MVO/VSG - Open Scientific Meeting
27 November 1996

The hydrothermal system of Soufriere Hills volcano,
Montserrat (West Indies): chemical, mineralogical, and microtextural
signatures in fluids, altered rocks, and 1995-96 tephra

G. Boudon1,2, B. Villemant4, J-C. Komorowski1, Ph. Ildefonse3, G. Hammouya5, M. Semet1,4

1Institut de Physique du Globe de Paris, Obs. Volcanol. et Dˇpt de Volcanologie, 4, Place Jussieu, 75252 Paris Cedex 05, France
2IPGP, Dˇpt des Gˇomatˇriaux, CNRS URA 734, 4 Place Jussieu, 75252 Paris Cedex 05, France
3Lab. de Mineralogie et Cristallographie, Univ. Paris VI et VII, CNRS UA S109, et IPGP, 4 Place Jussieu, 75252 Paris Cedex 05, France
4Lab. de Gˇochimie Comparˇe et Systˇmatique, CNRS URA 1758, IPGP et Univ. P. et M. Curie, 4 Place Jussieu, B109, 75252 Paris Cedex 05, France
5Obs. Volcanol. de la Soufriere de Guadeloupe, IPG, 97113, Gourbeyre, La Guadeloupe, F.W.I.

Determination of the physical and chemical characteristics of an active hydrothermal system are fundamental to the understanding of the nature of past and future eruptive processes. Interactions between the hydrothermal system and ascending new magma can modify the physical and chemical characteristics of the magma and host-rock thus influencing eruptive style (e.g. 650 B.P. Mt. Pele eruption; Villemant et al., 1996). Flank-collapse events occur frequently and repeatedly on volcanoes hosting a voluminous hydrothermal system weakening the edifice. Flank-collapse of Soufriere Hills Volcano (Wadge and Isaacs, 1988) produced the horseshoe-shaped English's Crater and previously undescribed debris avalanche deposits observed in the eastern seaside cliffs in 1988-89 and recently in the floor of English's Crater for brief periods (July 1996) as a result of denudation by hot pyroclastic flows and surges associated with dome growth (GVN, 1995, 1996).

Several fumarolic areas and hot-springs were active before 1995. Hot-springs, fumarolic gases, and products of hydrothermal alteration from the active areas of Tar River Soufriere (TRS) and Galway's Soufriere (GS) were sampled in detail in 1988-89 (Boudon et al., 1996) and since then on a semi-regular basis for fluids. Fluids have been analyzed on a collaborative basis at the Soufriere of Guadeloupe Volcanological Observatory. Rocks have been intensely hydrothermally altered (abundant white to subordinant ochre coloration, grey to black near vents or hot-springs) in fumarolic areas characterized by emission of gases (96-98C) saturated in H2O, rich in CO2 and H2S but poor in SO2 and H2, and very acidic condensates (pH< 2). The primary anion in hot-springs (ca. 81 to 36C; 2.4 4.

Primary host-rock minerals at TRS and GS have been transformed into polymorphs of SiO2 (opal-A, opal CT, cristobalite, quartz), titanium oxide (anatase), sulphates (natroalunite: (Na,K)Al2(SO4)2(OH)6, natrojarosite: NaFe3(SO4)2(OH)6 in ochre areas, gypsum, anhydrite). Pyrite is present in dark areas. Clay minerals are absent except as accessory phases at GS (kaolinite, Al-smectite, mixed layer kaolinite-smectite with 75% smectite). Alteration has led to a local reduction of the host-rock porosity. Native sulphur (from oxidation of H2S), alunogen (Al2(SO4)318H2O), halotrichite (FeAl2(SO4)4.22H2O), opal A, and cristobalite occur at the fumarolic vents. Such parageneses are typical of alteration processes of unsealed acid-sulphate systems in the outgrowths of deep hydrothermal systems. Acid-sulphate alteration implies:

  1. leaching of Al and Fe by the acid waters and gas condensates
  2. incorporation of residual Al and Fe in sulphates
  3. crystallization of pyrite from oxidized H2S and Fe released by alteration in areas with lower Eh
  4. precipitation of residual silica as silica polymorphs.

Clays minerals can only form when solution pH is less acidic than at TRS and GS. Fumarolic activity usually produces vertical and horizontal mineral zonation, i.e. siliceous residue (opal-A + quartz) - alunite+kaolinite - kaolinite + smectite - illite/smectite (at depth). Thus mineralogical assemblages at TRS and GS only represent the near-surface alteration zone of a more important hydrothermal system. Primary phases (in wt % from quantitative XRD analysis) in the < 63 micrometers size-fraction of phreatic ashes (Jul. 27-vent 1, Aug. 21-vent 3, Sep. 7, Sep. 17 1995) are: plagioclase (77-82), quartz (17-8), cristobalite (8-2), gypsum (4-0), and pyrite (2-1). Clay minerals are absent. Total silica polymorphs (quartz+cristobalite) contents are high and constant (17- 19 wt %) in all samples. Quantitative XRD determination of crystalline silica content in separates of the respirable fraction (< 5 micrometers) for accurate assessment of human toxicity of the airborne phreatic ash failed on our limited samples.

Castle Peak dome, 1995-96 domes, and 1995 phreatic tephra are chemically similar except for highly mobile elements (Au, As, Br, Mo, W, Ag). Tephra show systematic As, Br, Sb, Mo, W enrichment. Castle Peak dome and tephra have extreme Au contents. Thus tephra represent mixtures of dome magma and material strongly enriched in mobile and volatile elements. Hydrothermal material from TRS show a quasi-systematic strong depletion in all measured elements except highly incompatible Ta, Hf, to a lesser degree Th and U, which are little or unmodified. The enriched component of tephra is clearly distinct from the signature of hydrothermalized material and was likely inherited from deep hydrothermal fluids.

Under the binocular microscope, phreatic tephra (Jul. to Sep. 1995) consist of a heterogeneous mixture of porphyritic strongly to weakly altered (grey, reddish brown, greenish grey) vitreous and crystalline fragments in varying proportions. Minerals are magmatic (plagioclase, quartz, pyroxene, amphibole) and hydrothermal (sulphates, opal, cristobalite, quartz, pyrite) in origin. The most conspicuous characteristic of this tephra is the ubiquitous presence of irridescent to bright metallic-yellow pyrite as isolated subhedral crystals up to 250 micrometers in size or irregular specks on the surface of vitreous fragments. This is the distinctive criterion together with the abundance of hydrothermal phases that fingerprints the non-juvenile origin of this tephra. Vitreous grains in tephra from ash-clouds associated with gravitational destruction of the growing dome are texturally similar to vitreous grains from phreatic tephra (chiefly comminuted Castle Peak dome) but lack hydrothermal alteration phases.

Microtextural and microchemical analysis of polished resin-impregnated grains (1000-125 micrometers) from these samples using the SEM's Backscattered Electron Imaging mode confirm the presence of pyrite as the only abundant sulphide. Porphyritic vitreous grains show pervasive microcrystallization (plagioclase), alteration of the former vitreous matrix, and precipitation of silica (opal, cristobalite, quartz) associated with pyrite and other hydrothermal phases (gypsum, alunite) in most available relict vesicles, cavities, and fractures. Porosity of such clasts is almost nil. Micro-breccia fragments consisting of altered vitric clasts (100-300 micrometers) in a matrix of angular microclasts (10-50 micrometers) cemented by pyrite, banded and hydrated vuggy silica and other alteration phases are common. Similar textures and mineralogy have been described in the 1976 phreatic ashes from Soufriere of Guadeloupe (Heiken et al., 1980), in non-juvenile 1994-95 tephra from Popocatepetl (Siebe et al., in prep), and to a lesser extent in hydrothermally-altered material that surrounded the Mt. St. Helens 1980 dacite cryptodome (Komorowski, 1991; Komorowski et al., 1997). The well-developed hydrothermal textures and mineralization in the 1995 phreatic tephra indicate the presence of an active acid-sulphate hydrothermal system below Soufriere Hills Volcano whose alteration products were well and repeatedly sampled by the 1995 phreatic explosions and mixed with material from the Castle Peak dome complex. SEM studies of samples from Castle Peak dome and 1995-96 dome rocks show similar extensive microcrystallization and alteration textures in the glass matrix with abundant cracked cristobalite infilling voids and former irregular vesicles. However hydrothermal minerals such as sulphates and particularly pyrite are absent from fresh Castle Peaks dome rocks sampled in 1989 and "fresh" 1995-96 dome rocks.

Fine distal ashes (fractions =BE 125 micrometers) from the 17 Sept. 1996 dome explosion collected on Guadeloupe island (85 km) consist of a heterogeneous mixture of vitric porphyritic light grey to whitish fragments variably vesiculated (poorly to reticulitic) with numerous minute black oxide inclusions. They lack surficial pyrite in contrast with 1995 phreatic tephra. Angular fragments of hornblende, plagioclase and or quartz predominate. White hydrothermal minerals (gypsum, silica), and brown to orange clasts with a wavy appearance (natrojarosite ?) are relatively abundant. Pyrite is present in minor amounts and typically with white fragments (silica, opalescent lustre). Leachate analysis show an important sulphate and calcium component and a non-hazardous but marked fluorine enrichment. Distal tephra only provide a limited analysis of the products of the Sept. 17 1996 dome explosion. However these observations clearly indicate that this explosive event involved gas-rich parts of the interior of the dome and the hydrothermal system in the roots of Castle Peak dome and/or Castle Peak-conduit interface. Pervasive silicification (as described at Mount St Helens by Komorowski, 1991, and Komorowski et al., 1997, and Popocatepetl by Siebe et al., in prep.) has drastically reduced the porosity of older dome rock hosting the hydrothermal system and within which magma extrusion has taken place since 1995. This might be the explanation for the apparent lack of a significant leakage of magmatic gases (CO, SO2, H2) into the hydrothermal system as seen in long term monitoring of gas compositions from TRS (within Englishs Crater) and GS (outside Englishs Crater). An alternative hypothesis would see the aquifer of the hydrothermal system as an efficient chemical buffer for magmatic gases released during slow degassing of the rising magma (Hammouya et al., 1996; Young, 1996). The dome explosion of Sept 17-18 1996 (McGuire et al., 1996) and associated pumiceous products indicate that pressurization was maintained locally in the dome/conduit despite fracturing and degassing related to continued growth. Correlation of U-Th disequilibrium data (underway) with mobile element chemistry and microtextural data will better constrain the nature and extent of interaction at the hydrothermal-magma interface and their influence on eruptive style. Studies of the similar well-developed active hydrothermal system of Soufriere of Guadeloupe volcano through systematic long term sampling of the fluids and altered edifice (in debris avalanche deposits) will provide an interesting comparative database to understand the relationships between the hydrothermal system and eruptive processes at Soufriere Hills Volcano.

We thank the staff of the Montserrat Volc. Obs. for their hospitality and sharing information on the ongoing eruption. Logistics in 1995-96 were provided through joint MVO-SRU-BGS monitoring efforts. Supporting research funds came from the French DBT, PNRN and DIPCN programs, the Obs. Volcanol. of IPGP. Samples were kindly provided by MVO staff, W. Ambeh, S. Young, R. Robertson, G. Norton, P. Allard, G. Heiken , J.P. Viode, X. Sole, M. Feuillard, G. Hammouya, Meteo France Guadeloupe.


Boudon, G., Ildefonse, Ph., Komorowski J-C., Semet, M., Villemant, B. (1996) 2nd Caribbean Conference on Natural Hazards and Disasters, University of the West Indies, Jamaica, October 9-12 1996, abstract-poster
Hammouya, G., Allard, P., Clochiatti, R., Jean-Baptiste P., Parello, F., Semet, M. (1996) 2nd Caribbean Conference on Natural Hazards and Disasters,University of the West Indies, Jamaica, October 9-12 1996, abstract
Heiken, G., Crowe, B., McGetchin, T., West, F., Eichelberger, J., Bartram, D., Peterson, R., and Wohletz, K., 1980. Bull. Volcanol. 43:383-395
Komorowski J-C (1991) PhD, Ariz. State Univ., Univ. Microfilms, Ann Harbor, Michigan, 317 pp
Komorowski J-C, Hoblitt, RP, Sheridan MF. (1997) IAVCEI General Assembly, Abstract submitted
McGuire, W.J., Norton, G.E., Sparks, R.S.J., Robertson, R., Young, S.R., Miller, A.D. (1996) MVO, Special Report 01, URL site
Villemant, B., Boudon, G., Komorowski J-C. (1996) EPSL, 140/1-4:259-267
Wadge, G., Isaacs, M.C. (1988) Journal of the Geological Society of London 145:541-551
Young, S. (1996) IAVCEI News 1/2, 2-4.

Montserrat Volcano Observatory