Behaviour of chlorine prior and during the 79 AD Plinian eruption of Vesuvius (southern Italy) as inferred from the present distribution in glassy mesostases and whole-pumices
S. Signorelli et B. Capaccioni, Behaviour of chlorine prior and during the 79 AD Plinian eruption of Vesuvius (southern Italy) as inferred from the present distribution in glassy mesostases and whole-pumices, LITHOS, 46(4), 1999, pp. 715-730
This study deals with the distribution of chlorine in glassy mesostases and
whole-pumices from the 79 A.D. Plinian eruption (Somma-Vesuvius volcanic c
omplex. Italy). This explosive event produced a prominent Plinian fall depo
sit followed by flows and surges. The fall deposit can be divided into two
sub-units on the basis of an abrupt change in colour at approximately mid-l
evel: a phonolitic white pumice layer at the base and a tephriphonolitic gr
ey pumice layer at the top. Due to its hybrid nature (a mixture of k-tephri
tic, tephriphonolitic and phonolitic magmas), information on chlorine behav
iour in tephriphonolitic magma has been inferred only by means of mass bala
nce calculations. In the white pumices, chlorine concentrations show consta
nt whole-pumice values, whereas glassy mesostases display significant compo
sitional variations. These variations have been linked to the cryptocrystal
lisation of leucite in glassy mesostases, which affected the original melt
compositions just before and during the eruptive event. In this framework,
whole-pumice appears to better represent the pre-eruptive melt compositions
. Using chlorine concentration in whole-umices, a three-stage model of chlo
rine behaviour prior and during the eruptive event is predicted. (i) free v
ariation during Rayleigh fractionation, according to a system with variance
greater than zero; (ii) exsolution of volatile chlorine compounds (e.g., m
etal chlorides), when chlorine reaches its solubility limit in silicate mel
t, in coexisting hyper-saline and in dilute immiscible fluids; the variance
of the system is zero at a given temperature and pressure; (iii) residual
syn-eruptive variable enrichment of chlorine in the melt due to cryptocryst
allization of leucite, suggesting a very minor loss of chlorine in the gas
phase during the eruption. Although chlorine does not behave as a volatile
element during the eruption, it is present in the volcanic plume. This is d
ue to the postulated 'excess' fluid phase containing chlorine that formed i
n the magma chamber prior to the eruption. The homogeneous distribution of
chlorine in whole pumices, in contrast with a well-established chemical and
isotopic layering in Vesuvian magmas prior to Pompei eruption, suggests th
at the trace element zonation is not directly linked to chlorine distributi
on in silicic melts. (C) 1999 Elsevier Science B.V. ALI rights reserved.