South Flank Of Kilauea Volcano Research Articles

Aseismic deformation across the Hilina fault system, Hawaii, revealed by wavelet analysis of InSAR and GPS time series

The Hilina Fault System (HFS) is located on the south flank of Kilauea volcano and is thought to represent the surface expression of an unstable edifice sector that is active during seismic events such as the 1975 Kalapana earthquake. Despite its potential for hazardous landsliding and associated tsunamis, no fault activity has yet been detected by means of modern geodetic methods, since the 1975 earthquake. We present evidence from individual SAR interferograms, as well as cluster analysis and wavelet analysis of GPS and InSAR time series, which suggest an inferred differential motion at HFS. To investigate the effect of atmospheric delay on the observed differential motion, we implement a statistical approach using wavelet transforms. We jointly analyze InSAR and continuous GPS deformation data from 2003 to 2010, to estimate the likelihood that the subtle time-dependent deformation signal about the HFS scarps is not associated with the atmospheric delay. This integrated analysis reveals localized deformation components in the InSAR deformation time series that are superimposed on the coherent motion of Kilaueaʼs south flank. The statistical test suggests that at 95% confidence level, the identified differential deformation at HFS is not due to atmospheric artifacts. Since no significant shallow seismicity is observed over the study period, we suggest that this deformation occurred aseismically.

High‐resolution locations of triggered earthquakes and tomographic imaging of Kilauea Volcano's south flank

The spatiotemporal patterns of seismicity beneath Kilauea's south flank give insight to the structure and geometry of the decollement on which large, tsunamigenic earthquakes have occurred, and its relation to slow slip events (SSEs), which have been observed every 1 to 2 years since 1997. In order to record earthquakes triggered by a SSE that was predicted to occur in March 2007, a temporary network of 20 seismometers was deployed on Kilauea's south flank, termed the SEQ network. While the SSE did not occur until 17 June 2007, theSEQ network recorded over 3000 earthquakes, including those triggered by the SSE. We relocate hypocenters of volcano‐tectonic earthquakes and invert for P and S wave velocity structure using waveform cross‐correlation and double‐difference tomography using data from the SEQ network and the permanent Hawaii Volcano Observatory network (HVO) data, with additional data from other previous temporary arrays. The best‐constrained hypocenters, recorded by both the SEQ and HVO networks, indicate the decollement as a subhorizontal layer of seismicity at 8 km depth less than 1 km thick in most areas, with the western portion of the decollement dipping to the southeast. The seismicity triggered by the June 2007 SSE includes over 400 earthquakes overlapping with the southern edge of the decollement seismicity. A shallower swarm of earthquakes also occurred between 2 and 7 km depth in April 2007 near Apua Point, and may have been indirectly triggered by the Mw 8.1 Solomon Islands earthquake at ∼6000 km distance, which occurred 48 h prior to the beginning of the swarm.

Slow Slip Event at Kilauea Volcano

Early in the morning of 1 February 2010 (UTC; early afternoon 31 January 2010 local time), continuous Global Positioning System (GPS) and tilt instruments detected a slow slip event (SSE) on the south flank of Kilauea volcano, Hawaii. The SSE lasted at least 36 hours and resulted in a maximum of about 3 centimeters of seaward displacement. About 10 hours after the start of the slip, a flurry of small earthquakes began (Figure 1) in an area of the south flank recognized as having been seismically active during past SSEs [Wolfe et al., 2007], suggesting that the February earthquakes were triggered by stress associated with slip [Segall et al., 2006].

Kilauea slow slip events: Identification, source inversions, and relation to seismicity

Several slow slip events beneath the south flank of Kilauea Volcano, Hawaii, have been inferred from transient displacements in daily GPS positions. To search for smaller events that may be close to the noise level in the GPS time series, we compare displacement fields on Kilauea's south flank with displacement patterns in previously identified slow slip events. Matching displacement patterns are found for several new candidate events, although displacements are much smaller than previously identified events. One of the candidates, 29 May 2000, is coincident with a microearthquake swarm, as are all of the previously identified slow slip events. The microearthquakes follow the onset of slow slip, implying that they are triggered by stress changes during slip. The new slow slip event brings the total number of events on Kilauea, between 1997 and 2007, to eight, the smallest having MW = 5.3, and the largest having MW = 6.0. While the recurrence time between the four largest events is 2.11 ± 0.01 years, the repeat time for all eight events is 0.9 ± 0.6 years. We invert for the fault geometry and distribution of slip during the slow slip events. The optimal source depths of 5 km, assuming uniform slip dislocations in an elastic half‐space, are considerably shallower than the accompanying swarm earthquakes (6.5–8.5 km), which would place the earthquakes in a zone of decreased Coulomb stress. Inversions including the effects of topography and layered elastic structure in the forward models favor depths comparable to microearthquake depths, such that the earthquakes are located in a region of increased Coulomb stress. We also invert for time‐dependent fault slip directly from the 30 s GPS phase observations, constraining the source to the optimal uniform slip geometry. On the basis of these inversions, the larger events last between 1.5–2.2 days. The data are unable to resolve migration of slip along the fault. The temporal pattern of accompanying microearthquakes is consistent with the fault slip history assuming a seismicity rate theory based on rate and state‐friction, making the swarm earthquakes coshocks and aftershocks of the slow slip events.

Growth of the volcano-flank Koa’e fault system, Hawaii

Structural mapping of the Koa’e fault system, located on the south flank of Kilauea Volcano, Hawaii, has been carried out to study the relationship between fault propagation and re-surfacing by lava flows. The 1.9-km long White Rabbit Fault faces uphill towards the summit of the Kilauea, and has produced c. 8 m vertical offset of a young lava flow (500–750 years). The fault exhibits multiple peaks in its throw-distance profile, suggesting recent linkage between at least 2 separate faults at the surface. However, the width of a monocline associated with the fault shows a single maximum of c. 40 m, located near the centre of the overall fault trace. We suggest monocline width is related to heave across the fault at depth (a minimum of c. 40 m) beneath the most recent lava that re-surfaced the fault. The surface monocline records re-emergence of a fault that had previously propagated to the surface prior to recent lava deposition rather than upward propagation of a new fault. Overall, the fault displays a growth geometry with displacements increasing with depth. The above implies that the Koa’e faults pre-date the lavas exposed at the surface on the volcano flank, and are a long-lived feature of the volcano dynamics, despite the small offsets across the faults at the surface.

Microearthquake streaks and seismicity triggered by slow earthquakes on the mobile south flank of Kilauea Volcano, Hawai'i

We perform waveform cross correlation and high precision relocation of both background seismicity and seismicity triggered by periodic slow earthquakes at Kilauea Volcano's mobile south flank. We demonstrate that the triggered seismicity dominantly occurs on several preexisting fault zones at the Hilina region. Regardless of the velocity model employed, the relocated earthquake epicenters and triggered seismicity localize onto distinct fault zones that form streaks aligned with the slow earthquake surface displacements determined from GPS. Due to the unknown effects of velocity heterogeneity and nonideal station coverage, our relocation analyses cannot distinguish whether some of these fault zones occur within the volcanic crust at shallow depths or whether all occur on the decollement between the volcano and preexisting oceanic crust at depths of ∼8 km. Nonetheless, these Hilina fault zones consistently respond to stress perturbations from nearby slow earthquakes.

Argon geochronology of Kilauea's early submarine history

Submarine alkalic and transitional basalts collected by submersible along Kilauea volcano's south flank represent early eruptive products from Earth's most active volcano. Strongly alkalic basalt fragments sampled from volcaniclastic deposits below the mid-slope Hilina Bench yield 40Ar/ 39Ar ages from 212 ± 38 to 280 ± 20 ka. These ages are similar to high-precision 234 ± 9 and 239 ± 10 ka phlogopite ages from nephelinite clasts in the same deposits. Above the mid-slope bench, two intact alkalic to transitional pillow lava sequences protrude through the younger sediment apron. Samples collected from a weakly alkalic basalt section yield 138 ± 30 to 166 ± 26 ka ages and others from a transitional basalt section yield 138 ± 115 and 228 ± 114 ka ages. The ages are incompatible with previous unspiked K–Ar studies from samples in deep drill holes along the east rift of Kilauea. The submarine birth of Kīlauea volcano is estimated at < 300 ka. If the weakly alkalic sequence we dated is representative of the volcano as a whole, the transition from alkalic to tholeiitic basalt compositions is dated at ∼ 150 ka.

Seismic Velocity and Attenuation Structure of the East Rift Zone and South Flank of Kilauea Volcano, Hawaii

It has been proposed that a deep magma body beneath the east rift zone (ERZ) of Kilauea volcano, Hawaii, must be present in order to explain the observed deformation of the south flank. From November 1999 to June 2000, 29 IRIS-PASSCAL three-component seismographs were operated across Kilauea9s ERZ and south flank. Using local earthquakes recorded by these stations, supplemented with data from the Hawaiian Volcano Observatory stations, we investigated the deep structure of the ERZ and south flank through two approaches: seismic velocity tomography and seismic attenuation tomography. Results along two north-northwest-south-southeast profiles are examined. The western profile shows a low-velocity region beneath and south of the Hilina Pali, and earthquakes at depth are sandwiched between a high V p / V s , high Q zone and a low V p / V s , low Q zone. The velocity and attenuation differences are interpreted to reflect the contrast in material properties between the overlying volcanic pile and the underlying fluid-rich ocean sediments and fractured, altered ocean crust beneath the south flank thrust fault. The eastern profile shows an anomalous feature with low V p , low V p / V s , high Q p , and low Q s at about 7 km depth beneath the ERZ. The low V p / V s ratio is inconsistent with the presence of partial melt. Instead, this anomaly is attributed to a trapped reservoir of CO 2 . Our results do not support the existence of a sizable molten or partially molten body at depth beneath the ERZ. If the deep magma body exists, it must be relatively small. Based on synthetic resolution tests, an estimated limit on the maximum allowable cross-sectional size of the deep magma body is about 2 by 4 km.

Long-Period Source Characteristics of the 1975 Kalapana, Hawaii, Earthquake

The 1975 Kalapana, Hawaii, earthquake occurred under the highly mobile south flank of Kilauea Volcano. It has been interpreted variously as a normal-faulting earthquake, a thrust-faulting earthquake, and a landslide. Primary evidence for the landslide model has been the failure of previous faulting models to explain the observed Love-wave radiation pattern and the tsunami amplitudes generated by the Kalapana event. Here, we present a reanalysis of the long-period, digital seismic data for this event. Centroid-moment-tensor analysis shows that the seismic radiation pattern can be explained well by thrust faulting on a plane dipping shallowly landward. The seismic moment of 3.8 × 10 20 N m ( M W 7.7) that we determine is approximately twice as large as earlier estimates. The geometry and seismic moment of the focal mechanism determined here are consistent with the observed tsunami amplitudes. Inversion of long-period body-wave waveforms shows that the earthquake source duration (∼72 sec) is unusually long for an earthquake of this size.

Slope failure and volcanic spreading along the submarine south flank of Kilauea volcano, Hawaii

New multichannel reflection data and high‐resolution bathymetry over the submarine slopes of Kilauea volcano provide evidence for current and prior landsliding, suggesting a dynamic interplay among slope failure, regrowth, and volcanic spreading. Disrupted strata along the upper reaches of Kilauea's flank denote a coherent slump, correlated with the active Hilina slump. The slump comprises mostly slope sediments, underlain by a detachment 3–5 km deep. Extension and subsidence along the upper flank is compensated by uplift and folding of the slump toe, which surfaces about midway down the submarine flank. Uplift of strata forming Papa`u seamount and offset of surface features along the western boundary of Kilauea indicate that the slump has been displaced ∼3 km in a south‐southeast direction. This trajectory matches coseismic and continuous ground displacements for the Hilina slump block on land, and contrasts with the southeast vergence of the rest of the creeping south flank. To the northeast, slope sediments are thinned and disrupted within a recessed region of the central flank, demonstrating catastrophic slope failure in the recent past. Debris from the collapsed flank was shed into the moat in front of Kilauea, building an extensive apron. Seaward sliding of Kilauea's flank offscraped these deposits to build an extensive frontal bench. A broad basin formed behind the bench and above the embayed flank. Uplift and back tilting of young basin fill indicate recent, and possibly ongoing, bench growth. The Hilina slump now impinges upon the frontal bench; this buttress may tend to reduce the likelihood of future catastrophic detachment.

New insights into Kilauea's volcano dynamics brought by large‐scale relative relocation of microearthquakes

We investigated the microseismicity recorded in an active volcano to infer information concerning the volcano structure and long‐term dynamics, by using relative relocations and focal mechanisms of microearthquakes. There were 32,000 earthquakes of the Mauna Loa and Kilauea volcanoes recorded by more than eight stations of the Hawaiian Volcano Observatory seismic network between 1988 and 1999. We studied 17,000 of these events and relocated more than 70%, with an accuracy ranging from 10 to 500 m. About 75% of these relocated events are located in the vicinity of subhorizontal decollement planes, at a depth of 8–11 km. However, the striking features revealed by these relocation results are steep southeast dipping fault planes working as reverse faults, clearly located below the decollement plane and which intersect it. If this decollement plane coincides with the pre‐Mauna Loa seafloor, as hypothesized by numerous authors, such reverse faults rupture the pre‐Mauna Loa oceanic crust. The weight of the volcano and pressure in the magma storage system are possible causes of these ruptures, fully compatible with the local stress tensor computed by Gillard et al. [1996]. Reverse faults are suspected of producing scarps revealed by kilometer‐long horizontal slip‐perpendicular lineations along the decollement surface and therefore large‐scale roughness, asperities, and normal stress variations. These are capable of generating stick‐slip, large–magnitude earthquakes, the spatial microseismic pattern observed in the south flank of Kilauea volcano, and Hilina‐type instabilities. Rupture intersecting the decollement surface, causing its large‐scale roughness, may be an important parameter controlling the growth of Hawaiian volcanoes.

Shallow Normal Faulting and Block Rotation Associated with the 1975 Kalapana Earthquake, Kilauea Volcano, Hawaii

The Hilina fault system is a set of normal faults that accommodate extension of the mobile south flank of Kilauea Volcano. Large earthquakes ( M ≥6) and aseismic fault slip transport the flank southeastward along a basal detachment at ∼8–10 km depth. Both the 1975 M 7.2 Kalapana and the 1868 M 7.9 Great Kau earthquakes produced slip on the Hilina faults. We compare Kalapana earthquake fault offsets, ground displacements derived from analysis of geodetic surveys, and model displacements from a dislocation model to evaluate whether the central Hilina fault slip associated with the Kalapana earthquake was due to (1) shallow normal faulting independent of basal detachment slip or (2) deep normal faulting directly linked to basal detachment slip. Our analysis shows that observed site motions at the coast are significantly greater than model displacements expected from a dislocation model of basal detachment slip alone. To explain ground displacements and fault offsets, we require fault slip on shallow normal faults (as deep as 2–3 km) triggered by slip on the basal detachment. Leveling data along the Chain of Craters Road and vertical fault offsets across the central Hilina fault system suggest that the Kalapana earthquake produced block rotation of the hanging-wall blocks.

Faulting in the South Flank of Kilauea Volcano, Hawai'i

Models developed to explain the deformation of Kilauea volcano largely rely on seismic and geodetic data that cover only a short period of activity. Kilauea's active South Flank contains two normal fault systems — the Koa'e and Hilina systems — that must reflect longer-term deformation of the edifice. The scale and origin of these faults is poorly understood. Both fault systems consist of en echelon fault segments that have significant interaction and linkage. The presence of relay ramps between fault segments, systematic stepping senses, and slip and displacement data all suggest that the fault systems link down to individual larger faults. The Hilina Fault System (HFS) is ≥42 km long and ∼9 km deep, while the Koa'e Fault System (KFS) is underlain by two to three smaller faults, the largest of which is at least 14 km long and 4 km deep (and possibly extends down to the basal thrust at 9 km depth). We suggest that the HFS has developed as a result of ‘differential slip’ on the thrust at the base of the edifice and that footwall uplift on the HFS causes antithetic faulting in the South Flank, which provides a new explanation of the origin of the KFS.

Anomalously high b-values in the South Flank of Kilauea volcano, Hawaii: evidence for the distribution of magma below Kilauea's East rift zone

The pattern of b-value of the frequency–magnitude relation, or mean magnitude, varies little in the Kaoiki-Hilea area of Hawaii, and the b-values are normal, with b=0.8 in the top 10 km and somewhat lower values below that depth. We interpret the Kaoiki-Hilea area as relatively stable, normal Hawaiian crust. In contrast, the b-values beneath Kilauea's South Flank are anomalously high ( b=1.3–1.7) at depths between 4 and 8 km, with the highest values near the East Rift zone, but extending 5–8 km away from the rift. Also, the anomalously high b-values vary along strike, parallel to the rift zone. The highest b-values are observed near Hiiaka and Pauahi craters at the bend in the rift, the next highest are near Makaopuhi and also near Puu Kaliu. The mildest anomalies occur adjacent to the central section of the rift. The locations of the three major and two minor b-value anomalies correspond to places where shallow magma reservoirs have been proposed based on analyses of seismicity, geodetic data and differentiated lava chemistry. The existence of the magma reservoirs is also supported by magnetic anomalies, which may be areas of dike concentration, and self-potential anomalies, which are areas of thermal upwelling above a hot source. The simplest explanation of these anomalously high b-values is that they are due to the presence of active magma bodies beneath the East Rift zone at depths down to 8 km. In other volcanoes, anomalously high b-values correlate with volumes adjacent to active magma chambers. This supports a model of a magma body beneath the East Rift zone, which may widen and thin along strike, and which may reach 8 km depth and extend from Kilauea's summit to a distance of at least 40 km down rift. The anomalously high b-values at the center of the South Flank, several kilometers away from the rift, may be explained by unusually high pore pressure throughout the South Flank, or by anomalously strong heterogeneity due to extensive cracking, or by both phenomena. The major b-value anomalies are located SSE of their parent reservoirs, in the direction of motion of the flank, suggesting that magma reservoirs leave an imprint in the mobile flank. We hypothesize that the extensive cracking may have been acquired when the anomalous parts of the South Flank, now several kilometers distant from the rift zone, were generated at the rift zone near persistent reservoirs. Since their generation, these volumes may have moved seaward, away from the rift, but earthquakes occurring in them still use the preexisting complex crack distribution. Along the decollement plane at 10 km depth, the b-values are exceptionally low ( b=0.5), suggesting faulting in a more homogeneous medium.

Overthrusting and sediment accretion along Kilauea's mobile south flank, Hawaii: Evidence for volcanic spreading from marine seismic reflection data

A large, intermittently active, submarine landslide known as the Hilina slump has been interpreted along the south flank of Kilauea volcano. Seaward-dipping faults on land mark its headwall, and an offshore bench may define its uplifted toe. Geodetic data show that the entire south flank is also moving seaward, by a process referred to as volcanic spreading; this provides an alternative explanation for the bench, i.e., overthrusting along the edge of the sliding flank. The latter interpretation is consistent with new seismic reflection data across the submarine flank. A prominent reflection near the top of the oceanic plate suggests the decollement upon which the mobile flank slides. Landward-dipping reflections rise from this horizon and bound packages of bedded strata faulted and imbricated within the bench. The absence of correlative seaward-dipping faults and rotated strata on the upper flank suggests that the bench is not coupled to a slump. Moreover, kinematic reconstructions of the bench indicate that it has accommodated 15–24 km of displacement. This value is consistent with estimates for rift-zone extension but too high for shortening at the toe of a slump. We interpret the bench to result from overthrusting and accretion of volcaniclastic sediments to the edge of the mobile flank, and suggest that morphologic benches develop preferentially where landslide debris has accumulated near the base of the volcano and can be accreted to its sliding edifice.

Paleomagnetic constraints on fault motion in the Hilina Fault System, south flank of Kilauea Volcano, Hawaii

Movement of the south flank of Kilauea Volcano in Hawaii has been associated with catastrophic landslide events. The surface expression of this former movement is the Hilina Fault System with fault scarps as high as 500 m. Paleomagnetic directions for lava flows exposed in the Hilina Fault scarps at Puu Kapukapu and Keana Bihopa on the Hilina Pali are used to determine the average rate of movement along faults (slip surfaces) separating the two sections. This paper reports results from two independent paleomagnetic studies within the Hilina Pali area. Twenty-one paleomagnetic sites (143 cores) were sampled by the Michigan Technological University group from lava flows between the Mo’o Ash and Middle Pohakaa Ash at Keana Bihopa in the footwall block of the 500-m-high Hilina Pali fault scarp. Thirty paleomagnetic sites (152 cores) were collected by the California Institute of Technology group from lava flows between the Mo’o Ash and Middle Pohakaa Ash in the 300-m-high Puu Kapukapu section (the hanging-wall block). A comparison of site-mean directions show that lava flows in the lower part of the Puu Kapukapu section have been tilted more than lava flows in the upper part with respect to the Keana Bihopa section. The systematic steepening of remanent directions downsection at Puu Kapukapu indicates that slippage of this block occurs along listric normal faults. The average amount of backward tilt of the Puu Kapukapu block, based on a comparison of mean directions from the two sections, is 7.8°±7.7°. Using slope stability methods, the average rate of movement of the Puu Kapukapu block since deposition of the Middle Pohakaa Ash is 1.7–2.4 cm/yr, and the average displacement (subsidence) is 680–740 m. Assuming that the average displacement resulted from a series of earthquakes producing subsidence equal to that observed in the 1975 Kalapana earthquake (3 m of subsidence along a 40-km segment of coastline on the south flank of Kilauea Volcano), one Kalapana-size earthquake occurring every 200 years would account for this displacement. Lastly, overall mean directions for the two sections indicate that Puu Kapukapu has rotated counterclockwise with respect to the Hilina Pali by 14.8°±8.5° about a nearby vertical axis. This also suggests that slippage between the two blocks occurs along listric normal faults.

Mechanism of the 1975 Kalapana, Hawaii, earthquake inferred from tsunami data

We investigated the source mechanism of the 1975 Kalapana, Hawaii, earthquake (MS= 7.2) by modeling the tsunamis observed at three tide‐gauge stations, Hilo, Kahului, and Honolulu. We computed synthetic tsunamis for various fault models. The arrival times and the amplitudes of the synthetic tsunamis computed for Ando's fault model (fault length = 40 km, fault width = 20 km, strike = N70°E, dip = 20°SE, rake = −90°, fault depth = 10 km, and slip = 5.6 m) are ∼10 min earlier and 5 times smaller than those of the observed, respectively. We tested fault models with different dip angles and depths. Models with a northwest dip direction yield larger tsunami amplitudes than those with a southeast dip direction. Models with shallower fault depths produce later first arrivals than deeper models. We also considered the effects of the Hilina fault system, but its contribution to tsunami excitation is insignificant. This suggests that another mechanism is required to explain the tsunamis. One plausible model is a propagating slump model with a 1 m subsidence along the coast and a l m uplift offshore. This model can explain the arrival times and the amplitudes of the observed tsunamis satisfactorily. An alternative model is a wider fault model that dips 10°NW, with its fault plane extending 25 km offshore, well beyond the aftershock area of the Kalapana earthquake. These two models produce a similar uplift pattern offshore and, kinematically, are indistinguishable as far as tsunami excitation is concerned. The total volume of displaced water is estimated to be ∼2.5 km3. From the comparison of slump model and the single‐force model suggested earlier from seismological data we prefer a combination of faulting and large‐scale slumping on the south flank of Kilauea volcano as the most appropriate model for the 1975 Kalapana earthquake. Two basic mechanisms have been presented for explaining the deformation of the south flank of Kilauea: (1) pressure and density variation along the rift zone caused by magma injection and (2) gravitational instability due to the steep topography of the south flank of Kilauea. In either mechanism, large displacements on the south flank are involved that are responsible for the observed large tsunamis.

Type of faulting and orientation of stress and strain as a function of space and time in Kilauea's south flank, Hawaii

Earthquake focal mechanisms of events occurring between 1972 and 1992 in the south flank of Kilauea volcano, Hawaii, are used to infer the state of stress and strain as a function of time and space. We have determined 870 fault plane solutions from P wave first motion polarities for events with magnitudes ML ≥ 2.5 and depth ranging between 6 and 12 km. Faulting is characterized by a mixture of decollement, reverse, and normal faults. Most large earthquakes with magnitude M &lt; 7 slip on reverse faults striking NE at 40° and dipping SE between 60° and 70°. In Hawaii, the earthquakes with M &gt; 7 rupture the decollement plane, since it is the only surface large enough to generate magnitude 7 or larger earthquakes. The percentage of reverse faulting events is high compared to the decollement and normal faulting mechanisms for the period 1972–1983. The percentage of decollement type focal mechanisms becomes dominant after 1983. This pattern of faulting activity suggests that pressure was building up within Kilauea's rift zone prior to the 1983 Puu'Oo eruption. Overall, a single stress orientation with the maximum compressive stress oriented SE perpendicular to the rift and dipping at 45° is compatible with the coeval existence of decollement, reverse, and normal faults. However, in a crustal volume east of longitude 155°10′W, we find a change of the orientation of σ1 from nearly horizontal to plunging 45° SE occurring in 1979. This stress rotation suggests magma movements within the aseismic part of Kilauea's east rift zone. The strain and stress orientations are coaxial in the south flank except within the volume where the stress rotation is observed. We observe a change in the relationship between stress and strain directions caused either by the shifting of seismic activity from reverse faults to decollements, while stress stays constant, or by a rotation of stress, while strain remains constant. Assuming that the model of a noncohesive Coulomb wedge is appropriate for Kilauea's south flank, we find that high pore pressures are prevalent along the decollement and within the wedge for a coefficient of friction equal to 0.85.

Structure of the mobile south flank of Kilauea Volcano, Hawaii

How do huge landslides occur in Hawaii? We can begin to address this question by examining an incipient landslide structure on the mobile south flank of Kilauea volcano. Between 1970 and 1989 this part of the south flank moved in a southeast direction (seaward by 4 m on its eastern end and by 10 m on its western end), parallel to the western boundary of a large oifshore topographic bench. We show that abrupt decreases in cumulative flank movement correspond to abrupt decreases in hypocenter concentration. Most of the earthquakes occur in a 7‐ to 10‐km‐deep band beneath the south flank, and relative relocation of a subset of these events shows that they are derived from a single horizontal plane which appears to be a decollement. Focal mechanisms for this deep seismicity are dominated by thrust fault mechanisms, with the overlying block moving seaward. A comparison of south flank bathymetry with decollement structures on other volcanoes suggests that the distal end of this decollement coincides with the distal end of the offshore topographic bench (30–50 km from shore and 5 km below sea level). The fault system forming the Hilina pali is connected with this decollement, so that slip on the Hilina pali's normal faults coincides with slip of the decollement and with reverse faulting of the distal end of the bench. (This interpretation is supported by studies of the events associated with the M7.2 Kalapana earthquake in 1975.) When the Hilina pali is connected to the outer bench with a decollement inferred from seismicity, a wedge 8‐km thick adjacent to Kilauea's magma system and 1–2 km thick along the edge of the outer bench (an area over 3000 km2) is defined. The thickness adjacent to the magma system decreases eastward along Kilauea's East Rift Zone and southwestward from Kilauea caldera, whereas the thickness of the lip of the outer bench decreases only eastward. Geologic observations and geophysical studies of the subaerial flank reveal an extensive magma system beneath the Koae fault system as well as beneath Kilauea caldera and its East Rift Zone. We infer that the boundary of the proximal end of the mobile flank is the southern boundary of this magma system and that this boundary coincides with the southern boundaries of both the Koae fault system and Kilauea's East Rift Zone.

Estimate of the stress field in Kilauea's South Flank, Hawaii

SUMMARY We estimated stress and seismic strain tensors for the Kilauea volcano's south flank. The stress orientation inversion and the seismic strain calculation were performed using fault-plane solutions. The principal stress and seismic strain directions are approximately uniformly distributed in space and time during the interval covered by the data. However, the 01,03 plane is approximately orthogonal to the plane. Therefore, a weak layer may exist beneath the south flank. o1 has a plunge of 59 and an azimuth of 152, with a 10 95 per cent confidence range. We also developed a stress magnitude inversion to estimate magnitudes of boundary and interior stresses. In this inversion, the principal stress directions were taken as constraints in the seismic volume, and surface geodetic observations were used as data. The maximum magmatic pressure in Kilauea's rift zone is about 160 MPa. The direction of o1 can be interpreted as the superposition of hydrostatic stress (pgh) and magmatic pressure. Without the constraint imposed by the direction of ol, the estimated pressure is only 60 MPa. The distribution of magmatic pressure may be similar to that of pgh. In contrast, the upper rift zone may be in tension. The shear stress in the rift zone is about one order of magnitude smaller than the maximum compressive stress, supporting the interpretation of magmatic flow as fluid in dikes or channels. The combination of stress orientation inversion, seismic strain calculation, and stress magnitude inversion performed in this study provides a means by which to estimate the stress state in seismic areas.