Magma Ascent and the Pressurization of Mount Etna's Volcanic System

See allHide authors and affiliations

Science  28 Mar 2003:
Vol. 299, Issue 5615, pp. 2061-2063
DOI: 10.1126/science.1080653


After a period of deflation during the 1991–1993 flank eruption, Mount Etna underwent a rapid inflation. Seismicity and ground deformation show that since 1994, a huge volume of magma intruded beneath the volcano, producing from 1998 onward a series of eruptions at the summit and on the flank of the volcano. The last of these, started on 27 October 2002, is still in progress and can be considered one of the most explosive eruptions of the volcano in recent times. Here we show how geodetic data and seismic deformation, between 1994 and 2001, indicate a radial compression around an axial intrusion, consistent with a repressurization of Mount Etna's plumbing system at a depth of 6 to 15 kilometers, which triggered most of the seismicity and provoked the dilatation of the volcano and the recent explosive eruptive activity.

Since 1989, volcanic activity at Mount Etna has been monitored by seismic and geodetic networks, which have improved the capability of observers to track, in near–real time, subsurface magma movements (1–3). Seismicity and eruptions at Mount Etna are not randomly distributed in space and time. The eruptions of the past 30 years are mainly related to magma rising on the two NNW and NE primary structural trends (1, 4), recognizable both in the volcanic area and in the regional context (Fig. 1, top left), coherent with the trends of the eruptive fissures (Fig. 1, right) and the distribution of pyroclastic cones (1, 5). Studies of recent flank eruptions have shown preparation times lasting from months to several years (6). Modeling shows that a tensile mechanism, associated with the principal recent lateral eruptions (1989, 1991–1993, and 2001), occurred along the NNW-trending structures (6). Increases in seismicity and the inflation phases of the volcano, preceding eruptions, are usually interpreted as being due to stress changes caused by the movement and accumulation of magma. Since the 1991–1993 flank eruption, ground deformation and seismicity show a continuous dilatation of the entire edifice and a gradual increase of the seismic strain release, respectively, suggesting a progressive accumulation of magma beneath the volcano (2, 3). There are several geophysical evidences for a shallow magma reservoir at a depth of ∼3 to 5 km below sea level (7, 8). These depths match the horizon of neutral buoyancy at Mount Etna, as suggested by petrological studies (9).

Figure 1

(Right) Structural map of Mount Etna showing eruptive fissures (red lines) and major faults (black lines). Also shown are epicentral locations (blue squares) of the 647 best constrained earthquakes [average root mean square (rms) value < 0.1 s; average horizontal (Erh) and vertical (Erz) location errors < 1 km] recorded in the period from 1994 to June 2001 and localized by the 3D velocity model of Figs. 2 and 3. The elevation contour lines at 500-m intervals are also illustrated. (Top left) Structural map of eastern Sicily. Numbers represent the following: 1, volcanics; 2, front of the Appenninic-Maghrebian Chain; and 3, main faults. (Bottom left) The eruptive fractures (red lines) and the lava flows of the 18 July to 9 August 2001 flank eruption (yellow) and of the 27 October 2002 flank eruption (orange) (lava flows are updated to 25 December) are shown. Lava flows of the 1999 summit eruptions are also indicated (gray).

The size, extent, and depth of the magma chamber structure beneath Mount Etna are still poorly characterized. Recent geophysical and petrological studies (4, 10, 11), together with geochemical surveys (12), exclude the presence of a large midcrustal magma chamber. At a depth of 5 to 12 km, magma can be stored as a plexus of dikes and sills rather than as a unique magma chamber (11).

Tomographic images of Mount Etna reveal the presence of an almost aseismic high-VP body (VP , P-wave velocity) extended between a depth of 1 and ∼18 km (13), which is centered under the southern part of the Valle del Bove and also extends below the summit region. The three-dimensional (3D) velocity structure shown in Figs. 2 and3 is from the inversion of an augmented data set (647 earthquakes recorded in the period from 1994 to 2001), which allows us to improve even the most recent results (14–16). This new inversion has been obtained by the use of Thurber's method (17) and SimulPS-14 software. Within the crust at Mount Etna, there is no evidence of a low-velocity zone interpretable as a magma chamber (18). Therefore, the most important feature revealed by tomography is the presence of a central high-velocity body (HVB), which is interpreted as a main solidified intrusive body. The HVB shows a roughly ellipsoidal shape in the upper crust (depth < 10 km) with a NNW-SSE horizontal axis and a vertical axis extending between 1 and 9 km below sea level. The horizontal extension decreases with depth, from 8 to 10 km at a depth of 3 to 9 km, to 4 to 6 km below a depth of 12 km (Fig. 2).

Figure 2

Horizontal projections of the (left) P-axis and (right)T-axis plots (black lines) on the computedVP model at depths of 3 (top), 12 (middle), and 18 km (bottom). P and T axes refer to earthquakes (white dots) located at depths of 0 to 3 (top), 6 to 15 (middle), and 18 to 24 km (bottom). The P axes are radial to the main central HVB. Only axes with a plunge of <30° are plotted. The velocity contour interval is 0.2 km/s. A data set of 647 earthquakes, with 8474 P-wave arrivals and 850 S-wave arrivals, has been used for the inversion. Our final rms value is 0.09 s, whereas horizontal and vertical location errors are reduced to <0.50 and <0.70 km, respectively. ThisVP model does not show substantial differences with respect to the already published models (13–16), but the resolution is improved because of the use of more data. Red triangles indicate the positions of the summit craters.

Figure 3

(Left) A map of Mount Etna showing the displacement vectors (red arrows) calculated by GPS measurements for 1993–1997. The deformation pattern describes a marked expansion of the volcano edifice, which is due to the recharging phase that preceded the tens of lava fountains (from 1998 to 2000) at summit craters, the February to November 1999 summit eruption, the July to August 2001 flank eruption, and the October 2002 flank eruption. (Right) In the two N80°E cross sections of the VP model (Fig. 2), contoured every 0.2 km/s, the projections (white lines) of the (top) P and (bottom) T axes are reported for each earthquake (black dots). Angles that are <30° with respect to the projection plane are considered. The most important feature is that Paxes point away from the mainly aseismic HVB at depths between 6 and 15 km, whereas T axes are arranged tangentially. A and B indicate the extremities of the N80°E section reported on the map (white line A–B).

We propose that a region of magma storage exists below a depth of 3 km in the upper western part of the HVB, where fractionation processes can develop, as suggested by the depth of the pressuring source inferred by the deformation pattern (7). This was confirmed by positive gravity changes (19) recorded during 1994–1996. The deeper part of the HVB represents the plumbing system used by the magma during its ascent from the mantle source.

The 3D velocity model allows us to determine accurate hypocenters, azimuths, and takeoff angles to compute focal mechanisms. For the 647 earthquakes, we obtained 200 well-constrained double-couple fault plane solutions, by using the classical approach ofP-wave polarities [see (4) for details]. Most earthquakes have >20 polarities. Because focal mechanisms depend on earthquake location and velocity models, we performed several tests to explore how hypocentral errors affect the stability of the focal mechanism solutions.

The computed solutions offer a unique opportunity to define the seismic deformation of the volcano during inflation. In particular, by plotting the principal strain axes P(pressure) and T (tension), obtained by focal mechanism solutions, on the 3D velocity model, we can observe that, between a depth of 6 and 15 km, P axes point away from the mainly aseismic HVB, whereas T axes are tangential to the HVB (Fig. 3). The approximately N-S–trending regional compression (20) is not consistent with this seismic deformation pattern. This implies that earthquakes at Mount Etna mainly reflect local stress induced by magma pressure. The displacement vectors (Fig. 3) calculated by Global Positioning System (GPS) measurements for 1993–1997 show a pattern similar to that of the P axes, strengthening a connection with seismic deformation and supporting the idea that intrusive mechanisms mainly occur along the NNW-SSE direction. Conversely, we observe that P axes are NNW-trending below a depth of 15 km, suggesting a closer relation with the regional stress at greater depth (4).

A radial pattern of compression in the uppermost 10 km at Mount St. Helens has been interpreted as an effect of pressurization within a magma chamber, defined by a central aseismic zone (21). The radial compression observed at Mount Etna during the past years of inflation reveals that earthquakes were generated by repressurization of the plumbing system.

The magma would prevalently rise along the western and northwestern border of the HVB, where the two main NNW-SSE and NE-SW tectonic structures intersect (Fig. 1) and structural weaknesses develop between the HVB and surrounding rocks. Therefore, these two fault systems play a key role in magma rise and stationing beneath the volcano. Magma would be injected from the upper mantle into the crust, triggering seismicity due to the effect of magma pressure against the hosting rocks. Although the lateral dimension of the intrusive body is variable in the mid-crust, the similarity of focal solutions would indicate that the seismic response to the magma pressure is quite uniform. We therefore hypothesize that a huge intrusion of magma is mostly confined within the HVB at a depth of 6 to 15 km and that an increase of magma in this zone between 1994 and 2001 caused progressive repressurization of the deep plumbing system. At a shallower depth (0 to 5 km), seismic deformation indicates that P- andT-axis orientations are different when compared to the orientations of P and T axes at a greater depth. Also, the focal solutions mainly show normal faulting at a shallow depth (Fig. 3). This suggests that magma was stored mostly below 3 to 5 km, bending the overburden and generating shallow normal faulting events with vertical P axes. The concentration of stress at the crack tip, during dike intrusion, can also produce shear failure, so that the system operates as a network of magma-filled cracks connected by faults.

This hypothesis implies a continuous injection of magma from a depth of 6 to 15 km into the shallow (depth of 3 to 5 km) magma reservoir during 1994 to 2001, with magma accumulation in the upper part of Mount Etna's plumbing system. We suggest that the 2001 intrusive mechanism is consistent with overpressure of the shallow magma reservoir, which fractures the upper storage boundary, causing an upward intrusion of magma (16, 22). The 18 July to 9 August 2001 flank eruption (Fig. 1, bottom left) was heralded by several days of intense seismicity, ground deformation, and fracturing. Seismicity and ground deformation indicate that dike emplacement occurred in just 6 days before the onset of the eruption.

In the past 30 years, major eruptions at Mount Etna have normally been followed by a period of scarce seismicity that can last from a few months to 1 to 2 years. Conversely, after the end of the 2001 eruption, the seismicity rate remained high, suggesting, together with geodetic data showing a renewal of the areal dilatation, a new magma accumulation at shallow depths (3 to 5 km). This is confirmed by the reactivation of the summit craters from March 2002 onward and by the present flank eruption that began on 27 October 2002 (Fig. 1, bottom left). The onset of the 2002 eruption was abrupt, with only a few hours of premonitory seismicity. Also, this eruption appears to have been triggered by the overpressure in the shallow reservoir, caused by magma accumulation.

Geophysical evidences presented here indicate that a considerable quantity of magma is still accumulated beneath the volcano in the shallow storage zone. Considering that the volcano has been intensely fractured during the past two flank eruptions, Mount Etna's eruptive activity could become more frequent, voluminous, and potentially hazardous in the near future.

  • * To whom correspondence should be addressed. E-mail: patane{at}


View Abstract

Navigate This Article