Lava flows fed by a lateral eruption are seen spilling down the southwestern slope of Valle del Bove in December 1991. Photo by Carmelo Monaco

Mount Etna erupts both from the summit craters and from vents away from the summit craters, which are generally considered flank vents. It has been correctly observed that some activity on the flanks appears to be closely related to the summit craters, and thus to the central conduit system. In these cases, flank activity is almost always preceded by increased summit activity by days to months, and summit activity ceases once flank eruption has begun, or changes from Strombolian-effusive to strong degassing, often carrying some ash. Flank vents may form, in this type of eruption, very close to the summit and practically "take over" their activity, as has occurred in February 1999. This type of activity, which may continue for months or even years, is the classical "subterminal" eruption as defined by the volcanologist Rittmann, and is nothing else than some kind of summit activity, even though occurring from vents outside the summit craters. The output is generally fairly low (from less than one cubic meter per second to a few cubic meters per second) and is considered to represent the mean rate at which magma is ascending to the surface from depth. It should be noted, though, that this magma ascent rate is not all too regular.

"LATERAL" ERUPTIONS
The opening of vents still further downslope, but in a manner extremely similar to that of the "subterminal" eruptions, leads to what has been named "lateral" eruptions by Rittmann and other authors. As in the "subterminal" eruptions, the activity is very closely related to the central conduit system, and degassing frequently takes place through one or more of the summit craters rather than from the flank vents themselves (e.g., the 1989 eruption). The fissures of these eruptions are arranged more or less radially to the summit, but are strongly controlled by the tectonically induced fracture systems within the volcano, most notably the east-northeast and southeast systems.
The output during these eruptions is considerably higher than that of "subterminal" activity, especially during their initial stages. The higher output results mostly from the pressure of the hydrostatic head - that is, the magma still present in the central conduit system above the elevation of the new eruptive vents - but to some degree may also reflect the fact that magma rises easier to the lower elevation than to the summit area. As long as magma continues to drain from the central conduit system above the flank eruptive vents, the output remains quite high (considerally above the average of about 10 cubic meters per second for flank eruptions) while it drops to average levels or below once the level of magma in the central conduit system has sunken to the level of the active vents.
It is very difficult to make a clear distinction between "subterminal" and "lateral" eruptions because there is an uninterrupted gradation between both end members and in their characters (apart from the eruption rate), they are essentially identical. This difficulty becomes apparent in the case of the 1975-1977 effusive activity on Etna's north flank, often cited as a typical "subterminal" eruption (it was clearly related to the NE Crater with which it had alternating phases of activity). The eruptive vents lay at elevations as far down as 2600 m, not much higher than the March-July 1985 fissure (2600-2500 m) on the south flank. However, the effusion rate was very low (less than one cubic meter per second). So was this a flank eruption, as Murray (1990) and Azzaro and Neri (1992) believe, or is it rather to be considered an extension of the persistent summit activity at NE Crater, initiated in 1974? If the effusion rate is taken as the decisive parameter, then the 1975-1977 activity is summit eruption rather than flank eruption.

Dike intrusions near Valle del Bove since 1978 This map shows the traces of dikes emplaced on the upper southeastern flank of Etna between 1978 and 1991, as inferred from deformation measurements. The 1978-1979 dikes have been projected from the eruptive fissures of the four eruptive episodes during those years, no precise geophysical data about the position of the dikes are available. Note that all dikes begin at the Southeast Crater (SE Crater). The map originally appeared in McGuire et al. (1997) and has been slightly modified.

What causes such eruptions? There are two main reasons for these eruptions to occur. One is the response of the volcano and its fracture systems to variations in the volume of magma rising within the central conduit system. The uprise of a voluminous "batch" of magma, especially when gas-rich, may notably increase the hydrostatic pressure on the conduit walls and cause the mountain to fracture radially, in a process where magma injection into the flanks is active and fracturing is passive. The other process which may lead to "lateral" erputions is a change in the structural stability of the volcano, caused by tectonic movements. A decrease in the structural stability of the volcano may generate fractures at depth into which magma can intruce more or less passively, and there it may remain for weeks before it erupts, as supposedly occurred in 1991, before the 1991-1993 flank eruption.
The intrusion of magma into fractures radiating away from the central conduit system forms dikes (or dykes) which do not necessarily have to reach the surface to produce eruptions, but may also remain at depth and cool. It has been suggested by some authors that dikes may actually serve as temporary storage areas, for periods of up to many years, but Murray (1990) rejected such hypotheses based on the assumption that in a dike, magma cools relatively rapidly and therefore cannot remain available for eruption for more than a few months.
It is, however, well possible that in some areas within the mountain, close to the central conduit system, the environment is so hot that it would allow a magma much longer to remain fluid in a dike than close to the surface and farther away from the central conduit system.

The evolution of a "lateral" eruption from the initiation of dike intrusion to the arrival of magma in a newly opening eruptive fissure has been documented extraordinarily well in 1983 on Etna's south flank (Murray and Pullen 1984). The path of the subterranean dike, as it propagated towards the eruption site, could be followed by monitoring the ground deformation (a levelling network was set up in the area by John Murray and his colleagues months before, because an eruption was considered likely to occur there in the future). It was striking to see that, although there was no activity at the SE Crater before and during the 1983 eruption, the dike originatted from a point very close to that crater, at a depth of about 1 km. For the first kilometer, the path of the dike ran parallel to the fractures of 1978 and 1979 eruptions, then, east of Torre del Filosofo, it split into two branches, one continuing straight southeastwards while the other turned southwestwards, and then, still another kilometer downslope, southwards. The southeastern branch turned south at Belvedere and ended somewhere to the south without reaching the surface.
The western branch eventually passed below the Piccolo Rifugio (the building was traversed and heavily damaged by the non-eruptive part of the fracture) and little further downslope the top of the dike intersected the surface, and an eruptive fissure became active.

Although until 2001 the path of a feeder dike for a flank eruption was never again traced in as much detail as in 1983, geophysical monitoring has allowed the documentation of further intrusions in other cases thereafter. All flank eruptions during that period through 2001 thus seem to originate from the SE Crater, only one intrusion which did not result in a flank eruption, in September 1986, originated at the NE Crater (Murray 1990). Some were, in fact, characterized by the slow draining of the central conduit system (1983, March-July 1985, 1986-1987, 1991-1993) while others were caused by the fracturing of the volcano under the hydrostatic pressure of a newly arrived, voluminous and gas-rich batch of fresh magma (e.g., 1989). In 2002, eruptive activity on the Northeast Rift was again closely related to the NE Crater, while activity on the S flank during the same eruption was completely independent of the central conduit system.

A typical eccentric eruption at Etna occurred in early 1974 on the western flank of the volcano. These amateur photos, taken by Renato Bernardini (University of Catania), show a small cinder cone growing by intense Strombolian activity amidst a forest, February 1974.

"ECCENTRIC" ERUPTIONS
Eccentric eruptions are thought to be the result of dikes rising vertically from depth to the surface, which are not directly related to the central conduit system. Such eruptions are quite rare and therefore only two - in 2001 and in 2002-2003 - have been documented in much detail by modern geophysical monitoring techniques. Previous eruptions of this type occurred in 1974 on the western flank, forming two small cones named Monti De Fiore, and in 1892, when the Monti Silvestri craters came to existence. Many of the more remote pyroclastic cones on the flanks of Etna seem to be the products of such eruptions. In the case of the 1974 Monti De Fiore, the 1892 Monti Silvestri and 1763 Montagnola eruptions it is well documented that the summit activity was in no way affected by the flank eruptions; in July 1893 incandescent lava was seen flowing within the Central Crater only 7 months after the end of the voluminous Monti Silvestri eruption.
Eccentric eruptions are generally characterized by a much higher degree of explosivity and thus build pyroclastic cones of appreciable dimensions. Unlike "lateral" eruptions during which most degassing occurs though the summit craters, many eccentric eruptions are not accompanied by similar summit activity, and sometimes they are neither preceded by summit activity (e.g., 1892); in other cases summit activity that has begun before an eccentric eruption continues with little variation during and after that eruption, as in 1974. However, the 2001 and 2002-2003 eruptions were notable exceptions because in this case summit, lateral and eccentric activity occurred simultaneously. It seems that the uprise of an eccentric dike destabilized the volcanic edifice to the point that the opening fracture system affected the central conduit system, triggering a lateral eruption. This does not seem to be the case in most eccentric eruptions.
While eccentric eruptions may possibly occur in any spot on the volcano (that is, in all the area between its very base and the summit), eccentric vents are nearly always aligned along the same, essentially subradial, trends as lateral eruptive fissures. This is due to the fact that the uprise of eccentric dikes is facilitated along lines of structural weakness, which may be of regional tectonic origin or related to the more surficial volcano-tectonic stress field. Thus the eccentric vents of the 2001 and 2002-2003 eruptions on the south flank were formed on the same trend as some previous lateral eruptions as those of 1983 and March-July 1985.

In spite of the enormous amount of new data provided by the 2001 and 2002-2003 eruptions, the problem of eccentric eruptions remains full of mysteries. Do the conduits feeding such eruptions originate in a rather "individual" manner from the deep storage area, or do they depart, still at considerable depth, from the roots of the central conduit system? And which of the eruptions of the past centuries were eccentric, which were lateral? Is there a reliable method allowing to distinguish whether past eruptions were eccentric or lateral? Possibly the explosivity of those eruptions, which can be read from the size of the related cones, is an important indicator, and the chemical composition of their products may be another. Research in this field is not only of academic interest, but has important implications for hazard assessments at Etna. If eccentric eruptions are always as explosive as those of 2001 and 2002-2003, they represent a significant hazard to human infrastructures and traffic. It is therefore necessary to understand how frequent such eruptions are in the long and short term. During the 110 years before the 2001 eruption there had been only one single eccentric eruption, which furthermore was studied by an earlier generation of scientists, and before most of the current concepts of Etna's dynamics were developed. At the beginning of the new millennium eccentric eruptions thus received very little - if any - attention and were not even mentioned in most publications of the 20 years preceding the 2001 eruption.
In the opinion of researchers of the 19th century, some of the remote flank cones of Etna - such as the isolated Monte Moio cone complex on the northern flank, and Monte Barca immeidately south of Bronte, on the western flank - were true volcanoes on their own, a view that is still maintained by many inhabitants of villages on Etna's flanks close to the more remote of the flank cones. During a recent visit to the town of Randazzo, I talked with residents who recalled the dramatic 1981 north flank eruption, which seriously threatened their home town, as "a new volcano that exploded just over here" - and the 1981 eruption was not even an eccentric one! Anyhow, there are no other individual volcanoes than Etna in the region, and all cones are clearly products of Etnean volcanism. They merely represent a further complexity in the outstandingly complicated picture named "ETNA".

Copyright © Boris Behncke, "Italy's Volcanoes: The Cradle of Volcanology"