Earthquake Volcano Interactions Research Articles

Vesiculation of Rhyolitic Melts Under Oscillatory Pressure

Magma ascending in the Earth’s crust can undergo oscillations in pressure, from ultra-low frequency changes associated with tectonics, to relatively higher frequency oscillations associated with seismicity. Seismic waves travelling through shallow magma bodies can lead to a range of unrest phenomena and potentially trigger volcanic eruptions. The mechanisms by which pressure oscillations can induce unrest or eruption remain debated. Here, we experimentally impose pressure oscillations on magma and study how they affect vesiculation processes. We use cylindrical samples (4.00 mm long, 4.85 mm diameter) of hydrous rhyolitic obsidian (0.11 ± 0.01 wt% H2O) placed in alumina (AL23) crucibles and vary pressure by the uniaxial loading of an alumina plunger in a thermo-mechanical analyzer. We monitor vesiculation at temperatures of 950–990°C and confining pressure of 177 kPa. We perform two types of experiment: 1) “static” experiments (at constant pressure) and 2) “oscillating” experiments in which we impose sinusoidal pressure oscillations of up to 71 kPa upon the static pressure (i.e., between 106 and 250 kPa). In both cases, we dilatometrically observe sample expansion driven by vesiculation. Post-experimental bubble textures reveal that bubbles formed preferentially at the sample margins. For the oscillating experiments, the sample expansion rate is lower than in the static experiments, and there are fewer vesicles at the sample margins. We examine the constituent processes of bubble formation (nucleation, growth, coalescence) and gas loss (diffusion, permeable flow) occurring during static experiments and with the added element of pressure oscillations. The most likely mechanism responsible for reduced sample expansion is that pressure oscillations drive the sample in and out of water saturation conditions and thus reduce the fraction of residence time over which bubble nucleation and/or growth are driven. Future work will be needed to confirm this hypothesis. These results are relevant to the study of earthquake-volcano interactions, where a magma body that sits close to volatile saturation is subject to pressure fluctuations.

Numerical wave propagation study of the unusual response of Nevados de Chillán volcano to two aftershocks of the 2010 MW = 8.8 Maule earthquake

In 2012, two large aftershocks of the MW = 8.8 Maule, Chile earthquake occurred near the Central Chilean coast. Both events, with magnitudes MW = 6.1 and MW = 7.1 were located about 200 km away from the volcanic/hydrothermal Nevados de Chillán (NdC) complex. The system responded with a significant increase in seismicity to the MW = 6.1 event, and with a reduction in seismicity following the latter, larger MW = 7.1 earthquake. To understand the physical changes induced by both earthquakes at the NdC, we set up a numerical wave propagation study. Our results show that these observations likely result from dynamic changes at the NdC complex during the passage of the seismic waves. In the case of the MW = 6.1 aftershock, the extensive stress transfer was large enough to induce changes at the hydrothermal system, which could have later led to an increase in fluid-driven seismicity. In the case of the MW = 7.1 earthquake, the oblique arrival of seismic waves to the main fault of the NdC due to the MW = 7.1 earthquake induced slight compressions at the reservoir and created openings in the main fault, promoting fluid migration, which can explain the reduction in seismicity beneath the main craters of the volcano. Results also show that the geometry of an underlying fault system can influence the volumetric response, suggesting that the role of local structures in earthquake-volcano interactions influence the behavior of triggered systems and should therefore not be disregarded.

Surface displacements of Aso volcano after the 2016 Kumamoto earthquake based on SAR interferometry: implications for dynamic triggering of earthquake–volcano interactions

SUMMARYThe 2016 Kumamoto earthquake involved a series of events culminating in an Mw 7.0 main shock on 2016 April 16; the main-shock fault terminated in the caldera of Aso volcano. In this study, we estimated surface displacements after the 2016 Kumamoto earthquake using synthetic aperture radar interferometry analysis of 16 Phased Array Type L-band Synthetic Aperture Radar-2 images acquired from 2016 April 18 to 2017 June 12 and compared them with four images acquired before the earthquake. Ground subsidence of about 8 cm was observed within about a 3 km radius in the northwestern part of Aso caldera. Because this displacement was not seen in data acquired before the 2016 Kumamoto earthquake, we attribute this displacement to the 2016 Kumamoto earthquake. Furthermore, to estimate the source depth of the surface displacement, we applied the Markov chain Monte Carlo method to a spherical source model and obtained a source depth of about 4.8 km. This depth and position are nearly in agreement with the top of a low-resistivity area previously inferred from magnetotelluric data; this area is thought to represent a deep hydrothermal reservoir. We concluded that this displacement is due to the migration of magma or aqueous fluids.

Revisiting short-term earthquake triggered volcanism

It has been noted for centuries that earthquakes appear to trigger the eruption of volcanoes. For example, analyses of global volcanic and seismic records since 1500 AD have shown that explosive eruptions with Volcanic Explosivity Index (VEI) values ≥2 are preceded within days by nearby major earthquakes (magnitude M8 or larger) about 4 times more often than expected due to coincidence, suggesting that large earthquakes can trigger eruptions. We expand the definition of a triggered eruption to include the possibility of M6 or greater earthquakes within 5 days and 800 km of a VEI 2 or greater eruption. Removing pre-1964 records, to ensure complete and accurate catalogs, we find 30 volcanoes that at some point experienced a potentially triggered eruption and define these volcanoes as “sensitive” volcanoes. Within this group of sensitive volcanoes, normalized distributions of volcano-centric factors such as tectonic setting, dominant rock type, and type of volcano are practically indistinguishable from those of sensitive volcanoes in which the time of eruption is randomized. Comparisons of sensitive volcanoes and insensitive volcanoes (i.e., volcanoes that have never experienced a triggered eruption) reveal that sensitive volcanoes are simply more active than insensitive volcanoes: They erupt more frequently, are located solely in subduction zones, and erupt primarily andesites and basaltic-andesites. The potentially triggered eruptions do not show the magnitude-distance relationship expected for seismically induced responses (e.g., hydrologic responses), and eruptions that do occur within days of nearby earthquakes do so within rates expected by random chance. There is, however, a 5–12% increase in the number of explosive eruptions in the 2 months to 2 years following major earthquakes. We conclude that short-term seismically triggered explosive eruptions occur less frequently than previously inferred, an important conclusion when considering volcanic hazards and for understanding the nature of earthquake-volcano interactions.

Regional earthquakes followed by delayed ground uplifts at Campi Flegrei Caldera, Italy: Arguments for a causal link

Earthquake-triggered volcanic activity promoted by dynamic and static stresses are considered rare and difficult-to-capture geological processes. Calderas are ideal natural laboratories to investigate earthquake–volcano interactions due to their sensitivity to incoming seismic energy. The Campi Flegrei caldera, Italy, is one of the most monitored volcanic systems worldwide. We compare ground elevation time series at Campi Flegrei with earthquake catalogues showing that uplift events at Campi Flegrei are associated with large regional earthquakes. Such association is supported by (yet non-definitive) binomial tests. Over a 70-year time window we identify 14 uplift events, 12 of them were preceded by an earthquake, and for 8 of them the earthquake-to-uplift timespan ranges from immediate responses to 1.2 yr. Such variability in the response delay may be due to the preparedness of the system with faster responses probably occurring in periods during which the Campi Flegrei system was already in a critical state. To investigate the process that may be responsible for the proposed association we simulate the propagation of elastic waves and show that passing body waves impose high dynamic strains at the roof of the magmatic reservoir of the Campi Flegrei at about 7 km depth. This may promote a short-lived embrittlement of the magma reservoir's carapace otherwise marked by a ductile behaviour. Such failure allows magma and exsolved volatiles to be released from the magmatic reservoir. The fluids, namely exsolved volatiles and/or melts, ascend through a nominally plastic zone above the magmatic reservoir. This mechanism and the associated inherent uncertainties require further investigations but the new concept already implies that geological processes triggered by passing seismic waves may become apparent several months after passage of the seismic waves.

Numerical simulations (2D) on the influence of pre-existing local structures and seismic source characteristics in earthquake-volcano interactions

Earthquake triggering of hydrothermal and volcanic systems is ubiquitous, but the underlying processes driving these systems are not well-understood. We numerically investigate the influence of seismic wave interaction with volcanic systems simulated as a trapped, high-pressure fluid reservoir connected to a fluid-filled fault system in a 2-D poroelastic medium. Different orientations and earthquake magnitudes are studied to quantify dynamic and static stress, and pore pressure changes induced by a seismic event. Results show that although the response of the system is mainly dominated by characteristics of the radiated seismic waves, local structures can also play an important role on the system dynamics. The fluid reservoir affects the seismic wave front, distorts the static overpressure pattern induced by the earthquake, and concentrates the kinetic energy of the incoming wave on its boundaries. The static volumetric stress pattern inside the fault system is also affected by the local structures. Our results show that local faults play an important role in earthquake-volcanic systems dynamics by concentrating kinetic energy inside and acting as wave-guides that have a breakwater-like behavior. This generates sudden changes in pore pressure, volumetric expansion, and stress gradients. Local structures also influence the regional Coulomb yield function. Our results show that local structures affect the dynamics of volcanic and hydrothermal systems, and should be taken into account when investigating triggering of these systems from nearby or distant earthquakes.

Evolution of deformation and stress changes during the caldera collapse and dyking at Bárdarbunga, 2014–2015: Implication for triggering of seismicity at nearby Tungnafellsjökull volcano

Stress transfer associated with an earthquake, which may result in the seismic triggering of aftershocks (earthquake–earthquake interactions) and/or increased volcanic activity (earthquake–volcano interactions), is a well-documented phenomenon. However limited studies have been undertaken concerning volcanic triggering of activity at neighbouring volcanoes (volcano–volcano interactions). Here we present new deformation and stress modelling results utilising a wealth of diverse geodetic observations acquired during the 2014–2015 unrest and eruption within the Bárdarbunga volcanic system. These comprise a combination of InSAR, GPS, LiDAR, radar profiling and optical satellite measurements. We find a strong correlation between the locations of increased seismicity at nearby Tungnafellsjökull volcano and regions of increased tensile and Coulomb stress changes. Our results suggest that stress transfer during this major event has resulted in earthquake triggering at the neighbouring Tungnafellsjökull volcano by unclamping faults within the associated fissure swarm. This work has immediate application to volcano monitoring; to distinguish the difference between stress transfer and new intrusive activity.

Seismic activity of the Nevados de Chillán volcanic complex after the 2010 Mw8.8 Maule, Chile, earthquake

Couplings between large magnitude subduction zone earthquakes and the subsequent response of their respective volcanic arcs are generally accepted, but the mechanisms driving this coupling are not known. The 2010 Maule earthquake (Mw8.8) ruptured approximately 500km along strike in the south-central part of Chile, and provides an opportunity to investigate earthquake–volcano interactions. In an exploratory campaign, we deployed four broadband seismometers atop and surrounding the Nevados de Chillán volcanic complex because it is located directly behind one of the two primary slip patches of the Maule Earthquake. The data recorded (from December 2011 to April 2012) shows significant seismic activity, characterized by numerous volcano tectonic events and tremor episodes occurring within the volcanic complex. We recorded two strong aftershocks of the Maule earthquake (Mw 6.1 in January 2012, and Mw 7.1 in April 2012) and investigated the response of the volcano to the incoming seismic energy. We find that volcanic tremor increased significantly upon arrival of the seismic waves from the Mw 6.1 event, which was then followed a few hours later by a significant increase in volcano-tectonic events when tremors subsided. This delay between tremor and volcano-tectonic events suggests that fluid and/or magma migration (manifested as tremor) readjusted the local stress state that then induced the volcano-tectonic events. The increased activity persisted during the subsequent two weeks. In contrast, the Mw 7.1 event, which occurred at a similar epicentral distance from the volcano, did not produce any significant seismic response of the volcanic complex. Analysis of the particle velocity records of the two events shows that the volcanic system was perturbed in different ways because of the incidence angle of the incoming energy, inducing a back-elliptical vertical and fault parallel motion for the Mw 6.1 and no clear directional dependence for the Mw 7.1. This suggests dilatation-induced fluid migration within the complex, and a kinematic mechanism of the perturbation rather than the perturbation amplitude. Our results demonstrate the importance for continued monitoring of the arc behind Maule, with an increased seismometer and GPS array density to determine the style of deformation currently occurring in the arc.

Stress field change around the Mount Fuji volcano magma system caused by the Tohoku megathrust earthquake, Japan

Crustal deformation by the M w 9.0 megathrust Tohoku earthquake causes the extension over a wide region of the Japanese mainland. In addition, a triggered M w 5.9 East Shizuoka earthquake on March 15 occurred beneath the south flank, just above the magma system of Mount Fuji. To access whether these earthquakes might trigger the eruption, we calculated the stress and pressure changes below Mount Fuji. Among the three plausible mechanisms of earthquake–volcano interactions, we calculate the static stress change around volcano using finite element method, based on the seismic fault models of Tohoku and East Shizuoka earthquakes. Both Japanese mainland and Mount Fuji region are modeled by seismic tomography result, and the topographic effect is also included. The differential stress given to Mount Fuji magma reservoir, which is assumed to be located to be in the hypocentral area of deep long period earthquakes at the depth of 15 km, is estimated to be the order of about 0.001–0.01 and 0.1–1 MPa at the boundary region between magma reservoir and surrounding medium. This pressure change is about 0.2 % of the lithostatic pressure (367.5 MPa at 15 km depth), but is enough to trigger an eruptions in case the magma is ready to erupt. For Mount Fuji, there is no evidence so far that these earthquakes and crustal deformations did reactivate the volcano, considering the seismicity of deep long period earthquakes.

Monitoring radon emission anomalies at Stromboli Island as a tracer of eruptive events and “near field” earthquakes

Radon emission from soil at Stromboli Island has been monitored since 2002 utilizing a network of 25 stations and exposing two kinds of detectors: E-PERM and calibrated track-etches (LR115). We present and discuss the data from 2002 to 2007, thus including the last two major eruptive cycles. Earthquake-volcano interactions were detected providing evidence that radon emissions are somehow related to the occurrence of regional earthquakes. Single deep earthquakes related to active subduction, such as Salina event (ML =5 .1) of May 5, 2004, may be capable of increasing the ascent of geothermal fluids due to the passage of seismic waves. In addition, we observed major eruptions at Stromboli were preceded by anomalies that occurred at three summit stations that reached values above 20000 Bq m −3 : this is considered an anomalous value and could be regarded, under particular environmental conditions, as a potential precursory signal of a change in volcanic activity.

Earthquake–volcano interactions detected from radon degassing at Stromboli (Italy)

We investigated earthquake–volcano interactions by using a network for radon monitoring at Stromboli volcano. Radon is an alpha emitting radioactive gas produced from the decay of uranium bearing rocks, soils and magmas. Its spatial and temporal variations have been regarded as precursors of earthquakes and volcanic eruptions. Here we provide evidence of how radon emissions at Stromboli can be correlated to high magnitude ( M L > 5) regional earthquakes and erupted magma volumes. The Palermo earthquake of September 6, 2002 ( M L = 5.6), characterised by shallow hypocentral depth (∼ 15 km) and higher number of energetic aftershocks, enhanced more efficiently postseismic dynamic triggering that may have contributed to triggering the eruptions of Mount Etna and Stromboli by the end of 2002. A viscoeslastic relaxation mechanism seems to be compatible with the onset of both eruptions. The rate of erupted magma volume at Stromboli is positively correlated with the rate of radon degassing, and suggests a possible link between magma chamber volume, gaseous transfer and dynamic response of the volcano to near field seismic triggering. Single and isolated deep earthquakes related to active subduction, such as the Salina event ( M L = 5.1) of May 5, 2004, are capable of mobilising fluids (due to the passage of seismic waves at higher depths) but their dynamic effect on the fracture network and the feeding system of the volcanoes seems to be limited, likely due to the near-surface attenuation of seismic waves. It is not excluded that the coupling of two or more earthquakes of similar depth and/or higher magnitude in the near field, could play a more efficient role in the triggering process.

Earthquake–Volcano Interactions

New measurements, statistical analyses, and models support the conjecture that a large earthquake can trigger subsequent volcanic eruptions over surprisingly long distance and time scales.