Abstract
Abstract The Soufrière Hills Volcano on Montserrat has been in eruption since 1995. Before then, a number of hot springs and fumaroles high on the volcanic edifice attested to the presence of hydrothermal circulation systems. All of these features were buried beneath thick deposits of volcanic ash during the early years of the eruption. However, examination of low-lying sites in the vicinity of the abandoned town of Plymouth has revealed the persistence of a zone of hot (≤64 °C) groundwater discharge in a narrow valley that opens onto the northern beaches of the town. This small valley has not been significantly affected by ash deposition, with any airfall deposits having been subsequently washed downstream during periods of rain. The seepages and small springs in this zone all emit Na–Cl groundwaters, which appear from the concentrations of major ions and B and Li to be hydrothermally altered seawaters that have undergone extensive rock–water interaction at depth. Taken together with previously published volcanological observations, the persistence of this hydrothermal system throughout the sustained eruption period suggests that a robust hydrothermal circulation system exists at depths of several hundred metres in an area that closely adjoins the still-inhabited half of the island. Comparison with similar volcanogenic hydrothermal systems in the Caribbean and elsewhere suggests that there is significant potential for development of high-enthalpy geothermal energy resources in this vicinity. Using a portable well-head generating unit, a single successful borehole in this area could reasonably be expected to yield up to 5 MW of electrical power. This output significantly exceeds the current peak electricity demand on the island (c. 3.5 MW). In concept, it would be possible to exploit any excess capacity to operate an electrolysis plant to produce hydrogen for use in vehicles on the island. Establishing energy self-sufficiency would be a major step towards the economic and social regeneration of this devastated island, with the further possibility of increasing prosperity in the future through export of surplus energy to nearby non-volcanic islands.
This paper is one of a thematic set entitled Hydrogeology and Heat Engineering. Most of the papers in this set relate to thermogeology and the exploitation of low-enthalpy ground source heat in the UK (Banks 2009; Fry 2009). In contrast, this paper describes a high-enthalpy geothermal system in a British Overseas Territory: the small (106 km2) island of Montserrat, which lies in the westerly (‘Leeward’) volcanic island arc of the Lesser Antilles, East Caribbean. Montserrat ( Fig. 1) is effectively an amalgamation of three volcanic centres (Rea 1974), two extinct and one currently active. The most northerly of the extinct volcanoes corresponds to Silver Hill, where extrusive activity has been dated (Harford et al. 2002) to the late Pliocene to early Quaternary (2.6–1.2 Ma). The middle of the island is occupied by deposits of a volcano that was active in the early to middle Quaternary (0.95–0.55 Ma), corresponding to the present-day Centre Hills (Fig. 1). The southern half of the island is entirely dominated by the active Soufrière Hills Volcano. From the first recorded settlement of the island in 1632 until mid-1995, this volcano had never been known to erupt, and was generally assumed to pose little or no threat to the island's population. Hence the principal town of Montserrat, Plymouth, prospered and expanded over the lowermost western slopes of the volcanic edifice of the Soufrière Hills. Like the rest of the island, the Soufrière Hills were verdant, with lush pastures on the lower slopes and dense woodland at higher elevations. Taken together with the fact that the largest contingents of early European settlers were Irish people exiled in the 17th century by Cromwell, this verdance earned Montserrat the sobriquet of ‘Emerald Isle of the Caribbean’.
Topographic map of Montserrat Island, showing locations referred to in the text. HWP, Hot Water Pool, as discussed in Section 4 and Table 1. The shaded area shows the extent of the daytime access zone (DAZ) during the fieldwork period. Adapted from Druitt & Kokelaar (2002) with the permission of the Geological Society.
Hydrochemistry of the Hot Water Pool site (grid ref. [75164776]) on Montserrat
This green idyll was rudely shattered on 18 July 1995, with the onset of a sustained period of violent eruption, which has since continued to become one of the most enduring eruptions ever recorded of a supra-Benioff zone andesite volcano (S. C. Loughlin, pers. commun.). The eruption has been extensively documented (e.g. Druitt & Kokelaar 2002; Edmonds et al. 2003; Barclay et al. 2007; Le Friant et al. 2009), and intensive monitoring by the Montserrat Volcano Observatory (MVO) (which maintains detailed records of the eruption online at www.mvo.ms) has provided the basis upon which volcanologists from around the world have been able to substantially advance scientific understanding of the dynamics of a range of volcanic phenomena, including pyroclastic flows (e.g. Cole et al. 2002), lahars (e.g. Barclay et al. 2007) and tephra fallout (e.g. Bonadonna et al. 2002).
Although the 1995 onset of the eruption of the Soufrière Hills Volcano came as a shock to islanders, and as a surprise even to volcanologists, there had in fact been a number of prior indications of the potential for a return to extrusive activity (Druitt & Kokelaar 2002). In addition to scant records of earthquakes in 1897–1898 and 1932–1937, a well-documented seismic crisis in 1966–1967 was interpreted as heralding renewed magma movement at depth. In the event, the seismic crisis of 1966–1967 passed without any eruptive activity. However, a comparable but more intense seismic crisis from 1992 to 1995 preceded the onset of lava and ash extrusion in July 1995.
A more persistent indication of the continuing potential for eruptive activity was afforded by ‘one of the most active hydrothermal systems amongst Lesser Antilles volcanoes, with widely distributed hot springs and fumarolic vents’ (Boudon et al. 1998). Four major ‘soufrières’ (as hot springs and/or fumaroles are known in these islands) formed esoteric visitor attractionsprior to 1995. These were the Tar River Soufrière, the Galway's Soufrière, and the Upper and Lower Gages Soufrières (see Fig. 1 for locations). Although deposition of pyroclastic and epiclastic material during the current eruption has buried all four of these soufrières, at least for the time being, there is no a priori reason to assume that this means that hydrothermal circulation has diminished; indeed, evanescent steam emissions and ephemeral hot springs and pools are well attested in the areas of active lava extrusion (see MVO records at www.mvo.ms).
The human suffering caused by the eruption of the Soufrière Hills Volcano has been substantial. Within a year of the start of the eruption, the town of Plymouth had been entirely evacuated, and most of it has since disappeared beneath thick airfall deposits, which have ‘turned a green landscape grey’ (Pattullo 2000). Before the eruption, the island had a population of around 10 000. Although many people were relocated to emergency accommodation in the north of the island, up to two-thirds of the population eventually left for the safety of London and various cities in North America. Conditions in the cramped emergency shelters in the north of the island were poor, and some people were tempted to return temporarily to their homes near the edge of the areas of intense ash deposition. This practice contributed to an appalling tragedy on 25 June 1997, when 19 people were killed by a pyroclastic flow that over-spilled valley walls that had previously accommodated similar flows (Pattullo 2000).
An exclusion zone in the southern half of the island has successfully prevented similar tragedies. The extent of this exclusion zone has varied over the years in tandem with the violence of the eruption and the orientations of pyroclastic flows and lahars. At present, the exclusion zone corresponds to the land to the south of the Belham River (which drains to the west coast north of Plymouth) and a lookout point near Jack Boy Hill on the east coast (Fig. 1). During periods of relative quiescence, limited daytime access is permitted to a small area extending from the Belham River into central Plymouth. This ‘daytime access zone’ (DAZ) was the closest area to the volcano in which observations reported in this paper were made. It should be noted that the DAZ shown in Figure 1 is that which was current during the field work reported here; although the DAZ has usually included that area, its extent varies according to advice given by the MVO to the Government of Montserrat concerning the current level of risks posed by the volcano.
The work reported here was planned and undertaken with a view to attempting to derive some benefit for the people of Montserrat from the volcano, which has otherwise been such a negative force in their lives since 1995. One obvious possibility would be geothermal energy. Encouragement is drawn from a range of promising findings on the nearby islands of Dominica (Demange et al. 1985), St. Lucia (Williamson & Wright 1978; Goff & Vuataz 1984), and Martinique (Sanjuan et al. 2005). The most encouraging analogue of all is found on the neighbouring island of Guadeloupe, where the Bouillante Geothermal Field on Basse Terre has been gradually developed from a 5 MW source with a single borehole in the early 1970s, to the present system comprising three production boreholes linked to a common generating plant rated at 15 MW. It is noteworthy that the natural hydrothermal system of the Bouillante Geothermal Field, which has been described in detail by Brombach et al. (2000), is closely analogous to the hydrothermal system of Montserrat previously described by Boudon et al. (1998). These analogies provide abundant grounds for confidence in the suitability of the hydrothermal system of the Soufrière Hills Volcano of Montserrat as a serious prospect for geothermal energy.
This paper presents the findings of a reconnaissance investigation of this possibility, which has been undertaken by a combination of walkover surveys and water sampling (necessarily restricted to the inhabited part of the island and the DAZ), supplemented by an applied reinterpretation of scientific data relating to the Soufrière Hills Volcano, which were originally collected for other purposes. (Grid references in the text that follows refer to the UTM-based grid printed on maps published with the sanction of the Government of Montserrat.) Besides providing specific insights into the geothermal prospects on Montserrat, it is hoped that the findings of this study may prove useful for those planning similar reconnaissance studies elsewhere in the world, as the search for low-carbon energy sources intensifies in response to grave concerns over global warming.
Evidence from hydrogeological setting
Working from a collation of data on aquifer properties and spring yields from numerous small volcanic islands in the Eastern Caribbean, Robins et al. (1990) established a generalized classification of the predominant hydrogeological terrains found in this region. Within this classification, the Soufrière Hills Volcano falls into the category that Robins et al. (1990) identified as being most favourable for the presence of significant bodies of groundwater; that is, ‘abundant upland rainfall, coinciding with widespread Pleistocene and younger pyroclastics [which] ensures high water tables even at high elevations’. Taken together with the pre-eruption evidence of the existence of a vigorous hydrothermal circulation system on and around the Soufrière Hills Volcano (Boudon et al. 1998), this regional framework of Robins et al. (1990) is highly supportive of the notion that significant groundwater circulation is likely to be taking place within the volcanic edifice.
Given that it is currently difficult to access much of this edifice safely, it is worth considering whether the openly accessible flanks of the Centre Hills and Silver Hill might also be good prospects for geothermal exploration. No surface manifestations of active hydrothermal circulation have been found to the north of the line of valleys that separate the Soufrière Hills from the Centre Hills. Although it is important not to construe absence of evidence as evidence of absence, the lack of thermal springs or fumaroles in the Centre Hills is not encouraging. Nevertheless, the early to middle Quaternary pyroclastic deposits of the Centre Hills certainly retain significant permeability, as manifest in the flow rates of numerous non-thermal springs (such as that at Killiecrankie in the Waterworks Valley), which provide the public water supply for the inhabited part of the island (Davies & Peart 2003). Source protection considerations for these vital springs would be a powerful argument against geothermal exploration in their vicinity, even if there were favourable hydrothermal manifestations in the vicinity. The outlook with regard to Silver Hill is even less encouraging. Robins et al. (1990) have noted the relatively low permeability of Tertiary volcanic rocks in the Lesser Antilles, and the older volcanic rocks of Silver Hill are of Tertiary age. Furthermore, Silver Hill is relatively low-lying (c. 400 m above sea level (a.s.l.)) when compared with the Centre Hills (≤730 m a.s.l.) and the Soufrière Hills (≤1000 m a.s.l.); this in turn means that Silver Hill receives rather less orographic rainfall than the higher hills to the south. The lower rainfall rates on Silver Hill are reflected in the rather sparse scrub vegetation on its uncultivated slopes (see Davies & Peart 2003).
The South Soufrière Hills are not safely accessible at present, and are therefore not a realistic geothermal exploration target for the foreseeable future. However, it is probably worth noting that both the relative juvenility (c. 0.13 Ma; Harford et al. 2002) and mafic composition of the volcanic rocks of the South Soufrière Hills are suggestive of relatively high permeability (see Robins et al. 1990; Wood & Fernández 1988). Their proximity to shallow magma in the currently active volcano immediately to their north means that the South Soufrière Hills occupy a conceptual position between that of the Centre Hills and the Soufrière Hills Volcano, and as such may well be a rational target for geothermal exploration in the (still distant) future.
Evidence from seismic activity
Seismic emissions surveys have long been used in geothermal exploration to identify possible deep-seated hydrothermal activity (e.g. Butler & Brown 1978; Oppenheimer & Iyer 1979; Katz 1984). Obviously, around a highly active volcano such as that of the Soufrière Hills, seismic activity is predominantly due to magmatic movement and associated deformation of the enclosing rock mass. However, insofar as clusters of earthquakes attest either to magmatic ascent or to active brittle deformation of pre-existing rocks, such seismic activity is still recording processes that are relevant to the geothermal prospector's search for very hot water at depths of a few hundred to a few thousand metres.
Unsurprisingly, the majority of volcano-tectonic (VT) earthquake hypocentres detected during the current eruption of the Soufrière Hills Volcano have been located directly below the central volcanic edifice. However, a significant swarm of VT hypocentres was also detected beneath St. George's Hill (Aspinall et al. 1998). St. George's Hill, which lies NW of the active volcano in the current DAZ (Fig. 1), is a rather enigmatic feature. Aspinall et al. (1998) suggested that it may be an ‘inactive volcano’, whereas others have considered that it may be an ancient ‘cryptodome’ (S. C. Loughlin, pers. commun.); that is, a dome-shaped landform created by accumulation of viscous magma at shallow depth. Wadge & Isaacs (1988) pointed out that the domes of the Soufrière Hills, St. George's Hill and Garibaldi Hill (Fig. 1) are aligned along a zone trending WNW, possibly representing a zone of crustal weakness at depth. The same linear feature is believed to extend beneath the sea to the neighbouring volcanic island of Nevis (Chiodini et al. 1996). Aspinall et al. (1998) argued that the VT earthquakes beneath St. George's Hill in 1995–1996 are unlikely to have resulted from direct magma intrusion beneath the hill itself. Rather, they probably represent the response of the NW–SE stress field to the ascent of magma below the main volcanic dome of the Soufrière Hills nearby. Aspinall et al. (1998) commented that the earthquakes probably involve ‘a fault or faults which may have been in a state of incipient failure to begin with’. Subsequent gravity surveys did indeed reveal a linear NW–SE-trending feature, probably a fault, running between Garibaldi Hill and St. George's Hills (S. C. Loughlin, pers. commun.). Reactivation of faults beneath the hill in 1995–1996 is likely to have led to the enhancement of any hydrothermal circulation in its vicinity. Indeed, Kilburn & Voight (1998) argued that fracture development prior to the onset of magmatic activity in the Soufrière Hills Volcano was specifically due to ‘stress corrosion’, in which ‘circulating juvenile and hydrothermal fluids chemically attack the country rock and promote failure at stresses smaller than the rock's theoretical strength’. This in turn suggests that the eruption may well have enhanced the permeability and storativity of the geothermal reservoirs on the island. It is eminently possible that applied reinterpretation of existing seismic data gathered over the last decade (especially those from periods during which the Soufrière Hills Volcano was relatively quiescent) might yield further clues about the location and vigour of hydrothermal circulation beneath St. George's Hill.
Field observations made by the author on St. George's Hill suggest that recent seismic activity may have been associated with localized differential uplift. The destruction of the road bridge at grid ref. [76284776] is wholly ascribable to runoff and debris flows originating in the large western-facing gully (centred on point [773480]), which cuts the westernmost flank of St. George's Hill. Occasional torrential rainfall is not difficult to explain in this area, given the frequency of hurricanes and the localized increase of precipitation rates as a result of nucleation on airborne fine ash. However, inspection of this gully from above reveals it to have a very small upland catchment source area, of only a few tens of square metres. (It is considerably smaller, for instance, than the catchment of the Sand Ghaut, which drains the northwestern aspect of St. George's Hill; yet Sand Ghaut does not show anything like the same degree of erosion and sediment mobilization.) In appearance ( Fig. 2) the gully appears greatly over-steepened, with slopes far out of equilibrium. Standing above the large western-facing gully on a still, dry day, the sound of toppling rocks was continually audible. Given that the amount of water that drains down this gully cannot even equal that draining down the comparatively undisturbed Sand Ghaut nearby, it is not unreasonable to postulate that the current contrast between the two in both appearance and geomorphological response might be due to neotectonic uplift of the southern flank of the large western-facing gully.
View of the southern flank of the large west-facing gully on St. George's Hill, showing the effects of extreme recent erosion, which are difficult to explain on hydrological grounds alone. In the background a steam plume rising from the sea near Plymouth pier should be noted. Steam plumes are frequently seen in this area and are considered indicative of localized hydrothermal outflows into the sea.
Evidence from hydrothermal manifestations at surface
Soufrières and associated hot springs high on volcanic edifice
The very name of the Soufrière Hills Volcano emphasizes the impression that the four major zones of fumarolic and hot springs activity made on the early settlers of Montserrat. Three of these four soufrières (Fig. 1) emerged at around 500 m a.s.l., whereas Gages Lower Soufrière emerged between about 300 and 370 m a.s.l. All of the soufrières gave rise to vigorous gas emissions. High-temperature aqueous discharges were also found in close proximity to the gas vents at all of the soufrières except Tar River, in which case the nearest warm (42 °C) springs (apparently overlooked by Atkinson et al. 2000) were located about 1 km downstream of the fumarole field (Chiodini et al. 1996). In addition to the four major soufrières, a fifth zone of hydrothermal emanation also existed in the valley of the Hot River. Rea (1974) reported the existence of hot springs in the uppermost reaches of this valley at around 700 m, but this locality has not been safely accessible since July 1995. In March 1996, hot springs (50 °C) were found at an (unspecified) lower elevation in the same valley (Atkinson et al. 2000).
Tantalizing references to abrupt increases in temperature at some of the soufrières during earlier seismic crises have been reported (Hammouya et al. 1998): during 1934 at Lower Gages Soufrière, and in 1967 at Galway's. However, the temperatures of the soufrière emanations (gas and water) are reported to have remained virtually constant, in the range 96–99 °C, since then, right up to the last measurements at the start of the present eruption period (Hammouya et al. 1998). All four of the major soufrières have since been deeply buried beneath pyroclastic flows, ashfall tuff and/or lahars. Remotely observed pyroclastic flows have probably also buried the hot springs previously observed in the Hot River valley. Although Atkinson et al. (2000) reported some indications of re-establishment of the Galway's Soufrière a few weeks after its initial burial by pyroclastic flows, later burial appears to have put an end to any resurgence for the time being. At the time of writing, none of the soufrières high up on the volcanic edifice are extant (S. C. Loughlin, pers. commun.).
Before the soufrières were buried, gas emission records suggested little change in the hydrothermal dynamics of the Gages Soufrières during the eruption of the volcano. This is despite the fact that the first few weeks of the eruption (from 18 July to 25 September 1995; Gardner & White 2002) were dominated by explosive venting of superheated steam (c. 720 °C; Hammouya et al. 1998) rather than magma emission. Boudon et al. (1998) suggested that the evidence supports the conclusion that the Gages Soufrières are fed by a robust hydrothermal circulation system that survived the onset of magma venting without fundamentally changing its character from an ‘upper alteration zone of a high sulfidation system’. Quite how this hydrothermal system remained isolated from the magma conduit system that was developing nearby is unclear; Boudon et al. (1998) suggested that it is due to the development of a low-permeability seal between the two by rapid deposition of silica from magmatic gases.
Substantial changes in the chemistry of emitted gases were observed at Galway's Soufrière before it was lost to sight (Hammouya et al. 1998). Galway's had always been the most active fumarolic field, and changes in noble gas and isotope signatures were interpreted by Hammouya et al. (1998) as indicative of both increased boiling of the feeding hydrothermal waters and greater inputs of oxidizing magmatic gases. Nevertheless, although more interaction with magma is inferred, there are no suggestions of a diminution in scale of this hydrothermal system.
Notwithstanding the abundance of compositional data for the gaseous emissions of the soufrières and hot springs located high on the flanks of the volcano, only a few, incomplete, chemical analyses of their thermal waters were ever made, by Chiodini et al. (1996), Pedroni et al. (1999) and Atkinson et al. (2000). All available analyses show that these waters were acidic and of Ca–SO4 facies, a category that in volcanic settings is commonly associated with the dissolution of volcanogenic hydrogen sulphide gas in shallow groundwater (Henley & Ellis 1983). The wider significance of this finding will be discussed below.
Lowland and submarine thermal water emissions
In addition to the soufrières and hot springs high on the flanks of the volcano, hydrothermal emanations have also been reported previously in two other settings, as follows.
As hot springs, sub-tidal hot water outflows and hot water boreholes scattered along the coast between Plymouth harbour and Elberton (Tombs & Lee 1976; Wright et al. 1976; Chiodini et al. 1996; Atkinson et al. 2000). For instance, Tombs & Lee (1976) reported that ‘Groundwater was hot or warm in boreholes between Elberton and O'Garra's. There was some steam at the Emerald Isle borehole’. O'Garra's is the most southerly district on the island, beyond the end of the main coastal road, which formerly ran SE from Plymouth to Old Fort Point (Fig. 1). The ‘Emerald Isle borehole’ served a spa pool on the beach at the hotel of the same name, just north of Plymouth. Wright et al. (1976) recorded temperatures of up to 87.5 °C in this borehole, with an apparent geothermal gradient of the order of 3 °C m−1. A feature known to local residents as the ‘Hot Water Pool’ lies near the beach a few hundred metres NW of this hotel. Observations of this feature made before the onset of the post-1995 eruption were briefly recounted by Tombs & Lee (1976), Wright et al. (1976), Bath (1977) and Chiodini et al. (1996). Pedroni et al. (1999) and Atkinson et al. (2000) sampled the Hot Water Pool in the early years of the eruption. Only limited chemical analyses of the waters were reported by those workers (see Table 1), and no comparison of their findings has yet been made, for Atkinson et al. (2000) did not cite the work of the previous three groups of researchers.
A major submarine hydrothermal discharge zone associated with the Montserrat–Marie Galante Fault, the scarp of which forms the Shoe Rock escarpment on the sea bed between Montserrat and Guadeloupe (Polyak et al. 1992). At the toe of this scarp, hot water outflow into the ocean gives rise to anomalies in seawater temperature and chemistry (including clear peaks in concentrations of zinc and mantle-derived 3He) and precipitation of secondary minerals (todorokite, nontronite, etc.).
Since the onset of the post-1995 eruption, it has been possible to examine hydrothermal features only in the DAZ (Fig. 1). Plumes of steam sporadically rising from shallow seawater along the shoreline at Plymouth were observed from the vantage point of St. George's Hill (Fig. 2). This encouraged close inspection of the area, which took place on 28 September 2005.
The Emerald Isle Hotel beach spa, where Tombs & Lee (1976) previously reported hot water and steam, was located (grid ref. [75444754]) and visited. The hotel and spa were abandoned, being heavily cloaked in about 0.5 m of airfall ash. Although the borehole was not found during the site visit, a warning sign in the old spa pool read ‘The water approaches boiling point’. The beach was followed for several hundred metres NW from the abandoned hotel, until the ‘Hot Water Pool’ was encountered (grid ref. [75164776]). The Hot Water Pool is the mouth of a steep, fluvial ravine that is naturally dammed by beach sands. The barrier beach is sedimentologically indistinguishable from the cliff-foot beach either side of the ravine mouth, with dune bedform features that suggest that it has been emplaced (and maintained in position) by a combination of longshore drift and aeolian transport. At high tide, seawater was seen overtopping the barrier beach, introducing seawater directly into the pool. Although the water within the pool was noticeably warm to the touch, the incoming seawater was not warm; thus the elevated temperatures within the pool are due to inputs of terrestrial hot water. Given the undoubted marine influence on the quality of the water in the pool, this was not sampled directly. Rather, an upstream reconnaissance was pursued until free-flowing waters above the water level in the pool were encountered. The first freely flowing waters accessed upstream of the pool had a temperature of 47 °C. It soon became apparent that this water was a mixture of natural hot groundwater discharge (≤64 °C), which was clearly seen to be emerging from the streambed sediments at the margins of the valley floor, and cooler upstream water (32 °C; close to the ambient air temperature at the time). Temperature, pH, conductivity and oxidation–reduction potential were also measured on-site. Samples were also collected for later laboratory determinations of selected dissolved species. Results are given in Table 1, alongside previously published analyses (and some unpublished analyses from BGS and the MVO). Comparison of the analyses of the recently collected sample of hot spring water (sample 3 in Table 1) with those previously reported by other researchers (the four right-hand columns in Table 1) reveals that the hot spring has been remarkably constant in chemical composition over time, with only minor variations in most major ions (e.g. calcium, magnesium, sodium, potassium, sulphate and chloride) and close correspondence in many minor elements (e.g. iron, manganese, boron, strontium and lithium) across all sampling dates. This consistency is all the more striking given the inevitable differences in analytical techniques and precision between the various prior investigators. This consistency suggests that the hydrothermal system that feeds the Hot Water Pool has not undergone fundamental change in the 15 years since the earliest analysis. Without considering alternative explanations, the microbiologists Atkinson et al. (2000) casually suggested that the thermal waters here are barely modified seawater of local origin. However, all the hydrochemistry, including the ionic ratios of major cations and anions, and particularly the elevated concentrations of boron and lithium, strongly contradict this: these waters have all the geochemical hallmarks of hydrothermally reacted seawater, which has circulated fairly rapidly to sufficient depth to heat up and react strongly with easily altered volcanic rocks (see Henley & Ellis 1983; Wohletz & Heiken 1992).
Continuing the reconnaissance upstream from the head of the Hot Water Pool, free-flowing surface water (an acidic sulphate water of low ionic strength; sample 1 in Table 1) persisted for about 100 m, above which the streambed became dry. At this point, the valley narrowed to a steep-sided ravine cut into cemented, volcaniclastic debris flow material. Locally, the material is seen to have been pervasively hydrothermally altered, with mineral parageneses and fabrics typical of advanced argillic weathering. None of the altered zones was seen to be emitting steam or hot water at the time of inspection.
Synthesis: preliminary conceptual model
The evidence presented above demonstrates clearly the following.
There is a long-established hydrothermal system associated with the Soufrière Hills Volcano.
Notwithstanding the phreatic nature of the initial phase of the continuing eruption in 1995, evidence gathered before the soufrières were buried beneath volcaniclastic deposits indicates that this hydrothermal system was not destroyed by eruptive activity.
Persistence of the hydrothermal system to the west of the active volcano is corroborated by the continued hydrothermal emanations in the vicinity of the Hot Water Pool, which are of the same physical magnitude as implied in much earlier descriptions, and which are certainly of very similar chemistry to that previously determined.
The nature of seismic activity associated with the eruption suggests that the transmissivity and storativity of the hydrothermal system might even have been enhanced by deformation associated with shallow intrusion of magma beneath the main volcanic edifice.
Hydrochemical contrasts between the former soufrières and the extant Hot Water Pool are consistent with the classic model for hydrothermal systems associated with active island-arc volcanoes described by Henley & Ellis (1983), and summarized in Figure 3, in which high-level fumaroles represent recirculation of rainfall-fed recharge onto the active volcanic edifice, whereas the more distal, Na–Cl waters represent hydrothermally altered seawater, moving in accordance with thermal convection such that recharge is supplied by dense, cool marine groundwater sinking fairly rapidly towards the volcanic axis, where it is rapidly heated, thence rising back towards surface as a lower density, seaward flow of hot water, thermally stratified above (and often hydrostratigraphically separated from) the cooler, more dense marine recharge waters. Such circulation systems are not only known from andesitic island volcanoes similar to that of Montserrat (Henley & Ellis 1983); they are perhaps best documented from the Reykjanes Ridge geothermal field in Iceland (e.g. Sakai et al. 1980; Elderfield & Greaves 1981; Arnórsson 1995).
Fig. 3Schematic cross-section illustrating the conceptual hydrogeochemical–geothermal model derived for western Montserrat. Large block arrows indicate approximate directions of groundwater flow. Generic hydrochemical zonation and approximate isotherms are in accordance with the generic model for active island arc volcanoes proposed by Henley & Ellis (1983).
Corroboration that the Montserrat hydrothermal system conforms to the Henley & Ellis (1983) model can be found in details of the pre-eruption conditions documented by Wright et al. (1976), Bath (1977) and Chiodini et al. (1996). The two principal elements of the system, as described by Chiodini et al. (1996), may be summarized as follows: (1) acid to neutral, steam-heated waters, present together with several fumarolic vents, discharging vapours formed through boiling of hydrothermal aqueous solutions, in the four soufrières; (2) deep-seated Na–Cl waters in the vicinity of Plymouth, which are deduced to have equilibrated with the rock mass at a temperature of around 250 °C on the basis of widely used Na/K, K/Mg and quartz geothermometric calculations (Wright et al. (1976) had previously obtained a similar result, i.e. 220 °C, by application of the same suite of geothermometers).
Both elements of the system were inferred to have been influenced to some degree by direct contact with island arc-type magmatic water, on the grounds of high 3He/4He ratios in the soufrières, and on the basis of 18O and deuterium signatures in the steam condensates and Na–Cl thermal springs (Chiodini et al. 1996). That the thermal waters emerging in springs at the Hot Water Pool are ultimately derived from recharge of modern marine waters is further substantiated by the positive tritium value reported by Bath (1977).
All of the available evidence suggests that the present-day hydrothermal system between the summit of the Soufrière Hills Volcano and the coast at Plymouth conforms to the generic hydrogeothermal model for active island-arc volcanoes (Henley & Ellis 1983), as summarized in Figure 3. This is extremely encouraging from the point of view of geothermal exploration, as it is precisely in other areas that conform to this model that the majority of the world's most prolific geothermal reservoirs have been successfully explored, developed and economically exploited. The full roll-call of successful geothermal power plants exploiting hydrothermal systems of this type would fill a book; some of the more prominent examples include the major geothermal fields of Japan, New Zealand, the Philippines, Indonesia and Central America (DiPippo 2005). Within the Caribbean, the only successful geothermal development to date (i.e. the Bouillante Field, Basse Terre Island, Guadeloupe) exploits a natural hydrothermal system that is in every detail analogous to that described above on Montserrat (Brombach et al. 2000).
Exploration and development considerations
Although there is little doubt that hydrothermal systems exist on all flanks of the Soufrière Hills Volcano, considerations of safe access mean that permanent new-build of even as diminutive a structure as a 5 MW wellhead generating unit could be proposed to take place only in the current DAZ. As has been noted, natural and artificial manifestations of hot water (approaching 100 °C) are present in the immediate subsurface in the vicinity of Hot Water Pool and at several points along the coast towards Plymouth. The coastline between the Hot Water Pool and the westernmost tip of the island follows a strikingly linear NW–SE trend. This appears to be parallel to the inferred fault line (discussed above) running through St. George's Hill and Garibaldi Hill. If coastal morphology in the vicinity of the Hot Water Pool is controlled by a similar fault, then we may postulate a more general association between faults on a parallel trend and hydrothermal upwelling in this part of the island. From consideration of the generic hydrogeochemical–geothermal model of Henley & Ellis (1983), to which these manifestations have been shown to relate, it is an obvious postulate that higher temperatures are likely to be encountered in the subsurface further inland (Fig. 3). The earlier findings of Tombs & Lee (1976), who inferred a shallow body of hot water centred on a point in the upper Amersham Estate (grid ref. [788468]; a place now utterly inaccessible save by helicopter), can now be reinterpreted in the light of the above. If a line is projected northwestwards from point [788468] into the present DAZ, a preferred exploration location is defined close to the western flank of St. George's Hill, on the eastern edge of the valley that separates St. George's Hill from Richmond Hill. The southeasterly extremity of the proposed exploration zone would thus be at point [765483], and its northwestern extremity on a NW trend from this, around point [759486], which is in Elberton, near the southeastern foot of Garibaldi Hill. Logic would suggest that the more southeasterly the borehole, the more likely it would be to encounter water of suitably high temperature (i.e. in excess of 170 °C).
Were exploration to proceed here, it would not be the first such attempt in this part of the Caribbean. Some of these previous efforts have met with frustration. For instance, on St. Lucia, sufficiently high temperature resources were identified (Williamson 1979), but commercial development was hampered by a range of problems including cyclical alternations between steam and brine production in the same well, high gas contents preventing the use of conventional condensing exhaust turbines, and corrosion of pipework by low-pH waters. More encouraging experiences during the development of the Bouillante Field on Guadeloupe (Correia 2000) provide interesting insights into the likely magnitude of resources that could be expected from one or more successful geothermal wells drilled in this area of Montserrat. Following a series of geophysical and geochemical surveys, four exploratory wells were drilled in the Bouillante area during the late 1960s and early 1970s. Of these four wells, one was effectively dry, a second produced an intermittent wet steam resource, and two produced continuous steam discharges. The best of these (well BO-2) intersected a moderately transmissive geothermal reservoir at 320 m depth, which yielded a steady flow of 150 t h−1 of high-pressure steam at a temperature of 248 °C (which is about the same as the equilibration temperature inferred for the saline waters of the Montserrat Hot Water Pool by Chiodini et al. 1996). This well was not put into productive use until 1987, when it was fitted with a 4.7 MW wellhead turbine unit. In contrast to experience elsewhere with industry-standard portable wellhead generating units (Hudson 2005), this non-standard double-flash condensing cycle plant had a somewhat chequered history, and production was suspended between 1991 and 1996, when it was replaced with a more conventional and simpler atmospheric exhaust turbine. Since August 1996 the Bouillante plant has run without problems, with availability for over 90% of the year and an annual production of 23 GWh. As local confidence and experience with the technology grew, so plans for expansion of geothermal power generation were implemented, with a further 11 MW capacity being commissioned in 2004, fed by three new directionally drilled boreholes. The second-best of the four original wells has also been brought into use following a successful exercise in 1998 in which the well was stimulated (i.e. thermal cracking of the surrounding geothermal reservoir was achieved by injecting cold seawater into the well). The yield of this well was more than doubled, and currently peaks at 140 t h−1. This second well is now being deployed as a ‘make-up well’ connected to the original (Bouillante I) geothermal power plant, restoring the generating capacity of that plant to 4–5 MW, from the 3 MW rating to which it had slowly declined over the 30 years since borehole installation (Correia 2000).
On the basis of this Caribbean experience, and many more experiences in similar fields worldwide (DiPippo 2005; Dickson & Fanelli 2005), it is reasonable to propose that one or more wells properly drilled and completed in the Elberton–Richmond zone of Montserrat ought each to be easily capable of perennially yielding high-pressure steam at rates of about 150 t h−1, which is sufficient to support an electrical power plant rated at 5 MW. This comfortably exceeds the peak electricity demand on the island for the foreseeable future, which is estimated by Montserrat Utilities Ltd to be around 3.5 MW (T. Yeung, pers. commun.).
A single well yielding up to 5 MW of steam can be harnessed for electricity production using a portable, skid-mounted turbine generator unit (DiPippo 2005). Preliminary estimates of the capital and operating costs for the installation and operation of such a system have been derived by applying a generic costing model calibrated with data obtained from a large number of full-scale geothermal developments worldwide (Sanyal 2005). An initial exploration phase of geophysical surveys and exploratory drilling of up to six ‘slimline’ (i.e. <150 mm) boreholes with appropriate blow-out prevention and well-quenching safety measures is anticipated to cost around £500 000 to £1 million. (These figures allow for the fact that no suitable rig exists on Montserrat, so that one would need to be brought in from another island). The capital cost of a 5 MW atmospheric-exhaust wellhead turbine is estimated at around £7 million, and this would have operating costs in the region of 1.5p per kWh; allowing for strengthening of the island's power grid to cope with the relocation of the primary electricity source, the overall cost of electricity (amortized over a 30 year plant design life) is estimated at not more than 3.5p per kWh. This is less than half the current cost of electricity generated on the island using diesel generator sets.
Although a 5 MW plant would satisfy demand on the island, experience in similar geothermal fields elsewhere suggests that considerably greater resources might eventually prove to be available, not only in the Richmond–Elberton area, but elsewhere around the coast, once safe access is regained to the Tar River valley in the east and O'Garra's in the SW of the island. In the fullness of time, the prospect of generating 50 MW or more of electricity is not an unreasonable aspiration (see DiPippo 2005); this would place Montserrat in the favourable position of being able to develop an export business for excess power to neighbouring islands.
Another possibility, in the medium term, would be to increase the baseload operation level of the geothermal power generation plant and consume the excess of power production over local demand in producing storable hydrogen, which could then replace fossil fuels for vehicles on the island. Such a plan has recently been mooted for the Azores, where a 5 MW geothermal plant already exists. If implemented on Montserrat, this approach would essentially complete the transformation of Montserrat into a ‘carbon-neutral’ island, with bright green credentials to match its ‘Emerald Isle’ image; the benefits of this for eco-tourism ought to be considerable.
Conclusions
Although a number of hot springs and fumaroles high on the flanks of the Soufrière Hills Volcano were buried beneath thick deposits of ash during the early years of the post-1995 eruption, a zone of hot (≤64 °C) groundwater discharge persists on the northern edge of the abandoned town of Plymouth, in an area that closely adjoins the still-inhabited half of the island. The seepages and small springs in this zone all emit Na–Cl groundwaters, with major-ion, B and Li concentrations that all strongly suggest that they originated as seawater that underwent extensive rock–water interaction at high temperatures and pressures hundreds of metres underground. Taken together with previously published volcanological observations, the persistence of this discharge zone throughout the sustained eruption period suggests that a robust hydrothermal circulation system exists at depths of several hundred metres. Comparison with similar systems elsewhere suggests that there is significant potential for development of high-enthalpy geothermal energy resources in this vicinity. A single successful borehole here, equipped with a portable well-head generating unit, could reasonably be expected to yield up to 5 MW of electrical power, which significantly exceeds the current peak electricity demand on the island (c. 3.5 MW). Clearly, particularly high risks would attend such developments in the vicinity of an erupting volcano. Nevertheless, this study exemplifies the lessons that can be learned by a careful collation and examination of data in advance of exploratory geothermal drilling.
Acknowledgements
The author gratefully acknowledges financial support from the HSBC Partnership for Environmental Innovation. He would also like to acknowledge the hospitality and practical support in Montserrat of T. Yeung (then CEO of Montserrat Utilities Limited), J. Osborne (Chief Minister), J. Wilson (Minister for Communications & Works), G. Gray (Conservation & Environment Adviser, Ministry of Agriculture, Land Housing and Environment), R. Aspin, S. Loughlin (then Director of the Montserrat Volcano Observatory), R. Isles (who took me safely through the DAZ into central Plymouth), K. Lee (Montserrat Utilities Limited ) and V. Buffong, who showed me the location of the Hot Water Pool. At Newcastle University, P. Orme analysed hydrochemical samples, and N. Kruse helped find some literature recommended by referees. G. Darling of BGS was also most helpful with the latter, with references to analogous situations in Iceland, and with terminological definitions. I am also grateful to J. Hirosato (of Mitsubishi Heavy Industries Ltd) and H. Sasaki (of Fuji Electric Systems Ltd) for assistance with capital cost estimates for generating plant. Finally, I would like to acknowledge my debt to A. Bath and an anonymous referee, for incisive and highly constructive comments.
- © 2010 Geological Society of London
References
Reconnaissance assessment of the prospects for development of high-enthalpy geothermal energy resources, Montserrat
