International audienceA new, pre-Green Tuff (46 ka) volcanic stratigraphy is presented for the peralkaline Pantelleria Volcano, Italy. New 40Ar/39Ar and paleomagnetic data are combined with detailed field studies to develop a comprehensive stratigraphic reconstruction of the island. We find that the pre-46 ka succession is characterised by eight silica-rich peralkaline (trachyte to pantellerite) ignimbrites, many of which blanketed the entire island. The ignimbrites are typically welded to rheomorphic, and are often associated with lithic breccias and/or pumice deposits. They record sustained radial pyroclastic density currents fed by low pyroclastic fountains. The onset of ignimbrite emplacement is typically preceded (more rarely followed) by pumice fallout with limited dispersal, and some eruptions lack any associated pumice fall deposit, suggesting the absence of tall eruption columns. Particular attention is given to the correlation of well-developed lithic breccias in the ignimbrites, interpreted as probable tracers of caldera collapses. They record as many as five caldera collapse events, in contrast to the two events reported to date. Inter-ignimbrite periods are characterised by explosive and effusive eruptions with limited dispersal, such as small pumice cones, as well as pedogenesis. These periods have similar characteristics as the current post-Green Tuff activity on the island, and, while not imminent, it is reasonable to postulate the occurrence of another ignimbrite-forming eruption sometime in the future

Pantelleria volcano has a particularly intriguing evolutionary history intimately related to the peralkaline composition of its explosively erupted magmas. Due to the stratigraphic complexity, studies over the last two decades have explored either only the pre-Green Tuff ignimbrite volcanism or the post-Green Tuff activity. We here focus on the whole evolutionary history, detailing the achievements since the first pioneering studies, in order to illustrate how the adoption and integration of progressively more accurate methods ( 40 Ar/ 39 Ar, paleomagnetism, petrography, and detailed field study) have provided many important independent answers to unresolved questions. We also discuss rheomorphism, a distinct feature at Pantelleria, at various scales and possible evidence for multiple, now hidden, caldera collapses. Although the evolutionary history of Pantelleria has shown that each ignimbrite event was followed by a period of less intense explosivity (as could be the present-day case), new geochronological and geochemical data may indicate a long-term waning of volcanic activity.

The eruptive history of Pantelleria has been marked by the eruption of nine peralkaline ignimbrites, with inter-ignimbrite episodes from small, local volcanic centres. New whole-rock geochemical data are presented for seven ignimbrites and used with published data for younger units to track compositional changes with time. From ∼190 ka, silicic magmatism was dominated by comenditic trachyte to comendite compositions, evolving along generally similar liquid lines of descent (LLOD). The final ignimbrite, the Green Tuff (∼46 ka), was tapped from a compositionally zoned pantelleritic upper reservoir to a trachytic mush zone. Younger (20-7 ka) silicic magmatism has been relatively small scale, with compositions similar to the earliest pre-Green Tuff pantelleritic ignimbrite (Zinedi). These data suggest that the comenditic reservoirs may have been emplaced at deeper levels than the pantelleritic reservoirs. While both types of series evolved along similar LLOD dominated by fractionation of alkali feldspar, it is the fractionation of iron that determines whether comendite or pantellerite is produced. The deeper reservoirs were more oxidizing and wetter, thus leading to the crystallization of magnetite and therefore the fractionation of iron.

No abstract

No abstract