V51F-01 INVITED
Source Modeling and Seismic-Volcano Implications of the 2004-2007 Accelerated Deformation at Yellowstone Caldera
The youthful Yellowstone volcanic system is characterized by extensive earthquakes, episodes of cyclical uplift and subsidence, extraordinarily high heat flow, and widespread hydrothermal activity. In mid-2004, deformation of the 45-km-wide by 75-km-long Yellowstone caldera, measured by continuously operating GPS and InSAR, unexpectedly changed from subsidence to uplift at rates of up to 6.6 cm/yr that is three to four times faster than earlier deformation episodes. This pronounced uplift has continued to the time of this abstract submission, fall 2007, and was also accompanied by unusual subsidence of up to 4 cm/yr across the northwest caldera rim near the Norris Geyser Basin. Corresponding horizontal motions of 0.8-2.2 cm/yr and 0.7-2.0 cm/yr directed outward from the caldera and inward to the Norris area, respectively. Source modeling of the deformation data revealed an expanding sill-like structure 10 km beneath the caldera with a volumetric expansion rate of 0.11 km3/yr, consistent with the amount of magma required to supply the observed high heat flow of the caldera, and a contracting tabular body 8 km under the Norris area with a volumetric contraction rate of 0.01 km3/yr. The modeled expanding sill overlaps with the top of a tomographically imaged magma body, implying that the accelerated uplift is related to the inflation from the shallowest part of the magma chamber. The inflation of the caldera sill can furthermore induce dilatational strain in the surrounding volcanic rocks beneath the northern caldera rim, causing hydrothermal fluids to migrate into the caldera that can depressurize the Norris hydrothermal systems and cause the ground to subside. We also evaluate the elastic and viscoelastic stress evolution of this accelerated uplift to model the temporal changes of Columb failure stress on adjacent faults. The results will help us understand the interaction between the volcanic system and earthquake occurrence of the Yellowstone region.
V51F-02 INVITED
Seismic and GPS constraints on the dynamics and kinematics of the Yellowstone volcanic field
The seismically and volcanically Yellowstone hotspot resulted from interaction of a mantle plume with the overriding North America plate. This feature and related processes have modified continental lithosphere producing the Yellowstone-Snake River Plain-Newberry silicic volcanic field (YSRPN) system, with its NE volcanically active Yellowstone volcanic field. The size and accessibility of the Yellowstone area has allowed a range of geophysical experiments including earthquake monitoring and seismic and GPS imaging of this system. Seismicity is dominated by small-magnitude normal- to oblique-slip faulting earthquake swarms with shallow focal depths, maximum of ~5 km, restricted by high temperatures and a weak elastic layer. There is developing evidence of non-double couple events. Outside the caldera, earthquakes are deeper, ~20 km, and capable of M 7+ earthquakes. We integrate the results from a multi-institution experiment that recorded data from 110 seismic stations and 180 GPS stations for 1999-2004. The tomographic images confirm the existence of a low Vp-body beneath the Yellowstone caldera at depths greater than 8 km, possibly representing hot, crystallizing magma. A key result of our study is a volume of anomalously low Vp and Vp/Vs in the northwestern part of the volcanic field at shallow depths of <2.0 km. Theoretical calculations of changes in P- to S-wave velocity ratios indicate that these anomalies can be interpreted as porous, gas-filled rock. GPS-measured episodes of caldera kinematics reveals uplift and subsidence of the caldera at decadal scales with average rates of ~20 mm/yr but much higher short-term rates of up to 70 mm/yr of accelerated uplift, 2004-2007. The stress field inverted from seismic and GPS data is dominated by regional SW extension with superimposed volumetric expansion and uplift from local volcanic sources. Mantle tomography derived from integrated inversion of teleseismic and local earthquake data constrained by geoid, crustal structure, discontinuity structure reveals an upper-mantle low P and S velocity body extends from 80 km to ~250 km directly beneath Yellowstone and then continues to 650 km with unexpected westward tilt to the west at ~60° with a 1% to 2% melt. This geometry is consistent with the ascent of the buoyant magma entrained in eastward return-flow of the upper mantle. Some remaining issues to be discussed are: 1) the interaction dynamics and magma path from the tilted plume to the lithosphere, 2) the transfer mechanism of mantle magma through the lithosphere into the upper crust, 3) how the high potential energy of the large 12 m+ geoid high drives the dominant extensional strain and concomitant crustal magma emplacement, 4) how the crustal magma interacts with the surface hydrothermal features, and 5) how stress interaction of faults and volcanic features behave at short- to decadal time scales.
V51F-03 INVITED
The role of magma degassing on bradyseismic crises at Campi Flegrei inferred from the monitoring of the hydrothermal activity. Possible inferences on Yellowstone
The detailed processing of long time series of fumarolic compositions, available at Solfatara (Campi Flegrei) since 1983, carries a double thermobarometric signature. Temperatures of about 360°C, i.e. close to the critical point of water, are inferred by methane chemical-isotopic geoindicators, by gas equilibria in the H2O-CO2-CH4 gas system and by the H2/Ar geothermometer. These high temperatures are representative of a deep zone where magmatic gases flash hydrothermal liquid forming a gas plume where the kinetically fast reactive species (H2 and CO) re-equilibrate at temperatures of 200-240°C. The stable isotope compositions of H2O and CO2 shows that sampled effluents are mixture between magmatic fluids and the vapor generated by the vaporization of hydrothermal liquids of meteoric origin. A typical ‘andesitic' water composition has been inferred for the magmatic component, similar to that emitted by other volcanoes of the Mediterranean area. This mixing model suggests that bradyseismic crises, periodically affecting Campi Flegrei, are triggered by pulsing injections of CO2-rich magmatic fluids at the bottom of the hydrothermal system. A strong increase of the fraction of the magmatic component marked the bradyseismic crisis of 1982-84 and four minor episodes of ground uplift and seismicity occurred in 1989, 1994 and 2000 and 2006. The successful application of physical-numerical simulations asses the physical feasibility of the proposed conceptual model. Ground deformations, gravitational anomalies and seismic crisis can be well explained by the complex fluid dynamic processes caused by magma degassing events. Data on the fumaroles of other volcanoes suggest that magma degassing events frequently occur in dormant volcanoes causing crisis periods not necessarily linked to magma movement but rather to pulsating degassing processes from deep pressurized, possibly stationary, magma bodies. We can hypothesize that similar processes occur at Yellowstone that is characterized by the presence of a huge hydrothermal system and by the periodical occurrence of volcanic unrests episodes. Finally, resolving magmatic vs. hydrothermal fluid components needs for a better comprehension of the processes which occur at dormant volcanoes before an eruption.
V51F-04
Effects of the Yellowstone Hotspot on Western U.S. Stress and Deformation
The Yellowstone hotspot is a major source of regional deformation and driving stress of a large part of the western U.S. The high elevation and accompanying geoid anomaly of the Yellowstone Plateau, site of current hotspot volcanism, is due to buoyancy forces from low-density upper mantle material. Low-density mantle underlies much of the Intermountain West and is attributed to lithospheric modification associated with Basin- Range extension and crustal thinning, as well as locally high deviatoric stresses from mantle buoyancy forces. Along the Yellowstone-Snake River Plain, magmatic activity has reworked the crust, most notably through crustal melting and intrusions that result in a partially molten silicic magma chamber beneath the Yellowstone Plateau and a corresponding solidified, dense, mid-crustal sill beneath the eastern Snake River Plain. The reworked crust of relatively lower density adds to the local gravitational potential energy (GPE), leading to some the highest stress gradients in the western U.S. interior. The regional deviatoric stresses are predominantly tensional, corresponding to ongoing extension at the Yellowstone Plateau and Basin-Range as observed by GPS measurements and L. Quaternary fault slip rates. Our stress models, constructed from the CRUST2.0 velocity model modified by seismic data and a YSRP gravity-density model, allow the quantification of the relative hotspot stress contributions from mass variations in the upper mantle and crust. This deviatoric stress solution is combined with deformation models from GPS and fault slip data to estimate a stress field boundary condition to quantify effects of relative plate motions. Dividing the magnitude of estimated deviatoric stress by the magnitude of strain rate allows for the calculation of the effective lithospheric viscosity, an indicator of crustal strength. We account for effects of large earthquake post-seismic viscoelastic deformation to assess whether they improve the correlation between long-term deviatoric stresses and modeled instantaneous strain rates. Both the stress and strain rate models are compared with tectonic provinces and seismicity to identify microplates within the western U.S. The results of these models confirm the profound effect of the Yellowstone hotspot on the regional deviatoric stress field, and the predicted high stresses can explain the observed high strain rates of the YSRP.
V51F-05
Crustal Structure and Tectono-Magmatic Processes of the Yellowstone-Snake River Plain System From Gravity-Density Measurements and Strength Models Employing Seismic Constraints
The structure and composition of the Yellowstone-Snake River Plain (YSRP) system were analyzed from integrated modeling of gravity data from 33,209 stations in the YSRP and surrounding region. Recently derived tomographic velocity models, heat flow and temperature information, GPS-determined strain rates, and earthquake locations were used to constrain 3D density models. The density data were also constrained by velocity-density analyses based on petrologic information. These results were augmented by 1D strength profiles from representative tectonic and volcanic areas that were compared with earthquake focal depths. Results of this study suggest that the SRP lower crust has been thickened by the addition of an underplated layer 3 km thick composed primarily of clinopyroxene, with a density of 3.20 g/cm3. A mid crustal high-velocity sill occurs throughout the SRP, and is interpreted to be composed of a series of gabbroic lenses inter-fingering with the granitic upper crust. This geometry yields a bulk composition comparable to diorite and a density of 2.90 g/cm3. The sill varies laterally along the SRP from 4 to 11 km in thickness, resulting in the series of SW-NE trending gravity anomalies observed in the SRP. In Yellowstone, the density model is characterized by an upper- crustal partial melt 10 km beneath the caldera, 7 km beneath the northeastern side of the Yellowstone caldera, and extending up to 20 km north of the caldera boundary. The partial melt has a density of 2.52 g/cm3 for the caldera body and a significantly lower value of 2.47 g/cm3 for the northeastern caldera anomaly. Southwest of Yellowstone, the mid crustal SRP sill transitions to the Yellowstone partial melt. The transitional body has a density of 2.82 g/cm3. Strength models reveal that temperature has the greatest effect on crustal rheology of the region. The YSRP crust becomes progressively cooler with increasing distance from Yellowstone, and the shear strength increases from ~30 MPa in the Yellowstone caldera to ~50 MPa in the eastern SRP. The average thermal gradient in the upper crust decreases from 55 C/km in the Yellowstone caldera to 35 C/km in the eastern SRP (temperature data courtesy of David Blackwell, Southern Methodist University). In Yellowstone, the brittle ductile transition (based on mapping the 80th percentile maximum focal depths) is at 4 km depth and transitions into the Snake River Plain where the brittle ductile transition is at 8 km, coincident with the transition from the active volcanic Yellowstone system to the cool and stable SRP crust.
V51F-06
Numerical Simulations of Multi-phase, Multi-component Hydrothermal Fluid Flow: Implications for Heat and Mass Transport and Deformation of the Yellowstone Caldera
Ground surface displacements in large calderas such as Yellowstone have traditionally been modeled by the volume change of a discrete source embedded in an elastic or viscoelastic half-space, and the contributing roles of aqueous fluids and gases have largely been neglected. Because heat flux, hydrothermal fluid flow, and deformation are coupled in volcanically active areas, discrimination between crustal deformation resulting directly from magmatic intrusion and that induced by the associated injection of magmatic volatiles into the hydrothermal system has a direct impact on estimates of the magma source depth, geometry, and composition. The Yellowstone hydrothermal system contains multi-phase fluid and gas components extending to depths of several km, and has a very high efflux of CO2 (Werner and Brantley, 2003). We extend previous numerical simulations of fluid flow and rock deformation in an elastic porous medium using the coupled code TOUGH2-BIOT2 (Hurwitz et al., 2007) to include the effects of multi-phase (liquid-gas) and multi-component (H2O-CO2) flow. Simulations coupling hydrothermal fluid flow and mechanical deformation are sensitive to several hydrological parameters, including permeability, injection rate and depth of magmatic volatiles, and fluid composition. Phase distributions within the hydrothermal system are influenced by the water and gas depth and rate of injection, the temperature of the injected fluid, and the crustal permeability distribution. In simulations where a gas phase develops, ascent and expansion of a buoyant, high-enthalpy plume migrating towards the ground surface has a significant effect on the rate, magnitude, and geometry of ground surface displacements. For a plausible range of hydrologic parameters, water and gas injection yields simulated ground surface displacement rates (one to tens of mm/yr) and radii of deformation (tens of km) that are similar to those observed in Yellowstone, suggesting that ascent of magmatic gases towards the shallow crust and multi-phase dynamics within the hydrothermal system may explain some of the deformation observed in Yellowstone.
V51F-07
Seismic Source Evaluation of Possible Magmatic Related Earthquakes in the Yellowstone Volcanic System
The accelerated uplift of the Yellowstone caldera, 2004 to present (Chang et al., 2007) with uplift rates up to 7 cm/yr encouraged us to investigate the possible occurrence of earthquakes with fluid interaction. We initially chose an event with the deforming area, 1 September 2004, a Mc 3.3 normal-faulting earthquake. The magnitude is one of the largest since the beginning of the current uplift in Yellowstone. We examined the source characteristics of the event for such properties as earthquake as static stress drop, apparent stress, and radiation efficiency. The stress drop was estimated from the observed P-wave spectra by employing an empirical Green's function technique. We used the stacked spectra of 12 small collocated earthquakes ( Mc ~1.4) as a proxy for empirical Green's functions to correct attenuation effects along the source-station paths. Using a bootstrap technique, we determined the reliable range of the corner frequency at each station. Our evaluation revealed a stress drop is 47 MPa and its 95% confidence level is 8 to 109 MPa. To minimize the error in the estimation of stress drop, we used a 3-D local seismic velocity structure in Yellowstone (Husen et al., 2004). Given an appropriate density model and assuming hydrostatic pore pressure, the predicted maximum shear stress for normal faults at the focal depth is 26 MPa which is about half of the estimated stress drop. One of the possible explanations for the stress drop is pore pressure reduction around the focal depth. The total seismic radiated energy and the apparent stress were measured, assuming the ratio of P- wave to S- wave corner frequencies because signals for the direct S-wave are clipped. We found that the total energy ranged from 9.0 × 109 J to 9.6 × 1010 J, corresponding to apparent stress of 0.3 to 3 MPa, assuming the shear modulus of 32 GPa derived from local density and seismic velocity models. The estimated range of the apparent stress is consistent with similar earthquakes in the Lang Valley caldera. We also evaluated the radiation efficiency defined as the ratio of the apparent stress to the stress drop. Since estimates in stress drop and apparent stress are model dependent, we normalized the estimated radiation efficiency by the predicted one for a reference event with stress drop of 3 MPa. We found that the radiation efficiency for the Yellowstone earthquake is up to 30% less than one for the reference earthquake. Our result suggests that the rupture process of the initially evaluated earthquake involves relatively large energy dissipation. Additional analyses of several earthquakes occurring in close time and spatial proximity to the uplift area will also be shown.
V51F-08 INVITED
Yellowstone and Long Valley – A Comparison of Two Restless Calderas
Three large, silicic calderas in the conterminous United States have explosively erupted volumes > 300 km3 within in the last 2 million years -- Yellowstone caldera (Wyoming) Long Valley caldera (California) and the Vallez caldera (New Mexico) all located in extensional tectonic environments. All have shown varying levels of historic unrest. Pronounced unrest episodes at Yellowstone and Long Valley calderas over the past three decades stimulated extensive research on these two closely monitored calderas, and we explore some emerging similarities and differences. Yellowstone caldera is underlain by a long-lived (> 17 my) upper-mantle hot-spot that has fed a series of caldera-forming, extending to the southwest across southern Idaho to central Oregon including three caldera-forming eruptions from the Yellowstone caldera system in the last 2 my, the most recent at 600,000 ybp. It is marked by relatively low density and low seismic velocities extending to depths of at least 400 km and a regional topographic swell with elevations exceeding 2000 m. The extensive Yellowstone hydrothermal system has a thermal output of 5 GW. The most recent magmatic eruption dated at 70,000 ybp. By comparison, Long Valley caldera is underlain by a relatively modest "hot-spot", the locus of which appears to be influenced by a dilatational jog between the dextral Eastern California Shear Zone and the Walker Lane and westward delamination of the dense lithospheric root of the adjacent Sierra Nevada. The Long Valley system has fed multiple eruptions of over the past 4 my and a single caldera-forming eruption at 760,000 ybp. It is marked by a limited topographic swell but with the elevation of the caldera floor and adjacent basins comparable to the 2000-plus m elevation of the Yellowstone swell. Long Valley caldera hydrothermal system has a thermal output of 0.3 GW (including a 40 MW geothermal power plant). The most recent eruptions from the Long Valley Caldera- Mono Domes volcanic field occurred 600 and 250 ybp. Unrest at both calderas is characterized by strong regional seismicity, recurring intra-caldera earthquake swarms, deformation measured in decimeters, changes in the hydrothermal systems, and elevated CO2 emissions. Extra-caldera seismicity at Yellowstone includes the 1959 M=7.5 Hebgen Lake earthquake, and at Long Valley, four M~6 earthquakes in May 1980 immediately south of the caldera. In both cases, the extra-caldera seismicity is more energetic and persistent than intra-caldera activity underscoring the importance of tectonic-magmatic interactions. Long Valley seismicity includes long-period (LP) "volcanic" earthquakes, spasmodic bursts, and very-long-period (VLP) earthquakes. Seismicity in Yellowstone is limited to brittle-failure earthquakes. Deformation in Yellowstone has shown alternating episodes of inflation and subsidence while that in Long Valley has shown several episodes of rapid uplift with negligible subsidence. In both cases, peak uplifts reached 70 to 80 cm. Among the many outstanding questions for both calderas include 1) the nature of the process driving deformation and whether the active agent involves hydrous magmatic fluids and volatiles or the intrusion of magma into the upper crust, 2) the configuration of the magmatic plumbing system within the crust, and 3) regional tectonic-magmatic interactions.