Éditeur(s) : HAL CCSD
Résumé : International audience
Abstract A careful re-examination of the well-known written documents pertaining to the 2,750-year-long historical period of Mount Etna was carried out and their interpretation checked through the high-accuracy archeomagnetic method (>1,200 large samples), combined with the 226Ra-230Th radiochronology. The magnetic dating is based upon secular variation of the direction of the geomagnetic field (DGF) and estimated to reach a precision of  ±40 years for the last 1,200 years, and ±100 to 200 years up to circa 150 B.C. Although less precise, the 226Ra-230Th method provides a unique tool for distinguishing between historic and prehistoric lavas, which in some cases might have similar DGFs. We show that despite the abundance of details on ancient historical eruptions, the primary sources of information are often too imprecise to identify their lava flows and eruptive systems. Most of the ages of these lavas, which are today accepted on the geological maps and catalogues, were attributed in the 1800s on the basis of their morphology and without any stratigraphical control. In fact, we found that 80% of the “historically dated” flows and cones prior to the 1700s are usually several hundreds of years older than recorded, the discrepancies sometimes exceeding a millennium. This is proper the case for volcanics presumed of the “1651 east” (actually ∼1020), “1595” (actually two distinct flows, respectively, ∼1200 and ∼1060), “1566” (∼1180), “1536” (two branches dated ∼1250 and ∼950), “1444” (a branch dated ∼1270), “1408” (lower branches dated ∼450 and ∼350), “1381” (∼1160), “1329” (∼1030), “1284” (∼1450 and ∼700), “1169 or 812” (∼1000) eruptions. Conversely, well-preserved cones and flows that are undated on the maps were produced by recent eruptions that went unnoticed in historical accounts, especially during the Middle Ages. For the few eruptions that are recorded between A.D. 252 and 750 B.C., none of their presumed lava flows shows a DGF in agreement with that existing at their respective dates of occurrence, most of these flows being in fact prehistoric. The cinder cones of Monpeloso (presumed “A.D. 252”) and Mt. Gorna (“394 B.C.”), although roughly consistent magnetically and radiochronologically with their respective epochs, remain of unspecified age because of a lack of precision of the DGF reference curve at the time. It is concluded that at the time scale of the last millennia, Mount Etna does not provide evidence of a steady-state behavior. Periods of voluminous eruptions lasting 50 to 150 years (e.g., A.D. 300–450, 950–1060, 1607–1669) are followed by centuries of less productive activity, although at any time a violent outburst may occur. Such a revised history should be taken into account for eruptive models, magma output, internal plumbing of the volcano, petrological evolution, volcano mapping and civil protection.
Bulletin of Volcanology / Bull Volcanic Eruptions

hal-00315967
https://hal.archives-ouvertes.fr/hal-00315967
DOI : 10.1007/s00445-007-0121-x