V53A-1131
The Fathers Day Eruption of Kilauea, Volcano, Hawaii: a gas Emissions Perspective
Gas release events from Kilauea's summit magma reservoir and East Rift Zone (ERZ) before and during the June 17, 2007 Fathers Day eruptive sequence provide evidence of correlated summit and rift processes and illustrate how perturbations in magma supply from depth can disrupt steady processes at the surface. Changes in summit CO2 emission rates are thought to reflect changes in deep magma supply and the movement of magma into the rift zone. During the several years preceding the Fathers Day activity, Kilauea had undergone a surge of magma supply and effusive output, as evidenced by gas release and cross-caldera extension. In 2004, the CO2 emission rate, which had been quasi-steady, below 10,000 tonnes/day for the preceeding eight years, began to increase. By mid-2004 the rate had doubled, and by 2005, had tripled, reaching nearly 30,000 t/d. The peak and subsequent decline in summit CO2 attended a similar pattern in east rift SO2 emissions, which are used at Kilauea as one proxy for effusion rate. The surge in magma supply and lava effusion rate may have unsettled Kilauea's plumbing enough to set the stage for the subsequent Fathers Day intrusive and eruptive activity. As magma was withdrawn from the summit reservoir on 17 June 2007, forming the dike that intruded the Upper East Rift Zone, overburden pressure in the reservoir decreased allowing dissolved gases to exsolve and escape from the melt. The more soluble SO2, would be more affected by this pressure change than CO2, which exsolves at much greater depth. Summit SO2 emissions increased nearly four-fold as the summit deflated and summit and Upper East Rift tremor spiked. Increased ambient gas concentrations downwind of Halema`uma`u Crater resulted from the sharp increase in SO2 emission rates and caused the national park to close Crater Rim Drive in this area for several days. The Fathers Day intrusion, subsequent collapses at Pu`u `O`o, and the brief eruption at Kane Nui O Hamo (KNH) profoundly changed ERZ gases as well. SO2 emissions declined dramatically from Pu`u `O`o, dropping below 100 t/d by early July and remained low for several weeks even after lava reappeared in Pu`u `O`o. This overall decline in east rift gas release is reasonably consistent with the observed eruptive events: the dike likely robbed the magma supply but produced only the scanty KNH eruption, and the thinness of the dike geometry proved a poor conduit for SO2 release from depth. The beginning of the 21 July fissure activity reflected a return of magma and significant gas release to the Pu`u `O`o area as pressure within the shallow vent system exceeded structural integrity of the edifice. Following a brief increase in SO2 emissions at the onset of fissure activity, values declined, while effusion remained high. The low level of SO2 release, which persisted for the first several weeks of this fissure activity possibly reflected displacement of old rift-stored magma with newer magma coming from the summit and passing beneath the Pu`u `O`o system. Beginning 12 August, Pu`u `O`o re-emerged as Kilauea's primary SO2 degassing source, even though all extrusion as of this writing is occurring from the fissure system east of the cone. Currently, Kilauea appears to be following a pattern similar to that of the 1986-1991 Kupaianaha era, wherein most SO2 release was from Pu`u `O`o while active extrusion occurred down rift.
V53A-1132
Seismic and Infrasound Recordings from Kilauea Volcano: Volcanic Tremor, Lava Outbreaks, and Fissure Eruptions
The continuous effusion from the Pu‘u ‘O‘o crater complex, the active vent of Kilauea Volcano, Hawaii, produced nearly continuous tremor for years. Recently this tremor was recorded by two infrasound arrays, one at 12.5 km and one at 2.5 km, as well as a broadband seismometer at the closer array. These recordings exhibit significant temporal changes. A sharp, complex spectral peak of ~0.6 Hz is present in nearly the entire dataset, and tends to bifurcate and shift frequency over time. Although the seismic wavefield at Kilauea is complex and path effects appear to play a significant role, this spectral peak is also weakly manifested in the seismic recordings. Array processing of the infrasonic data reveals an abundance of broadband signal as well. Most of the signal appears to originate from the main crater region. However, the 2.5 km array detected the presence of a skylight with growing hornitos ~400 m south of Pu‘u ‘O‘o on the active lava tube system. On June 19th, 2007, the magmatic system at Pu‘u ‘O‘o changed. An intrusion of magma reached the surface 6 km west of the crater complex. The timing and location of the lava outbreak were determined acoustically using array processing. Two distinct acoustic pulses were recorded from the correct azimuth, both exhibiting harmonics. The 7/21 fissure eruption also produced clear infrasound signals. The onset of the fissure eruption east of P‘u‘ ‘O‘o was apparent beginning around midnight on 7/21 and was focused between ~1.5-5 Hz. Although the fissure eruption continued to produce infrasound, the character of the recorded signal changes over time. A third infrasound array was placed closer to P‘u‘ ‘O‘o and the fissure to help further constrain the eruption. More detailed results on acoustic signals from the Father's Day Intrusion and Fissure eruption will be presented.
V53A-1133
A Kknematic model for ground deformation at Kilauea volcano during 1992-2005
Kilauea volcano is actively deforming at rates of up to 10 cm/yr. Existing models for Kilauea well explain the ground deformation of the east rift zone and along the southflank. InSAR data for the 1992-2005 period (derived from Jers and Radarsat satellites) clearly show significant subsidence at rates of 5 cm/yr and more in the southwest rift zone. We present a kinematic model to explain the observed deformation consisting of deep rift intrusion and gravitationally-driven normal faulting in the southwest rift zone area.
V53A-1134
Modeling Deformation Sources From 2005-2007 at Kilauea Volcano, Hawaii Using InSAR
The Hawaiian Islands are home to some of the most active volcanoes in the world resulting in a constant state of deformation. After 2005, deformation at the summit caldera of Kilauea changed from deflation to inflation before returning to the typically observed deflation. In 2007, new fissure eruptions occurred in the East Rift Zone as a result of dike intrusions. Interferometric synthetic aperture radar (InSAR) data from Radarsat and Envisat are used to generate source models for the deformation, and, using the small baseline subset (SBAS) algorithm, InSAR time series are created for this period. The results allow for analysis of the deformation patterns both spatially and temporally to better understand the dynamics of the magmatic system.
V53A-1135
Mechanical modeling of the Father's day dike intrusion along the east rift zone of Kilauea, Hawaii, constrained from ALOS PALSAR interferometry
On June 17, 2007 a sudden increase in seismicity and rapid changes in surface deformation were detected
along the east rift zone of Kilauea volcano. This seismicity was associated with a dike intrusion that aligned with
the east rift zone and opened ~2 meters close to Makaopuhi crater by June 20, 2007. The surface
deformation caused by the dike intrusion event was captured by ALOS PALSAR interferograms. ALOS uses an L-
band signal whose wavelength is about 4 times longer than the C-band, and thus better penetrates dense
vegetation, resulting in better interferometric coherence. We formed interferograms from ascending and
descending orbits spanning 2007/05/05 – 2007/06/20 and 2007/02/28 – 2007/07/16, respectively. The
interferograms indicate maximum line-of-sight shortening (motion towards the satellite) of ~36 cm
(ascending) and ~86 cm (descending). Range increase in the summit region, suggesting summit
deflation, is ~11 cm (ascending) and ~9 cm (descending). These interferograms were used to model
the approximate dimensions of the dike and the deflation source at the summit. The modeled dike is sub parallel
to the east rift zone of Kilauea, dips about 81 to 85 degrees to the south, and has a volume increase of
~33×106 m
V53A-1136
Vector Deformation Maps of the Father's Day Intrusion at Kilauea: Constraints on Magma Injection
Between June 17 and June 20, 2007, the East Rift of Kilauea Volcano opened more than 1.9 meters. Here we report on two interferograms - ascending and descending - that were collected by the L-band synthetic aperture radar aboard the ALOS spacecraft that is operated by the Japanese Space Agency - JAXA. We find that the L- band interferograms retain coherence better than interferograms made from C-band instruments. The interferometric phase is completely unwrapped with a single seed point. In addition to the two line-of-sight (LOS) components, a third component (azimuthal offsets) can be derived from the reference and repeat images. These three components are nearly orthogonal and thus provide the full vector displacement for this event. Comparisons with continuous GPS data show that the two LOS components have rms precisions of 2.8 cm while the third azimuth component has an rms precision of 5.8 cm. This third component reveals a peak surface separation across the rift zone of 1.86 m, which is not fully captured (0.90 m) by the widely-spaced GPS measurements. These near-field InSAR data will be used, together with the more precise far-field GPS measurements, to constrain the models of dike opening and caldera deflation associated with this event. ALOS continues to monitor Kilauea with new acquisitions approximately every 23 days. http://topex.ucsd.edu/kilauea
V53A-1137
Deformation Associated With the July 21 Fissure Eruption at Kilauea Volcano, Hawai`i
Deformation measurements at the Pu`u `O`o cone on Kilauea volcano indicate that the vent is underlain by a shallow magma storage reservoir fed by magma transported from Kilauea's summit. The Pu`u `O`o reservoir was drained during the "Father's&pDay" intrusion of June 17-19, 2007, causing Pu`u `O`o's floor to collapse. Following the intrusion, the reservoir gradually refilled and lava reappeared on July 1-2. A lava lake grew in the crater during July 2-20, steadily raising the elevation of the crater floor. The crater interior and adjoining walls began to uplift on July 10, and 2 days later lava began to vent above the level of the lava lake along the margins of the crater. The number of crater margin vents and the magnitude of their activity increased until July 21, by which time crater uplift amounted to about 8 meters. Early that morning, the lava pond at Pu`u `O`o drained suddenly and an eruptive fissure opened on the east flank of the cone. The fissure propagated 2 km downrift, and within a few hours the eruption had localized on three fissure segments between 1 and 2 km east of Pu`u `O`o. Lava erupting from this series of vents formed a system of perched lava ponds feeding long `a`a flows. Deformation associated with the July 21 fissure was exceptionally well-documented by a borehole tiltmeter on the north flank of Pu`u `O`o and 7 continuous GPS stations within 2 km of the cone (the fissure propagated between two of these GPS stations). The time series of geodetic measurements suggests that deflation of Pu`u `O`o began at about 2250 HST (Hawaiian Standard Time - UTC minus 10 hours) on July 20, while a camera looking into Pu`u `O`o crater recorded draining of the lava lake starting at around 2355 HST. Deformation associated with opening of the fissure was apparent in the tilt and GPS records by 0012 HST, with the camera showing glow in the direction of the eruption site by 0039 HST on July 21. Localized deformation from InSAR suggests that the fissure has a shallow source, probably within 1 km of the surface; this interpretation is also supported by kinematic GPS results collected from points around Pu`u `O`o in July 2006 and July 2007.
V53A-1138
Source Models of the June 17th, 2007 Kilauea Intrusion: Spatio-Temporal Evolution
The June 17, 2007 intrusion in the upper East Rift Zone (ERZ) of Kilauea volcano, Hawaii was particularly well monitored with continuous GPS, tilt, and InSAR. The first indication of activity was increased seismicity from Kilauea's summit to the bend in the ERZ. The upper ERZ tilted ~70 μ rad sharply down to the south from 02:16 to 07:40 HST, and within minutes the summit crossing GPS baseline began to shorten and eventually decreased by ~14 cm over 2.5 days. Beginning between 08:00 and 0:900, the rift-spanning GPS baseline down rift lengthened by ~1 meter in just over two days. In the two following days, surface cracks were observed in the upper ERZ, and a small amount of new lava was seen on the northeast side of Kane Nui o Hamo. We test distributed source models with a variety of dike geometries that follow observed surface cracks, the axis of InSAR fringes, and the optimal uniform opening model (Sinnett et al., this volume). The sources included are the ERZ dike, the south flank decollement, and the summit magma reservoir. The dike and decollement are modeled as distributed rectangular dislocations, and the summit magma reservoir with a Mogi source. While several models explain most of the data, none fits the details of the data near the western end of the dike, perhaps indicating a complex source geometry, inelastic deformation, or differing mechanisms of pre-, co- and post- intrusion. In the preferred model, following the axis of deformation observed by InSAR, the maximum opening of 2.33 m is between 0-2 km deep under Kane Nui o Hamo, just north of Makaopui crater. Another opening maximum of nearly 2 m occurs about 2 km deep between Pauahi and Mauna Ulu. Seismicity during the intrusion was concentrated below the opening maxima, with some events between them. Unlike previous ERZ intrusions (e.g. Jan. 1997, Owen et al., GRL, 2000) the dike opening does not account for all of the deformation observed at the GPS sites on the southwest flank, nor does it agree with the tilt direction at Kaena Point, indicating possible decollement slip during the intrusion (B. Brooks, pers. comm.). Inversions including a decollement favor slip of up to 30 cm and improve data fits locally, especially in the western part of the network, although the overall amount of data variance explained (~ %70) is similar to the dike-only models. We use the Kalman filter-based Extended Network Inversion Filter (McGuire and Segall, GJI, 2003) to invert kinematic GPS solutions that have been smoothed to reduce multipath (Larson et al., JGR, 2001) and sampled every 4 minutes, and tilt, sampled every 1 minute, for the spatio-temporal slip evolution. The source parameters are the same as the preferred distributed model. Because the dike is short relative to station spacing, lateral resolution of propagation is limited, however, refined models my be able to determine the relative timing of the intrusion and slow slip.
V53A-1139
Source Models of the June 17th, 2007 Kilauea Intrusion: Monte Carlo Optimization
Father's Day, 17 June 2007, marked the beginning of the 56th episode of the ongoing eruption of Kilauea volcano, Hawaii. The episode culminated in a short-lived eruption approximately 6 km west of Pu\`{}u \`{}O\`{}o and 13 km southeast of Kilauea summit. The interruption of magma supply to, and withdrawal from, the reservoir beneath Pu\`{}u \`{}O\`{}o caused cessation of activity and ~100 m of crater floor subsidence there. The continuous and campaign GPS, electronic tiltmeter, and seismic networks, as well as InSAR captured the episode in fine detail. Visual inspection of the data show subsidence at Kilauea summit and Pu\`{}u \`{}O\`{}o, which fed the inflating dike. We began by modeling the intrusion with a Mogi source beneath Kilauea summit and a dislocation with uniform opening beneath the east rift zone embedded in an isotropic, homogenous, elastic, half space. We invert for the 12 source parameters (length, width, depth, dip, strike, horizontal position, and opening of the dike, and position, depth, and volume change of the Mogi source) using Monte Carlo optimization. The inversion used three component displacement data from 23 continuous and campaign GPS stations, diurnally and tidally filtered tilt from 6 stations, and an ENVISAT InSAR interferogram spanning 04/12/07 to 06/21/07 decimated using a quadtree algorithm. The optimum model included ~-4.1 * 106 m3 of volume loss from a reservoir 3 km beneath the summit, and a total dike volume of ~19*106 m3 (~4.84 km length x 2.45 km width x 1.6 m opening at 2.4 km depth). The discrepancy between summit volume loss and total dike volume suggests that other sources must have fed the dike. A crude estimate of volume loss from Pu\`{}u \`{}O\`{}o is 8.5*106 m3 accounting for ~ 66% of the volume of the dike. The eruption site lies inside the eastern edge of the model, and ~0.5 km to the south of the best fit dike top. The best fit dike top parallels the northern margin of an area of ground cracking near Makaopuhui and terminates at its western margin near Mauna Ulu. The western termination is ~2.5 km east of the westernmost observed ground cracks. Within 95% bounds the dike top may intersect the eruption area and extend to all regions of ground cracking. It is also interesting to note that this dike is located in an area between the 1997 and 1999 intrusions. The best fit single dislocation model explains only 35% of the variance in the data. This is in part due to the inadequacies of a single planar dike with uniform opening to explain surface deformation and perhaps to inelastic deformation associated with ground cracking near the western edge of the dike. Models with distributed opening, in which the dike plane honors the optimization results as well as the region of decorrelation in the ENVISAT interferogram, explain 69% of the data (Montgomery-Brown et al., this session).