G42A-01 INVITED 10:25h
Evolution of Deformation Studies on Active Hawaiian Volcanoes
Summarizing 1600 years of observations and interpretations into a brief presentation forces some difficult choices on highlighting the following techniques that are presented chronologically: Visual Observations, 400 AD to present: Missionary William Ellis' Hawaiian Guides told him that Kilauea "had been burning from time immemorial, or, to use their own words, `mai ka po mai', from chaos till now...that in earlier ages it used to boil up, overflow its banks, and inundate the adjacent country...and on occasions they supposed Pele went by a road under ground from her house in the crater to the shore". Observations of the nearly-continuous lava lake in Kilauea Caldera from 1823 until 1924 established that its surface level fluctuated from about 700 to 1100 m above sea level in 10 up-and-down episodes. Tilt Measurements, 1914 to present: Horizontal-seismometer drift and water-tube tiltmeters show that the range of long-term, ground-surface tilt radial to Halemaumau Crater exceeds 500 microradians. Triangulation and Leveling, 1920: R. M. Wilson measured deformation changes related to major Kilauea summit subsidence in 1924. The caldera area around Halemaumau subsided concentrically as much as 4 m relative to the Volcano House benchmark, and triangulation points moved toward Halemaumau by as much as 1.6 m in the caldera area. K. Mogi in 1958 modeled Kilauea leveling data and inferred 3-4 km-deep magma reservoirs. Gravity Measurements, 1959 to present: Changes were first measured during Kilauea summit subsidence related to the lower-east-rift Kapoho eruption. Surveys made before and after the 1975 M7.2 Kalapana Earthquake show that gravity changes are not a simple proxy for elevation changes. Electronic Distance Measurements (EDM), 1964 to present: D. A. Swanson, W. A. Duffield, and R. S. Fiske use EDM for trilateration proving movement of the south flank of Kilauea toward the sea. EDM show displacements as large as 8.7 m of Kilauea's south flank toward the sea related to 1975 M7.2 Earthquake. EDM provide forecasting data on 1984 Mauna Loa eruption, and infer 3-4 km-depth to Mauna Loa magma reservoirs. Global Positioning System (GPS), 1987 to present: Demonstrates changing rate of Kilauea's south flank movement toward the sea. GPS replaces most EDM monitoring of Mauna Loa's recent inflation episode. Satellite Radar Interferometry (InSAR), 1994 to present: NASA/JPL image of Kilauea Caldera area, April to October, 1994. Dilational Strain Measurements, 2000 to present: Early USGS results will be presented at this meeting.
http://www.lava.name/def.htm
G42A-02 10:45h
Recent Inflation of Mauna Loa Volcano, Hawai`i
The inflation of Mauna Loa Volcano that began in May 2002 continues. The inflation rate was highest during the first six months, then relatively low over the next year. Continuously recording GPS (CGPS) stations indicate that extension and uplift rates increased again in 2004. The deformation pattern defined by continuous- and survey-GPS data shows a strong asymmetry of motion about the summit, with greater horizontal velocities on the southeast flank. Vertical velocities, however, are higher in the summit area. Modeling of both the continuous- and survey-GPS data suggests two sources of deformation - inflation of a magma reservoir 4-5 km below the southeastern caldera (between the main topographic caldera and the outer ring faults) and slip on a fault beneath the southeast flank. The start of reinflation in 2002 immediately followed a small swarm of deep (30-50 km) short-period earthquakes in late April to early May. Seismicity was not notable thereafter until early July 2004, when a sustained flurry of deep, long-period seismicity began, averaging one located earthquake per day for the first three weeks, and gradually increasing to about 100 locatable events per week as of this writing. The Hawaiian Volcano Observatory has recently performed upgrades of the seismic network and expanded the deformation monitoring network on Mauna Loa. In early 2002, the deformation network on Mauna Loa consisted of 4 CGPS stations, 3 electronic tiltmeters, and 3 strainmeters. Just those few CGPS stations were enough to alert us, with unprecedented timeliness, of the onset of inflation. Survey GPS measurements of a large array of benchmarks confirmed the renewed inflation and helped characterize the deformation sources. Today, we have 13 CGPS stations, 6 tiltmeters, and 3 strainmeters on Mauna Loa, with another tiltmeter to be installed in October. The main expansion in GPS instrumentation occurred in 2004, with the installation of 6 new instruments. Several new GPS stations are positioned to monitor the potential movement of magma into the northeast and southwest rift zones and into the area of radial vents on the northwest flank. The network is also designed to monitor stress changes in the seismically active Ka`oiki fault zone on the southeast flank.
G42A-03 11:00h
Kinematics of the 1975 Kalapana, Hawaii Earthquake
Earthquakes on Kilauea Volcano can be even more hazardous than volcanic eruptions. The most recent large earthquake was the 1975 Kalapana quake, which generated a locally damaging tsunami, over 8 meters of horizontal displacement and 3 meters of subsidence at the coast. The magnitude estimates for this earthquake range from M$_{s}$7.2 to M$_{w}$7.7. The complexity of the seismic data is matched by the complex surface deformation that was observed with a combination of leveling, tilt and Electronic Distance Measurements (EDM). The geodetic data shows evidence of collapse of the caldera due to magma withdrawal, localized subsidence over the rift zone, shallow slumping along the Hilina Pali, and displacement on the basal detachment fault. We model the summit caldera, rift zones and basal detachment fault as dislocations in an elastic half-space to better quantify the sources of deformation related to this catastrophic event. Using inversion techniques that allow us to let the data constrain the geometry and magnitude of these sources, we find 0.04 km$^{3}$ of magma withdrawal at the summit, between 3-5 m of opening along the rift zone and 7.1 m of slip along the basal detachment at 8.3 km depth. Models that allow for finer spatial resolution of slip on the detachment show that the largest slip occurred west of the hypocenter, along the region of the fault below the coast, and that the majority of the fault slip occurred south of the region of aftershocks and background seismicity. Our best-fitting model has a geodetic moment of 4.1 10$^{20}$ N m (M$_{w}$7.7), which is consistent with tsunami models and recent analysis of long period seismic data. The residual displacements of sites located in the hanging wall of the Hilina Pali slump system, which were not used in the inversion of the other sources, suggest that triggered shallow slumping contributed several meters of horizontal displacement to the coseismic displacement at these sites.
G42A-04 INVITED 11:15h
An Ongoing Episode of Magmatic Inflation at the Three Sisters Volcanic Center, Central Oregon Cascade Range: Inferences from Recent Geodetic and Seismic Observations
Tumescence at the Three Sisters volcanic center began between summer 1996 and summer 1998 and was discovered in April 2001 using interferometric synthetic aperture radar (InSAR). Vertical surface displacement at a maximum rate of 3-4 cm/yr is centered about 5 km west of the summit of South Sister, a composite basaltic-andesite to rhyolite volcano that last erupted between 2,200 and 2,000 years ago. The deforming area is about 20 km in diameter and extends across much of the western part of the Three Sisters Wilderness. The background level of seismic activity is low, suggesting that temperatures in the source region are high enough or the strain rate is low enough to favor plastic deformation over brittle failure. A swarm of over 300 small, shallow (6-10 km) volcano-tectonic earthquakes (Mmax = 1.9) in the northeast quadrant of the deforming area on March 23-25, 2004, was the first notable seismicity in the area for at least two decades. Tilt-leveling and EDM networks were established at South Sister in 1985-1986 and resurveyed in 2001, the latter with GPS. Both networks have since been extended to cover more of the deforming area and 3 continuous GPS stations have been installed. In addition, 1 broadband seismometer and 4 short-period seismometers have been installed since 2001 to improve real-time seismic monitoring and earthquake-location capabilities. InSAR results together with annual GPS and leveling surveys constrain the location, geometry, and inflation rate of the deformation source as a function of time. A best-fit source model derived from simultaneous inversion of all three datasets through 2003 is a shallowly dipping tabular body located 6.5 ± 2.5 km below the surface with a volume increase of 0.0050 ± 0.0015 km3/yr. GPS and leveling data acquired in August 2004 will be used to update the model. The most likely cause of tumescence is a pulse of basaltic magma intruding the upper crust along the brittle-ductile interface - a process that must occur repeatedly beneath the Cascade Range but in the past would have escaped detection in the absence InSAR observations, unless accompanied by unusual seismicity. The likelihood that the current episode at Three Sisters will culminate in an eruption in the near term is judged to be low, but the impact of an eruption could be great. An updated volcano hazards assessment has been completed, concerned agencies and the public have been notified, and an emergency coordination and communication plan is being prepared.
G42A-05 11:35h
A New Method for Measuring Volcanic Deformation Using InSAR Persistent Scatterers
Persistent scatterer (PS) analysis of InSAR data has proven to be a very sensitive technique for measuring steady deformation in urban areas. Standard methods can also treat non-steady deformation if displacements follow a simple parametric function of time. Applying these methods to estimate deformation on volcanoes is, however, more challenging because a) the majority of volcanoes are not urbanized and therefore lack the man-made structures that are recognized by the PS algorithm, and b) deformation tends to proceed at an irregular rate. We present a new method for identifying PS pixels in a series of interferograms, based on a combination of their amplitude and phase characteristics, that is applicable to the study of natural targets. The phase-based method avoids one major problem with the existing algorithm: low amplitude pixels with actual phase stability are not identified. Our method also uses the spatial correlation of the phases rather than a specified phase history so that we can observe temporally-variable processes. The algorithm involves removing a residual topographic component of the phase for each PS, assumed proportional to the interferometric baseline, and then unwrapping the phase of the PS interferogram stack both temporally and spatially. Our technique finds scatterers with stable phase characteristics, even for pixels that do not contain man-made structures. It is applicable to areas where conventional InSAR fails due to complete decorrelation of the majority of scatterers, yet a few stable scatterers may be distributed amongst them. We created and analyzed a stack of 21 interferograms for Long Valley Caldera in California, and identified 23,000 PS pixels in the study region, as opposed to 300 found with Ferretti's (2001) algorithm. The resulting unwrapped phases, when transformed into estimates of line-of-sight displacements, agree with GPS, leveling and EDM measurements made over similar time intervals, validating the technique. Furthermore, the dense spatial coverage of the PS allows us to refine models of the sources of deformation within the caldera.
G42A-06 11:50h
Uplift, Subsidence, and Trapdoor Faulting at Sierra Negra Volcano, Galapagos Islands, from InSAR Observations and Mechanical Modeling
In the last 10 years, Sierra Negra volcano, on the island of Isabella in the Galapagos, has experienced rapid uplift, trapdoor faulting (Amelung and Jonsson et al., Nature 2000), renewed inflation, and subsidence (Geist et al., JVGR in press). Boundary element calculations based on the InSAR observations constrain the magma chamber geometry at Sierra Negra. The surface deformation during the periods of inflation was caused by pressurization of a sill-like intrusion increasing in thickness by a maximum of 0.5 meter (Yun et al., JVGR in press). However, for such a shallow intrusion only the top of the magma chamber can be resolved using surface deformation observations; the data are insensitive to the sides and bottom of the chamber. A simple thermal analysis shows that intrusion must be at least 40 meters thick to remain liquid during the period of observations, so that the magma chamber at Sierra Negra is likely a thick sill or a flat-topped diapir. We model the stress field in the volcano assuming magma chamber geometries and pressure changes found from analysis of the InSAR data. By simulating both the inflation and faulting events, we hope to gain insights into the stress state within the volcano, and the conditions that favor faulting on the intra caldera fault system versus dike propagation and eruption. The stress acting on the pre-existing intra-caldera fault is a combination of pre-inflation, gravitational, and magmatic contributions. We bound the change in excess magma pressure using InSAR observations prior to the trapdoor-faulting event, resulting in a lower bound on the shear stress that triggered the faulting. Our results will have important implications for the stress state within the volcano, the mechanics of induced faulting and dike propagation, and may lead to better forecasts of future behavior.
G42A-07 12:05h
Deformation and Eruption Forecasting at Volcanoes under Retreating Ice Caps: Discriminating Signs of Magma Inflow and Ice Unloading at Grimsvotn and Katla volcanoes, Iceland
Warmer climate is causing retreat of many of the world's ice caps that cover volcanoes. Such ice unloading may influence magma systems and lead to elastic/inelastic earth response such as glacio-isostatic uplift. In Iceland, a number of the most active volcanoes are under retreating ice caps, including the Grimsvotn and Katla volcanoes. Both volcanoes have calderas and shallow magma chambers, and are currently undergoing periods of unrest. At both volcanoes, repeated GPS measurements on nunataks (mountains sticking out of the ice) and around the ice caps, show uplift of about 1-3 cm/year, as well as horizontal displacements. At Katla, optical leveling tilt measurements, InSAR data, and continuous GPS measurements constrain the deformation field outside the ice-covered part of the volcano as well. For both volcanoes, the current uplift rates may eventually conform either to magma inflow or glacio-isostatic rebound. The origin of the deformation can be resolved by considering horizontal displacements and the ratio between horizontal and vertical displacements. Under ice caps and near their edges, the Earth response to ice unloading is mostly vertical, with horizontal displacements an order of magnitude smaller than vertical (ratio $<$ 0.3). This holds true both for immediate elastic response to ice unloading, and even more for the final relaxed state approximated as the response of an elastic plate (brittle part of the crust) underlain by fluid (relaxed ductile part of the crust or a magmatic system). For a magmatic source approximated as a point source of pressure, the ratio between horizontal and vertical displacements is $>$0.5 at distances $>$0.5D, where D is the depth to the source. At Grimsvotn and Katla, the observed ratio is close to 1, suggesting the deformation is mainly caused by magma movements. Horizontal displacements at rates of 1-2 cm/year occur at the caldera boundaries, away from the magmatic sources. In addition to inflow of magma to the volcanoes, both of them have elevated seismicity and geothermal activity. Pressure in a shallow magma reservoir at Grimsvotn is likely to have exceeded pre-eruptive limit for its last eruption in 1998. Katla volcano is also considered to be close to failure, and an eruption of at least one of these volcanoes is anticipated within a few years, as long as magma continues to flow into their shallow magma chambers.