plumbing system petrologically, focusing on a latest eruption in the middle of 6th century. The eruption started with Plinian phase (1.2 km3; Arai, 1993), followed by pyroclastic flow, and ended with lava dome emplacement. White pumices are found in all units of Plinian and flow phases, while gray and banded pumices are limited to the first Plinian phase. Involved magmas are felsic magma (60 wt% in SiO2; with phenocrysts of orthopyroxene, amphibole, Fe-Ti oxides and plagioclase; ca. 850C), aphyric mafic magma (52wt% in SiO2; ca. 1130C) and phyric mafic magma (with olivine and clinopyroxene; ca. 1130C). Mixing between the felsic and mafic magmas formed gray part of pumice and dome lava. The mafic endmember is aphyric in the pumice and is phyric in lava. The felsic magma appeared without mixing in middle of the eruption, but that in the first Plinian phase has heating record (-880C). Olivine (-100micrometer) in gray part of pumice is from the aphyric mafic magma. Their cores have similar compositions as olivine phenocryst rims in lava, indicating that aphyric mafic magma had the same composition as melt part of the phyric mafic magma and mafic magmas are from a storage system. Melts program yields more than 4kbar for olivine and clinopyroxene to be the first liquidus phases in melt with aphyric magma composition. On the other hand, storage pressure of felsic magma can be ca. 4 kbar. Thus, the felsic reservoir locates over shallower depth and is separated from the mafic reservoir. Separation of melt part from a phyric mafic magma ascending toward felsic reservoir may have caused earlier supply of the aphyric mafic magma. The felsic magma was nearly rigid due to high phenocryst abundance (ca. 50 vol%). So, the vent opening by the mixed and heated magmas with lower viscosity made the felsic magma erupt.
V53B-1561
Late Holocene Andesitic Eruptions at Mount Rainier
Holocene Mt. Rainier erupted much more frequently than is recorded by its 11 pumiceous tephras. In the 2.6-2.2 ka Summerland eruptive period, 6 groups of thin (1-5 mm) Sparsely Vesicular Glassy (SVG) ashes were deposited (S1-S6), followed by the 0.3 km3 C-tephra. Two groups of andesitic lava flows and one andesitic block-and-ash flow (2.45 ka) also erupted in the Summerland period (ice conceals any other products). Based on glass composition the pyroclastic flow correlates with S4 ashes that also contain pumiceous grains and rare pumice lapilli. The first of the lava groups, exposed in windows through the Emmons and Winthrop glaciers, is Sr-rich for Mt. Rainier eruptives and correlates with S5 & S6 ashes based on similar high-Sr plagioclase. The ensuing C-tephra formed by plinian eruption of mixed and mingled magma comprising 4 juvenile components: mixed porphyritic andesite pumice, crystal-poor andesite scoria, vesicular high-Sr dacite blebs in pumice and scoria, and poorly inflated crystal-rich high-Sr dacite. High-Sr components were probably entrained conduit linings and segregations from the preceding high-Sr eruptions. The youngest lava group, exposed at the summit, is normal-Sr andesite lacking mixing textures of the C-tephra, and represents eruption of another small batch of andesitic magma perhaps just after the C event. SVG ash grains have blocky-to-fluidal shapes, are rich in plagioclase microlites, and their glasses are high-SiO2 (66-78%) and low-Al2O3 (15-11%). Melting experiments yield apparent equilibration pressures <50MPa for SVG liquids. SVG ashes likely result from shallow hydromagmatic explosions as largely degassed magmas transited the upper-edifice hydrothermal system during effusive eruptions. Rare pumice lapilli codeposited with S1, S2, and S4 ashes have microlite-free dacitic glasses, one with nonreacted hbl phenocrysts. These pumice formed from magmas that ascended rapidly from reservoir depths, synchronous with or closely between effusive-hydromagmatic eruptions. Mt. Rainier's late Holocene activity was typified by repeated arrival and eruption of slightly different andesitic magmas. Most eruptions were effusions of largely degassed magma, accompanied by near-surface explosions that blanketed the proximal region with fine-grained glassy ash. Associated rapidly ascended magma led to sparse pumice, pyroclastic flows, or plinian tephra fall, depending on amount.
V53B-1562
The Basalt of Yellowjacket Butte, Another Large and Interesting Lava Flow at Medicine Lake Volcano, N. California, USA
The late Pleistocene basalt of Yellowjacket Butte (BYB) covers ~300 km2 on the SE flank of Medicine Lake volcano (MLV) in the N CA Cascade Range. Combined field mapping, geochemistry, paleomagnetism, and Ar dating provide constraints on growth of this extensive basalt flow in space and time. Multiple vents on the upper S flank fed lobes via lava tubes to >30 km. Estimated volume of this compositionally-zoned plag-ol basalt is 4-5 km3, most in the earliest lobes. The E lobe is largest, covering >one 7.5' quadrangle. It contains 50.0-51.4% SiO2, 7.4-8.3% MgO, and ranges in Mg\# from 62-66. The smaller SW lobe overlaps the E lobe in composition, but is somewhat less mafic (50.7-51.7% SiO2, 7.0-7.6 MgO, Mg\# 61-64). Both display late, low-volume, high-Sr facies (400- 465 vs. 350-400 ppm for the main lobes). Overlying the high-Sr E and SW lavas are the Hole-in-Rock (HR) lobe and the youngest lava, the NW lobe. The HR lobe typically has higher and more variable SiO2 (50.4-53.1%), lower MgO (5.2-6.3%) and lower Mg\# (50-55), whereas the NW lobe has the lowest SiO$_{2} $ (49.3-50.7%) despite its mid-range MgO (6.8-7.8%) and Mg\# (56-61). Sr concentration is similar in both youngest lobes (370-430 ppm), which are partly buried by late Holocene lava, but have ~subequal volumes, larger than either of the late high-Sr facies of E and SW lobes. BYB is distinct from other large MLV compositionally-zoned basalts in having little compositional variation in SiO2, despite considerable (although complex) variability in TiO2, FeO* (7.5-10.2%), K$_{2} $O (0.2-0.9%), and other elements. Paleomagnetic sampling at 19 sites indicates a correlation between stratigraphic position, chemical composition, and magnetic inclination (I) value. A >10° shift to shallower I values was found, while declination (D) values were essentially constant as the eruption progressed. The farthest-traveled early E lavas have the highest I ($78-80°), whereas the SW and near E lavas have I=$75-77°. TiO2 contents for these earliest lavas range from 0.85-1.02%. Three sites in late high-Sr lavas yielded I of $72-74° and higher TiO2 (0.96-1.16%). Three sites in the HR lobe and five in the last-erupted NW lobe yielded indistinguishable I averaging $67 °. HR lavas have the highest TiO2 (1.24-1.45%), while the NW lavas have 1.06-1.20%. The large range in I values combined with unchanging D values from earliest to latest lavas indicates that eruption duration may have been as much as a few hundred years. BYB is unusual in displaying significant variability in remanent magnetic direction, in contrast to limited variation in magnetic direction in other large-volume basalts we have studied at MLV, Newberry Volcano, and elsewhere. The 40Ar/39Ar dating of one BYB sample yielded a 1-σ plateau age of 86±14 ka, an isochron age of 76±21 ka, and an inverse isochron age of 73±31 ka. These ages overlap the ~80 ka age of the Norwegian Sea paleomagnetic event and its inferred on-land equivalent, defined as the Badlands Excursion at Newberry Volcano OR (Champion et al., 2004 AGU abstract). It is possible that BYB took no more time (decades) to erupt than most basalts of similar volume in similar tectonic environments. Thus, this basalt may record faster geomagnetic secular variation during an early or late part of the Badlands Excursion. This leads us to speculate that eruptive events may have been roughly synchronous at MLV and Newberry Volcano, two volcanic systems located in similar extensional environments at the transition between the Cascades and the Basin and Range.
V53B-1563
Quantitative Analysis of 0-2 Ma Volcanic Vent Distribution in Space, Time, and Composition Centered on Medicine Lake Volcano, Northern California, USA
525 volcanic vents 0-2 Ma were identified during geologic mapping of approx. 2500 sq km centered on Medicine Lake volcano (MLv), northern Calif. More than 95% of the mapped area consists of volcanic rocks 0-2 Ma. Vent locations were digitized, and each vent assigned to 1 of 5 time slices (0-12 ka, 12-25 ka, 25-120 ka, 120-780 ka, and 0.780-2 Ma) based on radiometric dating and one of 5 chemical composition groups. In this preliminary study, no attempt is made to interpret locations of vents buried by younger units or relate vents to areas or volume of individual volcanic map units. Vent data were analyzed using a GIS. 53% of vents are basalt and mafic andesite (<57% SiO2 but not classified as low-K tholeiite), 20% are low-K tholeiite (SiO2 <49%, K2O2 always <0.5% and commonly <0.39%), 11% are andesite (57-62% SiO2), 3.5% are dacite (62-70% SiO2, and 8.5% are rhyolite (SiO2 >70%). Considering only mapped vents, one vent erupted on average every 3800 yrs during the last 2 m.y. This frequency is clearly too low because vents are unknown or buried for 56 volcanic map units and, therefore, not included in calculating this eruption frequency. If one assumes one vent for each of these 56 units, the interval between vents falls to 3400 years. This frequency is still too low because 1.) Most mapped units with identified vents have multiple vents and 2.) There is a significant decrease in the average time between vent formation in successively younger time slices. About 90% of vents are widely distributed (vent densities <1.3 per sq km) west of -121.48 deg which marks an abrupt N-S trending eastern boundary for all vents in the MLv area. Four conspicuous areas with vent densities between 1.3 and 6.5 vents per sq km form a trend parallel and approximately 5 km W of this abrupt boundary. A second trend of similar vent densities defines a trend of 030. Both trends are present in the 4 youngest time slices, although stronger in some than in others. The two trends intersect near long. -121.5 and 41.8 N in the northwestern part of Lava Beds National Monument, which is also the area of highest vent density, 6.5 vents per sq km. There is no particular concentration of vents in the vicinity of the summit of MLv with vent densities below mean vent density (0.234 vents per sq km, due, at least in part, to subsidence and younger volcanic units covering older ones.
V53B-1564
Distributary Systems in Basaltic Flow Fields: Insights from Mauna Ulu, Kilauea Volcano, Hawaii
This investigation examines the significance of lava transport in features intermediate in scale to small, toe-dominated pahoehoe lobes and large lava tubes to assess controls on flow field development. Field observations show that channel systems of moderate size are responsible for emplacement of a significant portion of the final surface lava of the Mauna Ulu flow field (1969-1974) at Kilauea Volcano (HI). Channel networks and individual channels with segment widths of a few meters across or less control the distribution and morphology of flow surface units in parts of the medial zone of the Mauna Ulu flow field. Complex flow surface unit patterns can be interpreted in terms of local conduit evolution, including generation of sequences of lateral levees, production of zones with slabby and clinkery lava textures along flow fronts and margins, and later stage breakouts of pahoehoe at flow margins and within interior zones, sometimes burying the primary channel. In order to understand conduit evolution, breakout history, and local flow stratigraphy, morphometric and topographic surveys of flow conduit systems (channels, levees, and breakouts) were completed to quantitatively characterize levees and breakout lobes and measure channel dimensions, flow thicknesses, and levee volumes. This work utilized a dGPS unit coupled with a high-resolution laser profiling system providing sub-meter horizontal (~10 cm) and vertical (~30 cm) accuracy. Preliminary analyses of areal resurfacing from four conduit systems with channel segments between 55 and 166 m in length have been completed. Total resurfaced areas are between ~500 and 2500 square meters, with 70 percent or more in levees and lateral and distal rubble zones and the remainder in late-stage pahoehoe breakouts. Along a given channel segment, ~10-30 square meters are resurfaced for each meter of channel length, suggesting the duration of activity in small to moderate conduits can be a key factor in understanding flow field evolution.
V53B-1565
The Highway Flow in the Craters of the Moon Lava Field; A New Interpretation of Eruption History
The Highway flow is a trachyandesitic lava flow that lies west of Sunset Cone in the Craters of the Moon lava field, and is part of a complex series of eruptive events that took place in the vicinity of North Crater during the most recent eruptive episode of the field (2,500-2,000 years ago). The trachyandesitic Devil's Orchard and Serrate flows originated from the same vicinity during this period, and carried large masses of rafted cinder cone blocks and disaggregated cinder cone material up to 12 km to the east. The rafted blocks have long been thought to be the remnants of a disintegrating North Crater cinder cone. However, recent work has established that the flows carry a cumulative volume of rafted blocks too large to have come from the existing breach in North Crater. Previous studies had postulated the Highway flow vent to be on the north side of North Crater, and suggested the lava had flowed north, and later drained back to the south when the Highway fault downdropped the vent area. This study examined lava flow characteristics and flow direction indicators based on flow fold patterns interpreted from aerial photographs and field mapping to provide new insight into the complicated history of the Highway flow. We found two vents on the northern side of the flow, and evidence that the lava flowed to the south and ponded against a formerly present cinder cone, which we have named "South Highway Cone". The following sequence of events for eruption of the Highway flow is hypothesized: 1) eruption from two linear vent systems in the northern part of the present day flow 2) flow from the second vent southward, down the valley between Sunset and Grassy Cones 3) ponding of the lava flow at the base of the South Highway Cone and 4) downdrop to the south along the Highway fault due to collapse into a lava chamber beneath South Highway Cone. The rafted blocks in the Devil's Orchard and Serrate flows represent pieces of North Crater and the now hypothesized South Highway Cone, which were carried away, but the exact manner in which these flows fit into the sequence of events has yet to be determined.
V53B-1566
Temporal Geochemical Variations in Lavas From Sakurajima Volcano, Japan
In historic times, Sakurajima volcano has erupted effusively four times, in AD 1471, 1779, 1914, and 1946. Yanagi et al. (1991) examined the major element contents of these four flows and showed that from 1471 to 1946, lavas became progressively more mafic with time from an initial dacitic composition. On the basis of data from the historic eruptions alone, they proposed a two-chamber magmatic system in which a wedge-shaped plagioclase-pyroxene cumulate plug sinks between a dacitic upper chamber and a basaltic lower chamber to drive the mixing process. In our study, we collected samples from all exposed Sakurajima lavas, extending coverage to approximately 12,000 years of eruptive history. Major and trace element contents of the lavas determined by ICP-MS and XRF reveal that the trend observed in the historic eruptions is only the most recent part of a cyclic variation in magma composition. Major and trace element contents define sinusoidal curves with time, varying systematically between more mafic and more felsic compositions, in two cycles. The oldest prehistoric lavas exhibit increasingly mafic compositions with each subsequent eruption, ranging from 67 to 61 wt.% SiO2. This pattern then temporarily reverses such that lavas become progressively more felsic prior to the 1471 eruption. After 1471, the mafic character of erupted material increases once again with time, as reported previously. Linear correlations of major and trace element compositions with SiO2 concentration and disequilibrium phenocryst assemblages indicate that magma mixing is contributing to the chemical changes observed in Sakurajima lavas. Mass balance calculations reveal that Sakurajima lavas have incorporated up to 55% mafic material into a parental dacitic magma. The cumulate-plug-settling model is consistent with the historical trends but does not explain the cyclicity. Instead, we propose a model with a shallow chamber of dacitic magma that is repeatedly flushed with mafic magma from a deeper reservoir, resulting in progressively more mafic compositions. A quiescent period between injections of basaltic magma re-starts the system, and the magma chamber evolves towards the parental dacitic melt composition again.
V53B-1567
Quantitative Flow Morphology, Recent Volcanic Evolution and Future Activity of the Kameni Islands, Santorini, Greece
The fundamental importance of careful field investigation, and the long term value of detailed published volcanic eruption reports, means that much can be learned about eruption processes even many decades after an eruption has ceased. We illustrate this with reference to the young dacite lava flows of the Kameni islands, Santorini. We have created a new, high resolution digital elevation model (DEM) for the intra-caldera Kameni islands, Santorini, based on new data from a recent airborne laser-ranging (LiDAR) and aerial photography mission. This DEM reveals a wealth of surface morphological information on the dacite lava flows that comprise the Kameni islands. When combined with a re-analysis of contemporary eruption accounts, these data yield important insights into the physical properties and flow behaviour of dacite magma during slow effusive eruptions. Kameni island lava flows exhibit the classic surface morphologies associated with viscous aa: levees, and compression folds. Levee heights and flow widths are consistent with a Bingham rheology, and lava yield strengths of (3 to 7)× 104 Pa. Analysis of the shapes of flow edges confirms that the blocky aa dacite lava flows show a scale-invariant morphology with a typical fractal dimension that is indistinguishable from Hawaiian aa. Dome-growth rates during eruptions of the Kameni islands in 1866 and 1939 are consistent with a model of slow inflation of a dome with a strong crust. Lava domes on the Kameni islands have a crustal yield strength (4×107 Pa) that is lower by a factor of 2 to 4 than the domes at Pinatubo and Mount St Helens. The dome height model, combined with the apparent time-predictable nature of volcanic eruptions of the Kameni islands, allows us to predict that the next eruption of the Kameni islands will last for > 2.6 years (in 2005) and will involve formation of a dome ca. 115 to 123 m high.
V53B-1568
Integrating laser-range finding, electronic compass measurements and GPS to rapidly map vertical changes in volcanic stratigraphy and constrain unit thicknesses and volumes: two examples from the northern Cordilleran volcanic province
We present preliminary results of laser-range finding-GPS surveys from two separate locations in northern British Columbia, in the south-central northern Cordilleran volcanic province: Hoodoo Mountain volcano and Craven Lake cone. This technique, described in detail below, is appropriate for rapidly measuring changes in vertical thicknesses of units that either would be difficult or impossible to measure by most other techniques. The ability to accurately measure thicknesses of geologic units in otherwise difficult-to-access locations will aide in generating better quantitative estimates of deposit geometries and eruption volumes. Such data is particularly important for constraining quantitative models of magma production and eruption dynamics. The deposits of interest in this study comprised at least partly inaccessible, largely pyroclastic units, although the technique could be used to map any vertical surfaces. The first field location was the northern side of Hoodoo Mountain volcano (56deg47'23.72'N/131deg17'36.97'W/1208m-asl), where a sequence of welded to unwelded, trachytic-phonolitic tephra was deposited in a paleovalley. This deposit is informally referred to as the Pointer Ridge deposit, and it comprises at least 7 distinct subunits. The horizontal limit of the exposures is approximately 1.5km, and the vertical limit is approximately 250m. Three different GPS base stations were used to map the lateral and vertical variations in the deposit. The second field location is north of Craven Lake (56deg54'44.55'N/129deg21'42.17'W/1453m-asl), along Craven Creek, where a sequence of basaltic tephra is overlain by pillow lava and glacial diamicton. This exposure is 200m long and approximately 30m high, much smaller than the area mapped at Hoodoo Mountain. The basaltic tephra appears to comprise 4 distinct sequences (measured thicknesses vary from 3-4m) not including the overlying pillow lava (measured thickness varies from 2 to 10m), and measurements of the sequences give average thicknesses of 5-10m. The laser-range finding-GPS setup used in these studies comprises an IMPULSE LR laser-range finder, a MapStar Module II electronic compass, and a Trimble ProXL global positioning (GPS) unit attached to a tripod specially designed to hold all three pieces of equipment. The three units communicate via a variety of cables. The maximum distance over which the laser can used is 500m (this distance varies for different lasers); the tripod was set at a relatively short distance from the exposure of interest at Craven Lake (95-115m), but further away at Hoodoo Mountain (up to 450m). The range finder was used to 'shoot' bottom and top contacts of each unit within the vertical faces. The distance and relative bearing were automatically transferred into compass and then to the GPS unit, producing in a map of the vertical face with horizontal and vertical coordinates. Analysis of the data provides detailed estimates of unit thicknesses across the vertical faces. The data collected can be imported into ArcGIS as a SHAPE file and overlain on DEM models for the areas of interest. ArcGIS extensions such as Spatial Analyst and 3D Analyst can be used to estimate surface areas and volumes for units mapped within the laser-GPS setup.
V53B-1569
Sources and Flow Directions of Dikes at Sinker Butte Volcano, Western Snake River Plain, Idaho
At least thirteen basaltic dikes are exposed in a 140° arc along the south and east sides of the Pleistocene Sinker Butte Volcano. The dikes are exposed for up to 500 m along strike and are between 28 and 115 cm in width. Most have nearly vertical dips. Greater abundances of plagioclase and olivine phenocrysts in the interiors of many of the dikes suggest they were concentrated by flow differentiation, although complex zoning and internal chilled borders in a few of the dikes are attributed to multiple magma pulses. The dikes are best exposed along the walls of the Snake River Canyon, where they intrude a thick series of phreatomagmatic tuffs sandwiched between sequences of lava flows. Some of the dikes cut the lower sequence of lavas, and several merge with the capping flows; however, the relationship between the dikes and the overlying lavas is not yet clear. Anisotropy of magnetic susceptibility (AMS) was determined for about 80 oriented drill cores collected from nine of the Sinker Butte dikes. Seven showed strong anisotropies (up to 9.7%). Average magnetic lineations, which are assumed to reflect flow directions, are nearly horizontal for two of the dikes, plunge about 45° downward and away from the presumed vent area in two others, and are nearly vertical for the other three. The AMS data indicating horizontal and inclined flow directions are consistent with the propagation of radial dikes outward from a magma column or lava lake located in the upper part of the edifice. Vertical flow directions may have been upward in response to the impermeable barrier at the base of the tephra section, or may have been downward in response to draining of magma from the central conduit or lava lake. Our model for the emplacement and transport of magma in the radial dikes at this center should be further refined by detailed field mapping and studies of crystal alignments currently in progress.
V53B-1570
Deformation of Scoria Cone by Conduit Pressurization
A simplified mechanical model is used to simulate the deformation of a scoria cone due to pressurization of magma in a feeder conduit. The scoria cone is modelled as consisting of a cone of stabilized scoria with an axial region of loose scoria (height h$_{1}$), all overlying a vertically oriented cylindrical conduit intruded into rhyolite tuff country rock. For our analyses, the conduit is filled with basalt magma, usually with the upper length (h2) solidified. The style of deformation of the cone depends on both h$_{1}$ and h2. If magma is prevented from hydrofracturing out of the conduit (as, for example, might be the case if the magma is surrounded by a solidified, but plastically deformable layer acting as a gasket backed up by the brittle country rock) pressures in the magma can build to 10s of MPa. When h$_{1}$ is 100 m, not unusual for a small isolated basaltic cinder cone, the magma pressure needed to destabilize the cone when molten magma extends all the way to the original ground surface (h2 = 0) is only about one-third of the pressure when the upper part of the conduit is solidified (h2 = 25m). In the former case, almost the entire upper third of the cone is at failure in tension when the configuration becomes unstable. In the latter case, small portions of the surface of the cone are failing in tension when instability occurs, but a large volume in the central core of the cone is failing in shear or compression. These results may provide insight into the status of volcanic plumbing, either past or present, beneath scoria cones. Field observations at the Lathrop Wells volcano in southern Nevada identify structures at the outer edge just below the crater rim that appear to be inward-dipping listric normal faults. This may indicate that, near the end of its active stage, the cone was close to failing in this fashion. Such a failure could have been quite energetic had it occurred.
V53B-1571
The Explosion of March 2004 at Montserrat: Constraints From Borehole Strain Data
The dome-building volcano Soufriere Hills on Montserrat in the Carribean has been active since about 1995. By the end of January 2003, the CALIPSO group had established a small network of 4 borehole sites with the instrumentation including Sacks-Evertson dilatometers. Major dome collapse in July 2003, reported elsewhere, was accompanied by increased pressure in a deep (~5km) reservoir. In March 2004 a well observed and recorded explosion took place (Green and Neuberg, 2005). This event resulted in clear, coherent signals (peak amplitudes up to ~100 nanostrain) on the 3 strainmeters then being recorded. The amplitudes and polarities of the strain signals from this event cannot be satisfied by a single Mogi-like source: two closer sites at similar distances (but different azimuths) record similar amplitudes but opposite polarities while the most distant site has amplitude less than one-fifth that of the nearer sites. The data are well satisfied by a small shallow dike striking N55W, consistent with geologic fabric, together with corresponding loss of pressure in the deep reservoir. Seismic activity accompanying the explosion does not commence until some time after the onset of the strain event, i.e. after the dike begins to form. Later, the magma chamber is re-pressurized.
V53B-1572
The Intrusive Process Leading to the Mt. Etna 2001 Flank Eruption: Constraints From 3D Attenuation Tomography
We have analyzed 291 shallow seismic events, leading to the July 17-August 9, 2001 flank eruption at Mt. Etna volcano, to obtain high resolution three-dimensional (3D) images of seismic attenuation of P waves (Qp) down to 2 km depth. 2293 P wave spectra have been inverted for seismic attenuation to calculate t* values, considering a frequency-independent Q, and a w2-type source model. The 3D tomographic images reveal an anomalous region of very low Qp values (between 10 and 30), located in the same place where the 2001 dike emplaced. We interpret this anomaly as the effect of fluid intrusion (magma rich in gas) in the uppermost part of the Etna volcano, just before the 2001 eruption. This confirms that attenuation is a physical parameter sensitive to the thermal state of crustal volumes through which seismic waves travel.
V53B-1573
Rerouting of Volcanic Heat Fluxes in the Mt.~Wrangell Caldera Glaciation after the 1964 Alaska Earthquake
The ice-filled caldera at the 4000 m summit of Mt. Wrangell, Alaska underwent significant changes in the years after the magnitude 9.2 Good Friday earth quake in 1964. Measured surface velocities in the Caldera center changed both in magnitude and direction between 1965 and 1976, but remained stable since. Since the ice is very cold ($-20^{\circ} $C 20 m below the surface) these changes can only be due to altered boundary condition at the glacier base. There, ice is melted at a high rate (about 0.6 m/a) which was inferred from observed accumulation rates and velocity fields. Between 1965 and 1976 the surface flow field changed from horizontal convergence to divergence in the Caldera center. Between 1976 and 2004 the submergence velocities decreased by 20 - 50 %, while the ice thickness increased by 8m. Both observations lead to the conclusion that the melt rate at the Caldera base has significantly decreased. Also within 12 years after the earthquake $40· 10^6 $m$^3$ of ice melted in the North Crater. Simultaneously, the ice volume increased in the West Crater which had been the most active of the three craters along the Caldera rim. We speculate that the changes of ice volumes in the craters and of the flow velocity field in the Caldera resulted from changes in the hydrothermal system of the volcano.
V53B-1574
Tomographic Images of Volatile Rich Magma Intrusions Leading to the 2001 and 2002-2003 Mt. Etna Eruptions
Seismic monitoring with proper sampling in space and time, and repeated tomographic analysis (4-D tomography) on active volcanoes, can be considered useful tools for the detection of volatile-rich magma intrusions, responsible of explosive eruptions. Seismic Vp/Vs tomographic images at depth less than 5 km b.s.l., obtained analyzing 2001 and 2002-2003 Mt. Etna eruptions datasets, highlights both volumes where gas-saturated magma intrusion occurred (Vp/Vs negative anomalies) and regions enriched in fluids (Vp/Vs positive anomalies). The 4D images of the Vp/Vs velocity ratio suggest that these eruptions can be related to intrusions of volatile-rich basaltic magma emplaced some months before their onset, as supported by recent seismological, geodetic and geochemical evidences. These results constitute the first unambiguous detection of Vp/Vs changes at Mt. Etna volcano, representing a possible tool to recognize high level of explosive eruptions.
V53B-1575
Origins and Implications of Zigzag Rift atterns on the surface of lava lakes
The distinctive rift patterns observed on newly formed lava lakes are very likely a product of interaction between heat transfer (cooling of lava) and the ability of the solid crust to deform in response to applied stresses. Ragnarsson et al. ({\it Phys Rev Lett.} 1996) observed similar features in analog wax experiments. The experimental setup consisted of a layer of liquid wax heated from below and cooled from above to create a solid crust, which is then pulled apart to form a rift filled with liquid wax. Of particular interest for lava lakes is a regime in which "zigzag" shaped rifts form. We performed a similar series of analog wax experiments designed to characterize the symmetric zigzag rift patterns associated with the cooling and deformation of a newly formed crust layer. The experimental setup is similar to that in Manga and Sinton ({\it JGR} 2004). We vary the wind speed, which controls the cooling rate, and pulling speed. The rifts are characterized by two quantities: the angle between rifts and a line perpendicular to the pulling direction $\phi$, and the amplitude of the zigzags $A$. We find experimentally the relationship between $\phi$, $A$ and cooling rate. We also develop a model to explain the observed relationships: $\phi$ is determined by a balance of pulling and solidification speeds, the latter being governed by wind speed (the primary control on cooling rate); the amplitude $A$ is limited by the thickness of the solid wax crust. The theoretical model is based on a local energy balance in which the conductive heat flux through solidifying crust is balanced by convective and radiative heat loss to the overlying air. Model predictions agree well with experimental results. We can scale this model to basaltic lava lakes. Observed angles are consistent with model predictions. The critical thickness of the solid crust, limiting the amplitude of zigzags (and rift propagation) is a few centimeters, consistent with other estimates (Harris et al., {\it JVGR} 2005). The model enables us to calculate crust spreading speed, or crust thickness, given the measurable $\phi$ and $A$. This makes it potentially useful in other applications where these quantities are important.