J. Barclay1, M.D. Murphy1, M.R. Carroll1, R.S.J. Sparks1, A.-M. LeJeune1, J. Toothill2 and R. Macdonald2
1Department of Geology, University of Bristol, Bristol BS8 1RJ, UK
2Environmental Science Division, I.E.B.S., University of Lancaster, Lancaster LA1 4YQ, UK
In order to understand the current eruption of the Soufriere Hills Volcano it is essential that the pre-eruptive conditions (P,T, aH2O) of the magma are quantified. This paper presents the preliminary results of a series of experiments designed to reproduce the observed phenocryst assemblages and mineral reactions in the Soufriere Hills magma. Pre-eruptive dissolved volatiles have also been measured via melt inclusions trapped in phenocryst phases. These data are used to quantify both the depth of the Soufriere Hills magma chamber and the ascent of the magma prior to eruption.
Experimental Phase Equilibria
The Soufriere Hills andesitic magma is highly porphyritic (typically 40 vol%) and contains phenocrysts of plagioclase (plag), orthopyroxene (opx), amphibole (amph), titanomagnetite (mt) and minor quartz (q). Plagioclase phenocrysts exhibit a wide variety of textures and compositions (An65-80), (Murphy et al., this volume), but in general plag microlites are relatively more sodic than earlier grown microphenocrysts. The groundmass is characterised by an anhydrous assemblage with clinopyroxene (cpx) occurring in place of amph. Amphibole phenocrysts usually exhibit some degree of reaction or breakdown. These mineral reactions reflect changing magmatic conditions during magma storage and ascent (Murphy et al., this volume).
Experiments conducted on a natural composition (MON6aR) in the pressure range 50-250MPa (where PH2O = Ptotal) and temperatures from 825-875C° with fO2 varying from NNO to NNO+3 have been used to constrain the conditions at which the original phenocryst assemblage was stable. At PH2O from 150-175MPa and fO2 equivalent to NNO+1, the observed natural phenocryst assemblage (with the exception of quartz) has been reproduced with equilibrium mineral and residual glass compositions close to those found in the Soufriere Hills andesite. At these pressures quartz is projected to become stable at temperatures of ~810±20C°. Amphibole is replaced as an equilibrium mineral phase by clinopyroxene at PH2O<125-150MPa and amph and cpx have been found to co-exist at 150MPa and 875C°. The presence of amph (and absence of cpx) in the Soufriere Hills phenocryst assemblage therefore indicates that the main body of magma resides at PH2O>150MPa. Isothermal experiments conducted at pressures ranging from 50-250MPa show that the equilibrium plag compositions become more albitic with decreasing PH2O (An55 at 250MPa to An40 at 50MPa). The more sodic microlite population can be attributed to their growth during magma ascent.
Pre-eruptive volatiles have been measured in glass inclusions in quartz, plagioclase and orthopyroxene from both pyroclastic flow deposits and pumice from the 17th September explosive event. Quartz-hosted melt inclusions, measured by FTIR, have water contents of ~4±0.5wt% H2O, and CO2 <50ppm. Cl and S concentrations are 2400±400 and <150ppm respectively. No significant difference in pre-eruptive volatiles has been found between inclusions measured in the earlier pyroclastic flows and the 17th September explosive event. Comparison of these values with H2O concentrations measured in glasses from the experimental runs suggests that these concentrations of water are equivalent to PH2O of ~100-125MPa; a value slightly less than that suggested by the experimental phase equilibria. This may suggest that the magma chamber was water-undersaturated prior to the injection of the basalt magma. Comparison of concentrations of S and Cl dissolved in the melt inclusions and SO2 and Cl measured in the volcanic plume (MVO Scientific Reports) suggest that the degassing of the ascending andesite magma cannot be solely responsible for the SO2 flux. Another source must be found for the measured SO2; it would seem unlikely that this is the result of the presence of an excess vapour phase prior to eruption.