Mt Etna’s east flank surface displacements & the siting of future eruptions

Scientific Report 929

This report is also available as a PDF document .

Leica system 530 GPS set up on the unstable east flank, with summit behind, 20^th September 2010

Abstract

This project is aimed at determining not the time, but the position of the next eruption of Mt Etna volcano, by measuring the detailed relative displacements of a dense network of GPS stations.

Background

In 2009, Etna completed its longest flank eruption in 17 years (13th May 2008 - 6th July 2009). Like the eruptions of 2004 and 2006 October, the main eruptive vents were associated with the Southeast Crater, and were situated at the head of the Valle del Leone in its southern part. The activity was largely effusive, with comparatively little explosive activity. The lavas from these three eruptions have covered much of the lower part of the Valle del Leone, and the central Valle del Bove. In 2010 there was an ashy emission at SE Crater on April 10th.

The Etna ground deformation network is the longest-lived and by far the densest of such networks on Etna, with 95 stations at present, and it is the only one with stations within the Valle del Leone and the Valle del Bove, (where the recent eruptions have taken place) installed in 1983 and 1987. These stations were measured initially by trilateration, and from 1994 by dualfrequency GPS. Three of these long-lived stations were destroyed by the lavas of 2008-9, so one objective for 2010 will be the installation of new stations to replace these lost stations, and to add 3 more to increase the detail in this region.

Previous results

The results of these measurements have shown a marked difference in behaviour between the east side of the volcano and the west. Those stations on the east side have shown movements towards the sea (mainly easterly or southeasterly) of more than 2 metres in some cases, i.e. an average of 10 cm per year, though the rate of movement has varied greatly from year to year. Those on the west, on the other hand, have shown radial outward movements about one third of the magnitude of those on the east. On the lower flanks they have been comparatively stable, showing movement of about 1 cm per year or less. The dividing line between the two halves is marked by the northeast rift, continuing into the Pernicana fault to the north, and by the southern rift to the south, with other discontinuities within the unstable sector.

Underlying rationale

The apparent consequence of these movements has been the persistent opening of new fissures as a result of the tensional movement between the two halves of the volcano. The 1985, 1989, 1991-3, 2001, 2002-3 and 2008-9 eruptions all included north-south to NE-SW fissures, some of them intrusions from which no magma was erupted. There is also clear evidence from the data that once magma pathways had been determined by these large scale events, magma pressure greatly added to the movement, and that this magmatic component had a different proportion of the total movement for each eruption.

Another factor in the determination of eruption sites and supply channels has been the loading effect of erupted lava. This was first noticed in regard to the levelling measurements (Murray 1988¹) and was evident after the 1991-3 eruption (Massonet et al. 1995²). In its simplest form, shear stress following flow emplacement reaches a maximum at the flow edge, causing fissures to open here which may later be used by erupting lavas.

Application of the three-dimensional measurements of movements at Etna to determine likely fissure positions and directions using mathematical modelling of the stress regime was attempted by Dowden et al (1995)³. As an extension of this work, Dowden et al. (1997⁴) took the three dimensional movements as surface boundary conditions, then integrated downwards to determine internal displacements, from which stress distributions within the mountain could be calculated. These studies pinpointed an area at depth from the 1989-1990 displacements that later marked the site of the 1991-3 eruption, and an area beneath the Southeast Crater and another SW of the Torre del Filosofo visible in the 1991-2 displacements marked the site of the February 1999 fissure eruption and the initial fractures of the 2001 eruption respectively (La Delfa et al. 2001⁵, Bonaccorso et al. 2002⁶).

Specific objectives

1. To determine the effect of recent surface displacements and lava loading on subsequent eruption position.
2. To measure the surface displacement 2009-2010 to determine internal stress distributions and likely positioning and orientation of future eruptive fissures.
3. To measure the 96-station dual-frequency GPS network, using 7 Leica System 530 sets.
4. To install and measure new GPS stations, to replace those destroyed 2008-9 and intensify the network on the upper eastern flank.
5. To occupy and extend the 236-station precise levelling network.
6. To measure the 27 Dry Tilt stations on the flanks.

All the practical objectives were achieved, and substantive advances in objective 1 are ongoing.

Volcanic activity

Since the previous trip in Sep-Oct 2009, Mt Etna had been essentially inactive, apart from quiet degassing from the summit and the 2008-9 eruptive fissure. Earlier in the year there had been occasional collapses and emissions of ashy clouds, such as that on June 19th, when ash was emitted from a vent at the southeast foot of the Southeast Crater for about an hour in the early morning.

During the trip, there was little sign of activity apart from quiet degassing from the Chasm, Northeast & Southeast craters and the Bocca Nuova. During visits to the summit, no noises could be heard from the Chasm or Bocca Nuova, but on September 12th strong prolonged rumbling could be heard from the Northeast Crater, indicating gas escape from the magma.

Survey procedure

We reoccupied all the ground deformation networks first established in 1975 (levelling and dry tilt) and between 1981 and 1987 (trilateration to 1994, GPS from 1994-2008). At the present time, the ground deformation stations consist of 95 dual-frequency GPS stations which are habitually measured as a static survey with occupation times 30 minutes to 18 hours, depending upon line lengths. Results give error ellipses to <9mm accuracy, precise levelling benchmarks which yield precisions of <1mm per km, and 27 dry tilt stations at widely scattered locations around the summit and flanks of the volcano, which measure changes in ground tilt of 3 to10µ.

Data quality was good, similar to previous years, though some stations on the lower western flank experienced satellite problems, apparently due to the ever increasing height and density of conifer trees in the Corpo Forestale territory. Data were post-processed using Noto, Cagliari and Matera permanent GPS stations, using Leica GeoOffice software.

Two new GPS stations were installed, and during October, we again occupied the new precise levelling branches linking the southern traverse to the Rifugio Sapienza, and the one linking the northern traverse to the Piano della Concazze. We also established a new loop following the Corpo Forestale track, linking the Sapienza to the Piano Provenzana, to create a new giant levelling network extending the previous 236 station network to 343 stations, and more than doubling the distance levelled from 34 km to 78 km. The aim of this levelling is to provide more accurate height control of GPS stations, 32 of which are common to both networks.

Personnel

The measurements were carried out by the following volunteer assistants:

Toby Balaam, University of Sheffield.

Guillermo Caravantes-Gonzalez, Open University

Anne Forbes, Open University

John B.Murray, Open University

Anne Peterson, Open University

Andrew Pitty, University College London

Phil Sargent, Nottignham Trent University

Julia Scott, Cambridge University

Richard Wall, University College London

Xiomara Gabriela Villagomez, Tarragona, Spain.

Figure 1. Vectors of horizontal movement 2009-2010, derived from GPS values at 95 stations

Preliminary findings

On previous occasions, it has been noted that whilst horizontal movements of benchmarks prior to eruptions could broadly be described as radial to the summit region, in detail the movement could be broken up into groups of stations with movements almost parallel to each other. This suggests that whilst the volcano as a whole is behaving elastically and/or plastically (Murray & Pullen 1984⁷), at a smaller scale the volcano surface is breaking up into slabs, with increased tensional strain or fissures between. This was particularly the case in 1987-1988 in the Valle del Leone, where such movements indicated tensional strain nearly one year prior to the 1989 eruption which included a voluminous eruptive fissure there, and also in 2005-6 prior to the flank eruption which began on 2006 October 13th, right at the end of our trip that year.

The measurements 2009-2010 (see fig. 1, opposite) suggest a similar difference between the movements in three areas of the volcano: the northeast, southeast and the lower northern sectors. Most stations in a block approximately 5 x 5 km northeast of the summit moved parallel to each other 4-6 cm ENE, whereas those southeast of a line running diagonally across the middle of the Valle del Bove have moved in parallel 2-3 cm ESE. Another block is represented by those stations north of Monte Maletto, which effectively comprise the most northerly quarter of the volcanic edifice, and which have moved 2-3 cm NNE. The remaining stations show a slight outward spreading from the summit of 0.5-1.5 cm.

Discussion

The most commonly used models of volcano deformation, developed initially by Anderson (1936)⁸ and later applied to deformation observed in active volcanoes by Mogi (1958)⁹, describe elastic displacement in an infinite half-space, and other more recent models are largely based on the same premises (e.g. Okada 1992¹⁰). Although observed displacements in real volcanoes follow the predicted models reasonably well on a scale of several kilometres, the models break down at smaller scales when fractures develop. The above observations of 1987-8, 2005-6 and 2009-10 show that observed departures from elastic behaviour prior to eruption might be used to determine the most likely location of future fracture development, and therefore of future eruption locations.

Interpretation to date

An initial interpretation of the 2009-2010 data suggests that the northeast sector and the lower northern block could be moving transcurrently in a sinistral direction, but without accumulating much tensional strain. This interpretation is supported by the fact that the Pernicana fault runs between these two sectors. Similarly there is no appreciable increase in tension indicated at the junction between the summit spreading and the southeast sector, since movements at the join would tend to take up most of the strain. However, the line between the northeast and southeast sectors, where tensional strain is accumulating, runs east from the summit, and may represent a future eruptive fissure. The division between the lower northern sector and the summit is also a site of increased tensional strain, and so may represent a future eruptive fissure running northwest from the summit.

Preliminary findings

From the work carried out so far, the most likely position and orientation of the next flank eruption is from a fissure east of the summit, oriented east-west. This has partly been fulfilled in that on 11th January 2011, the first of a series of short-lived but violent paroxysmal pyroclastic eruptions lasting a few hours took place from a new vent at the eastern foot of the southeast crater. Since then there have been 20 such paroxysms, the most recent on 9th February 2012. At the time of writing, strombolian activity has started again within the new vent, where a new cone has built up over the past year now more than 200m high. However, it remains to be seen whether the next flank eruption fissure follows the same trend.

Instrument deployment

Table 1. List of static observations carried out

NERC GEF

This is probably my last loan from NERC GEF, and I would particularly like to thank Alan Hobbs, Paul Kearney and the other personnel working there for their assistance in building up the Etna project over the years, and also the panel that have supported this project since 2001. They have always been helpful, cooperative and prompt, and have at times gone out of their way to make sure that equipment has been delivered in time.

References

1. Murray, J.B. 1988, J.Volcanol. & Geotherm. Res. 35, 121-139. ↩

2. Massonet et al. 1995, Nature 375, 567 - 570. ↩

3. Dowden, J. et al 1995, Tectonophysics 249, 141-154. ↩

4. Dowden, J. et al. 1997, Tectonophysics 269, 299-315. ↩

5. La Delfa, S. et al 2001. J.Volcanol. & Geotherm. Res. 105, 121-139. ↩

6. Bonaccorso, A. et al 2002, Geophysical Research Letters, 29, No. 13, 10.1029/2001GL014397. ↩

7. Murray, J.B. & Pullen, A.D. 1984. Bulletin Volcanologique, 47-4 (2), 1145-1163 ↩

8. Anderson, E.M. 1936, Proceedings of the Royal Society of Edinburgh 56, 128-157 ↩

9. Mogi, K. 1958, Bulletin of the Earthquake Research Institute, University of Tokyo 36, 99-134. ↩

10. Okada, Y. 1992. Bulletin of the Seismological Society of America, 82, 1018-1040 ↩