V14A-01 INVITED
Assessing Lava Flow Hazards from Mauna Loa: A Natural Laboratory
The primary goal of the U.S. Geological Survey's Hawaiian Volcano Observatory is to provide scientific information that can be used to reduce risks from volcanic activity. With detailed geologic mapping, we are using GIS to assess lava flow hazards for Mauna Loa. Mauna Loa makes up 51 percent of the surface area of the island of Hawai"i. Its lava flows extend 50 km or more from source vents and have reached the sea in less than 24 hours. Mauna Loa has been showing signs of inflation and will undoubtedly erupt again. Anything in the path of a flow will be buried, crushed, or ignited. Emergency managers need to know the areas threatened with inundation, the frequency of inundation, and the people, property, and facilities at risk. We have prepared several different types of analyses: topographic, inundation, economic, and recurrence, to assess the potential hazards that lava flows present to communities on the island of Hawaii. GIS has greatly facilitated our ability to provide hazards analysis which should serve as a guide for planning by emergency managers and the public. It has enabled us to quantify volcanic risk on Mauna Loa in ways never before attempted for any volcano.
V14A-02
Theoretical Model for Lava Flows in Confining Channels
Active lava flows down confining valleys are modelled as Newtonian viscous gravity currents. We determine the structure and the rate of propagation of the flow as functions of the form of the pre-existing topography, the effusion rate at the vent and the effective lava viscosity. When the confining boundary of the flow obeys a power- law with respect to the cross-slope coordinate, and the total volume of fluid obeys a different power-law with respect to time, the extent of the resulting flow can be determined analytically as a function of time for different channel shapes. The channel can be either horizontal or inclined, and its shape can vary gently along the downstream direction. The theoretical results are in excellent agreement with data from laboratory experiments, in which glycerine was released with either a constant volume or flux at one end of a long channel with either a semi-circular or V-shaped cross section. We will summarize the various different propagation rates that arise on different topographies. Applications to interpreting field data will be also presented.
V14A-03
Model studies for the length of a lava tube
Laboratory experiments with flowing wax over a cooled disc show that the wax forms channels of melt. For certain conditions, the channels allow steady flow indefinitely, and for others the channels freeze, forcing the wax to flow elsewhere or stop. This has motivated the study of the fluid mechanics and thermal balances of a viscous melted substance flowing in a cold circular tube. As the melt flows it cools and solidifies at the tube radius and we investigate the question "how far can the melt flow and remain liquid?" A theoretical solution is derived for the tube radius (which varies in space and time) and the temperature profiles in liquid and solid. If the melt source is maintained at a constant flux the distance can be infinite, but if the melt source is maintained at constant pressure difference across the length of the tube, then there is a maximum length which depends on the Peclet number and a dimensionless temperature. Conditions are derived for which the radius is unstable. This produces a bound on the distance the melt can travel. These predictions are further investigated with numerical and laboratory experiments. The formula for distance is use to predict the longest possible distance for lava tubes in volcanic systems.
V14A-04
How do Continental Flood Basalt Lava Flows Attain Their Great Length and Size?
I examine the conditions required to explain the great length of continental flood basalt (CFB) lava flows on Earth. Current estimates suggest that this is in excess of 1000 km in length, with total flow-field volumes approaching 10,000 cubic kilometers. Examples from the ~ 65 Ma old Deccan and ~ 15 Ma Columbia River CFB provinces will be presented. This makes them the largest eruptive units yet reported on Earth. The two simple requirements for a long lava flow are 1) sufficient magma held in a chamber to supply a large volume of lava, and b) an emplacement system that can transport the liquid lava from vent to flow front without the lava freezing. Other environmental conditions of lava emplacement also constrain lengths. For the common lava composition in CFB provinces (tholeiitic basalt) this is a cooling range of only 100 + 50 deg. C. During flow a combination of cooling and degassing promotes groundmass crystallization that is crucial in determining the resulting flow characteristics. An inventory of long lava flows in flood basalt provinces shows them to be exclusively pahoehoe flow fields, or closely related varieties, in type. This is mandated by the fact that `a`a flows are too thermally inefficient, losing ~ 2-5 deg. C per km in the feeder channels and rapidly developing groundmass crystals, to attain lengths much in excess of 100 km on Earth. Only the insulated flow model for pahoehoe flow fields where the dominant lava body is the inflated sheet lobe fits all the quantitative criteria required to explain the great extent of CFB flows. In this model, total effusion rates at the vent(s) are 1000s of cubic meters per sec, lava travels from the vent to the flow front in days to weeks, heat losses are restricted to as little as 0.001-0.1 deg. C per km, and eruption durations are years to centuries. These values dictate that eruptive volumes must be 10s to 1000s of cubic kilometers, a condition met solely by lava flows in CFB provinces. A variety of this lava flow type sometimes found in CFB provinces, informally named rubbly pahoehoe, also meets model criteria for yielding long lava flows.
V14A-05
Flow Fields of the 3.5 Ga Komati Formation, South Africa: Geochemical, Stratigraphic, and Temporal relationships between Massive, Vesicular, and Spinifex flows
A challenge of Archean volcanology is to reconstruct submarine flow fields by mapping and analyzing vertically dipping sequences of lavas. Some flow fields are bound by sediments and/or seafloor alteration that mark clear gaps in volcanism. Flow fields in the Lower Komati Fm are defined by alternating layers of komatiite (26% MgO) and komatiitic basalt (15% MgO). Five komatiite flow fields (100-200m thick) repeat the same stratigraphic zoning of spinifex overlying massive komatiite, and each flow field has a distinct Al2O3/CaO, a ratio unaffected by olivine fractionation, consistent with the contention that each komatiite flow field represents a distinct batch of mantle melting. Although massive and spinifex komatiite form distinct stratigraphic units on a map scale, detailed outcrop mapping reveals that the change in flow type represents a transition within a single flow field. In one type of transition, thin massive flows alternate with spinifex flow lobes of a compound flow unit. In another, a vesicular flow along the boundary links the underlying massive komatiite and overlying spinifex flows in time. The vesicular flow has alternating spinifex and vesicular layers that form a distinctive crust above a thick massive interior. Locally, this crust is tilted, intruded by massive komatiite from the interior, and overlain by a thick breccia including a spinifex flow broken into blocks and rotated like dominoes by the tilting. These outcrop relations indicate that spinifex flow lobes were starting to flow over the vesicular flow before it had undergone differential inflation, a temporal link between the lower massive and upper spinifex komatiites consistent with their belonging to the same flow field. The transition in flow type may reflect 1) an overlap of proximal and distal facies of komatiite flows as eruption rates waned and/or 2) thermal maturation prior to eruption. Early, cooler, crystal-rich, massive lava, flowing out as thick sheet flows, was replaced by hotter, crystal-poor, less degassed lava, flowing out as spinifex flows.
V14A-06
Spreading and Degassing of Lava Flows
Lava flows are complex types of viscous gravity currents. They are comprised of a mixture of several components : lava, gas and crystals, which influence the spreading of the flow. The physical properties of the flow, density and viscosity, depend on gas and crystal contents. Relative motions between phases due to density difference and compaction of the matrix also introduce strong complexity relatively to the single-phase gravity current theory. Using the 2-phase theory of Bercovici et al (2001), we have developed the general equations for the spreading of a two-phase mixture composed of matrix (lava) and fluid (gas). A loss of gas is considered through the upper boundary of the current. We compare the spreading of a 2-phase mixture, composed of lava and gas, to the spreading of a single-phase lava flow. We consider both constant volume and constant volume rate lava flows. As the current spreads, its surface increases and degassing becomes easier, which, in turn, increases the density and the viscosity of the flow. When the loss of fluid becomes significant, the evolutions of the peak thickness and radius diverge significantly from the single-phase evolution. It returns to it when degassing is complete. We derive the characteristic time for complete degassing of lava flows as a function of gas and lava physical properties. We also find a characteristic thickness for degassing to become significant and a characteristic radius for complete degassing. The self-similar shape of a two-phase viscous gravity current does not differ significantly from the shape of a single-phase flow.
V14A-07 INVITED
Spherulites Record Crystallization, Degassing, and Oxidation-reduction Mechanisms in Obsidian Flows
Insight into the formation conditions of obsidian lavas may be gained from studying crystals that grow in their interiors during eruption and cooling. We analyzed spherulites in an emergent obsidian dike from Hrafntinnuhryggur, near Krafla volcano, Iceland, in order to quantify their crystallization rates, the relations between spherulites and lithophysae, and how the growth of spherulites may affect the oxidation state of the magma. We measured water concentration profiles around spherulites in rhyolitic obsidian with synchotron Fourier Infrared spectroscopy. The concentration of OH- groups in the glass is elevated by up to 50% above the background level near the spherulite-glass border, and decreases radially away from the spherulite. This pattern reflects expulsion and diffusion of water away from the spherulites as they crystallized an anhydrous assemblage of albite+magnetite+quartz. We modeled the advective and diffusive transport of the water away from the growing spherulites by numerically solving the diffusion equation with a moving boundary. Numerical models fit the natural data best when a small amount (~10%) of post-growth diffusion is incorporated in the model. Comparisons between models and natural data constrain the spherulite growth rate to between approximately 0.1 to 0.3 mm/day for model eruption temperatures of 800 and 850°C. The expulsion of water during spherulite growth may affect the oxidation state of the melt, specifically by causing local reduction of melt adjacent to the spherulite. Optical evidence of this reduction includes clear glass zones enclosing the spherulites; outside of these halos the glass is uniformly brown. Near-infrared spectra and Micro- XANES measurements of the Fe3+/Fe-total ratio in these different color zones confirms that there is an enrichment of ferrous iron in the colorless glass adjacent to the spherulite. We suggest that reduction is driven by the release of water during growth of anhydrous minerals however the exact reduction mechanism remains unclear. A nearly one-to-one correspondence between the molar concentrations of hydroxyl groups and ferric iron that was reduced to ferrous iron in the clear glass zones suggests that the reduction may be due to an interconversion reaction involving molecular water and hydroxyl groups bound in the silicate glass.
V14A-08
The Evolution of Water Concentration in Rhyolitic Lava Flows During Emplacement and Solidification and Effects on Development of Flow Textures
Rhyolitic lava flows typically host spherulites, consist of radiating skeletal crystals of feldspar +/- quartz that nucleated on a crystal or a vapor bubble and/or flow bands. We have examined the association of mineral and rock microtextures with variations in water concentration in one flow banded, spherulite-bearing rhyolitic lava flow and two spherulite-bearing non-flow banded rhyolitic lava flows. All of the flows are approximately 24 Ma and are part of the Atascosa volcanic complex of southern Arizona. Fourier transform infrared microspectroscopy was used to analyze water concentrations and to map variations in water concentration across zones of interest in the samples. The Bartolo Mountain lava flow is flow banded, with gray thicker flow bands hosting larger, water-richer spherulites and glass, and orange thinner flow bands hosting smaller, water-poorer spherulites and glass. Skeletal crystals vary in their water concentrations, but water preferentially was partitioned into the surrounding glass during spherulite formation, which occurred during flow of the lava. Textures and water concentration variations suggest that flow banding reflects primary variations in water concentration in the melt, possibly associated with stretching of vesicles as the magma flowed. Spherulites from the Hell's Gate lava flow consist of two or more generations of skeletal radiating crystals, with each successive generation nucleating on the end of crystals of the previous generation. Single skeletal crystals are up to 300 microns in length, and are wider nearer the core of the spherulites. Water concentrations generally increases along the length of each generation of sanidine needles, although oscillation of water concentration has been observed. Water concentration also generally increases from the innermost sanidine generation to those that successively overgrow the spherulite. Overall, water concentration increases from approximately 600 ppm in the cores of spherlites to approximately 5000 ppm in the rims of spherulites. Coexisting quartz phenocrysts are deeply embayed, possibly reflecting resorption due to increased concentration of water in the melt when sanidine crystallization in spherulites caused partitioning of water into the surrounding melt. Melt inclusions in quartz phenocrysts inside of spherulites generally have higher water concentrations (6000 ppm+) than melt inclusions in quartz phenocrysts outside of spherulites. Perlitic glass that surrounds the spheulites contains 2.4 to 4.1 wt.% water, but water is heterogeneously distributed in the glass. Spherulites in the Atascosa Peak lava flow consist of up to six generations of skeletal sanidine growth, and record complex variation in water concentration. Fourier transform infrared analysis, with microscopy and mapping of water concentration, provides a local monitor of evolving water concentration during the quenching and crystallization sequence of the lava flows.