Volcano Hydrology II
Presiding: S Hurwitz, U.S. Geological Survey; J Major, U.S. Geological Survey Cascades Volcano Observatory
V14A-01 INVITED 15:35h
Volcano-ice Interaction Under Glaciers, Interplay of ice Rheology, Volcano Construction and Subglacial Hydrology
Eruptions under temperate glaciers can be complicated events, controlled by the interplay between ice thickness, surface and bedrock geometry, hydrological conditions at the base, eruption rate and magma composition. However, the most important characteristic is rapid melting of ice during extremly fast heat exchange from magma to ice. Usually this occurs with almost simultaneous subglacial drainage of the meltwater. Ice thickness seems to be a controlling factor in determining the rheological response of a glacier in a volcanic eruption. Where the glacier is thin (100-200 m or less) it responds to the rapid deformation rates caused by melting and drainage mainly with brittle fracturing, leading to rapid penetration of the glacier and subaerial phreatomagmatic eruption. Where the ice is several hundred meters thick, ductile deformation appears to be dominant, leading to rapid inflow of ice towards the eruption site. In this case the rate of melting is controlled by the eruption rate. The best-studied eruption of this type was Gjalp, Iceland, in 1996. At Gjalp, basal water pressure at the vents was 3-6 MPa during early phases of the eruption and all indications suggest that activity was explosive. The buildup of the 6-km long NNE-SSW trending edifice under the glacier modified the subglacial hydraulic regime which in turn influenced the form of the edifice. The southern and central parts of the edifice are steep and narrow as a consequence of ice confinement; water pressure was always much lower than glaciostatic. In the northern part, where the water pressure may have exceeded glaciostatic, the edifice is low and wide, suggesting only minimal confinement by ice. Removal of subglacially-erupted tephra by meltwater was relatively minor at Gjalp. However, where conditions are favourable, a large part of the tephra may be flushed from the eruption site. This tephra may eventually be deposited outside the glacier, on outwash plains or in the ocean.
V14A-02 INVITED 16:00h
Response of peakflow discharges to the 1980 Mount St. Helens eruption: Seasonality and effects of channel flow resistance
Decades of streamflow measurements that precede and follow the major 1980 eruption of Mount St. Helens (MSH) provide an exceptional opportunity to examine the response of peakflow discharges in large (300-1300 km2 drainage area) basins to devastating landscape disturbance. About 100,000 ha of landscape across multiple basins surrounding MSH were reconfigured by volcanic processes that included a huge debris avalanche, a widespread lateral blast and associated pyroclastic flow, large (to 108 m3) debris flows, and thick (>5 cm) plinian tephra fall. We determined the nature, longevity, and seasonality of discharge responses by linearly regressing logarithms of seasonal unit-area peakflow discharges compiled from pre- and post-eruption hydrographs from disturbed basins against paired discharges from a nearby control basin relatively unaffected by the major and smaller 1980 eruptions, and then compared regression models. We tested the null hypothesis that seasonal pre- and post-eruption regression models were coincident versus an alternative hypothesis that they were unequal. From 1980 to 1984, autumn discharges increased by a few to many tens of percent in basins heavily disturbed by the eruption. In contrast, pre- and post-eruption regression models for other seasons and later time periods do not differ significantly. Hydrologic responses to the eruption were thus strongly seasonal and short lived at the scale examined. Contrary to model predictions that smaller discharges would increase proportionately more than larger discharges, based chiefly on postulated changes in soil moisture and hillslope runoff, we found discharges across a range of magnitudes increased nearly proportionately relative to predisturbance regression models of unit-area discharges in several basins that sustained significant channel disturbance. Proportionate increases among small and large discharges are inconsistent with changes simply in hillslope hydrology. We attribute the proportionate increases in discharges to eruption-induced variations in channel geometry and flow resistance. We contend that transport of large amounts of easily erodible sediment created conditions that differentially affected flows of various magnitudes and that those conditions played as key a role in magnifying discharges as did perturbations to hillslope hydrology.
V14A-03 16:15h
Experimental Investigation of Water-Magma Mingling
Experimental laboratory simulations are investigating the hydrodynamics of submerged volcano under varying flow conditions and vent geometries to answer questions about phase mingling that precedes phreatomagmatic explosions. Experiments concentrate on understanding the flow structure in the absence of heat transfer, a reasonable initial assumption under film boiling conditions. General flow patterns indicate that the overall flow structure is predominantly a function of a vent Richardson number. Phase mingling is a function of Richardson number Reynolds number, viscosity ratio and Morton number (dimensionless surface tension). Particular emphasis is placed on determining threshold flow conditions under which phase mingling occurs and mingling characteristics including volume fraction and diameter. Phase mingling is correlated to the flow structure by simultaneously capturing particle image velocimetry (PIV) and laser induced fluorescence (LIF) images and superposing them with a beam splitter. LIF images identify phase boundaries while PIV images map flow velocity field.
V14A-04 INVITED 16:30h
Numerical Simulation of Flow and Deformation in Volcanic Area
Although the uplift of land surface in volcanic terrains is commonly interpreted to result from the inflation of a magma chamber at depth, an alternative explanation is the pressurization caused by upwelling of deep hydrothermal fluids. To explore the latter mechanism, the nonisothermal, multiphase flow model TOUGH2 is used in combination with a thermo-poroelasticity model BOIT2 to simulate flow and deformation. At this preliminary stage of investigation, the two codes are not fully coupled. Instead, pressures and temperatures computed by the TOUGH2 model are imported to the BIOT2 model to compute displacements. This modeling approach is illustrated by a cylindrical subsurface domain that is 50 km in diameter and 5 km thick. The permeability and porosity are assumed to be 1 x 10-15 m2 and 0.2 respectively. Atmospheric pressure and 15 oC are assigned to the top surface, which represents the water table. At all other boundary surfaces, there is no flow of fluid or heat across the surface. Hydrothermal fluids are injected at a rate of 5 x 105 kg (or 500 tons) per day at the center of the base of the cylindrical domain. The simulated land surface uplift after 5 years of continuous injection is of the order of 10 cm. This uplift rate is similar to those observed in volcanic areas such as Long Valley of California.