Drilling of an Ore-Forming Shallow Hydrothermal System, Santorini Volcano, Greece

INTRODUCTION

Numerous occurrences of metalliferous hydrothermal systems and associated submarine metal-rich sediments have been reported from most spreading or mid-ocean ridge systems (MOR), which, furthermore, are characterized by high heat flows. Such MOR-systems have been studied extensively in the last 25 years, partly through DSDP (Deep Sea Drilling Project) and the following ODP (Ocean Drilling Project); these studies have resulted in numerous publications, which have caused a major revolution in our thinking about submarine volcanism; for reviews, see Boström (1980), Cronan (1980), Rona et al. (1983) and Rona (1988).

In contrast to this situation most occurrences of metalliferous hydrothermal systems and submarine metal-rich sediments at island arcs and related environments have received fairly restricted interest. Such systems are known from several areas in the world, though, like the Aeolian island arc in the western Mediterranean Sea (Stromboli volcano), Banu Wuhu (Indonesia) and from the Kuriles; such occurrences have been discussed by Zelenov (1964), Honnorez (1969) and Cronan (1980); a more detailed discussion of the global distribution patterns for such process and supporting data are presented by Zelenov (1972) and by Boström et al. (1990b). It is obvious that the Santorini hydrothermal system also shows distinct metallogenetic tendencies, as indicated by the occurrence of iron-rich solutions and muds, although they are poorer in trace metals than sediments at deep-sea spreading systems.

Since hydrothermal processes and metal-rich sediment formations occur both at spreading systems and at subduction zones, it would be of great interest to compare the metallogenetic processes in these two geotectonic environments. Such a study is of importance not only from a purely scientific point of view, but also has a substantial significance for applied aspects of economic geology and mineral exploration. A cross-section study of a recently active system should therefore greatly advance our understanding of such ore-forming processes and hydrothermal systems. The Santorini island group in the southern Aegean sea (Fig. 1) is an ideal site for such a study, both from a scientific and from a logistical point of view.

SURFACE STUDIES OF THE HYDROTHERMAL ACTIVITY AT SANTORINI

Early observations of surface manifestations of the hydrothermal processes were decisive for the selection of the present drilling target; a summary of these indications is therefore of interest.

The Hellenic volcanic arc was formed by episodic volcanic activity from Late Pleistocene to recent times as the African plate became subducted beneath the Aegean microplate (see Fig. 1). Present hydrothermal activity is best displayed at Santorini, but other sites also show indications of hot-spring activity, e.g. Methana, Milos and Nisyros; locally such processes have produced conspicuously rich deposits of manganese and barium in the Vani caldera on Milos (V. Galanopoulos, unpubl. data).

The Santorini islands obtained the present morphology (Fig. 2, bottom) due to a Minoan-age eruption sometime between 1400 and 1600 BC; this produced a large part of the present caldera, which is open to the sea through channels between the islands Thera, Therasia and Aspronisi. The islands Palaea Kameni, Mikra Kameni and Nea Kameni formed in AD 197, 1573 and 1711 respectively, the last two later merging into a larger Nea Kameni. However, the development history of the caldera and the Kameni islets and the mechanisms involved has been only partly resolved (Fouqué 1879; Hammer et al. 1980; Pichler and Kussmaul 1980; Heiken and McCoy 1984; Aston and Hardy 1990; Druitt and Francaviglia 1990; Heiken et al. 1990; Sigurdsson and Carey 1990).

The Palaea Kameni and Nea Kameni islands consist of dacitic flows and domes that rise a few hundred metres above the caldera bottom. The presence of metalliferous sediments and the high temperature (~ 40° C) of the hydrothermal emanations suggest a pronounced hydrothermal circulation and associated ore deposition. At least eight hydrothermal discharge sites are known around these is lands (Butuzova 1969), depositing iron-rich sediments (Fe₂O₃ ~ 30-40 wt %), but if all sites with weak hydrothermal deposition are included the number of discharge zones may be at least eighteen (Boström et al. 1990a).

The top layer of the metalliferous sediments are mainly orange to brown muds of iron hydroxides, turning greenish black at depth where sulphate-reducing bacteria are present, but the total iron content varies little with depth in the sediments. Other constituents in the muds are rock detritus from eroded volcanics, pumice, and opaline silica, much of it deposited because of a volcanic supply of silica (Fouqué 1879; Brun 1911; Butuzova 1966; Bonatti et al. 1972; Puchelt 1973; Puchelt et al. 1973; Schroll 1978; Smith and Cronan 1983; Boström et al. 1990a, b).

In addition to iron, these hydrothermal solutions deliver minor amounts of manganese, zinc, barium and copper into the sea water, i.e. the iron-manganese ratio is high. Furthermore, the hydrothermal spring waters at Nea Kameni are much poorer in uranium and show higher calcium-sodium ratios than ordinary sea water, suggesting spilitization and uranium uptake at depth (Smith and Cronan 1983; Boström and Widenfalk 1984, Boström et al. 1990b).

This hydrothermal system displays significant differences relative to its ocean floor counterpart in its geometry and in some of the element fluxes. Harder (1964) thought carbon dioxide was a major transport agent for iron, but some volcanic emanations are very rich in carbon dioxide, and are yet poor in iron (White and Waring 1963). To explain these particularities Boström and Widenfalk (1984) proposed a model (Fig. 3) whereby descending oxygen is mixed with hydrogen sulphide that has been expelled from a degassing magma chamber. This mixing leads to oxidation of sulphide, and the sulphuric acid formed may cause digestions of the dacitic carapace hosting the hydrothermal convection. Such a leaching process could account for the low abundances of Mn, Zn, Cu and Ba relative to Fe in the discharging solution, since the leaching, if fairly complete, should yield similar metal proportions in the final solutions as in the leached rock. This model has been somewhat modified during this project in accordance with new data (Fig. 4) insofar as the main rock-water reactions probably occur at larger depths than assumed earlier.

The sediments far way from the Hellenic arc also show regional elemental trends due to the distance from the debouching springs (Varnavas 1989). Furthermore, the sequence of sediments away from the submarine hydrothermal outlets off Santorini with the differential precipitation of iron and manganese, and the possible pre-segregation of sulphides, may show several analogues with the situation around the Atlantis II Deep in the Red Sea (Cronan 1980).

THE GREEK-SWEDISH JOINT DRILLING PROJECT

The studies referred to above have relied on surface data. Consequently the processes at depth could be treated in an indirect manner only. Such indirect conclusions must often be used in geological research, but involve a serious risk of misdirected speculation. The paucity of data regarding processes at depth at Santorini therefore led to the obvious conclusion that a drill hole was needed to attack several of the unsolved problems (Arvanitides et al. 1988).

Such a bore hole, some 200-300 m deep, could penetrate a significant part of the hydrothermal cell and allow the monitoring of the down-hole changes in the ore-forming processes, as the degree of dilution of the ascending metal-rich solutions by admixed cold sea water decreases. Furthermore, logistically this hydrothermal system is fairly accessible, which means that a drilling operation could be achieved at low cost. This situation is in remarkable contrast to that for some analogous sites, like the submarine volcanoes Banu Wuhu or Kick'em Jenny, which are less easily reached.

In 1986 K. Boström (Department of Geology and Geochemistry, Stockholm University: US) suggested a joint Greek-Swedish drilling project, to be run by IGME (Athens) and US. This proposal was readily accepted, not the least because similar plans had been circulated also among leading scientists at IGME.

The aim of this Greek-Swedish venture was the direct or indirect down-hole measurement of several critical physico-chemical parameters, such at T, pH, fO₂, fH₂S, and fCO₂, and to correlate them with corresponding variations in solution chemistry and mineral parageneses. This would permit the modelling of the ore-forming processes in this particular tectonic environment. The targets of the project are summarized in Table 1.

The rugged topography of the volcanic terrain and the necessity to be as close as possible to a hydrothermal discharge zone imposed considerable constraints on the choice of drill site. The north-eastern bay of Palaea Kameni (Fig. 2) appeared to be a promising site, since there was a small plateau only some 40 metres away from hot springs discharging close to the shoreline at shallow water depths.

This plateau could accommodate the drilling equipment, but the transfer of the drilling rig proved difficult. Due to the rocky shores and vicious shoals, transport by boat proved impractical. An airlift therefore had to be made, which was made possible by the constructive assistance of the Greek Army, which put a helicopter at our disposal.

The wire-line core drilling operation started in May 1987, and proceeded at a rate of 30 m per month. The hole diameter was 92 mm for the first 24 m, 76 mm for the next 72 m and 60 mm for the remainder of the drilling operation. The difficulty in maintaining large hole diameters even at shallow levels was due to the intense fracturing of the rocks, which also necessitated continuous casing of the hole.

Operations were occasionally impaired by bad weather conditions, preventing the commuting of the drilling crew and the transport of supplies to and from the island of Palaea Kameni. On the whole, however, the drilling was kept on schedule and well within budget.

The total on-site expenditure on the project, including labour costs, overheads, travel expenses, transport of equipment and supplies, consumables and equipment repair costs, was 10,600,000 drachmas, equivalent to 53,000 drachmas per metre. This is a very low budget considering the logistical difficulties involved during this operation.

In addition, travel between Sweden and Greece by several Greek and Swedish participants for co-ordination and planning sessions, for the down-hole logging and for the logging equipment involved another 2,500,000 drachmas. Costs for running various analyses are not included in these amounts.

PRESENT DRILLING RESULTS AND CONCLUSIONS

The materials and solutions obtained to date have been only partly studied, but results of an initial report nature have been presented (see further below). Other studies, using isotope relations, for example, have just begun, and additional geochemical studies of the pumice-rich sections are still in progress. However, several results have been presented at this conference; for this reason, the following is only an outline of the results with respect to core lithology, chemistry of the fluid phase etc., indicating the present state of the on-going phase of the project.

1. The geochemical studies of the sediments from the open Santorini caldera suggest that the caldera sediments have a substantial exhalative-sedimentary input. In the areas north-east of Nea Kameni this input is very pronounced, indicating the discovery of a new submarine hot spring (Perissoratis et al. 1990; Boström et al. 1990d; see Fig. 2).
2. Drilling in Palaea Kameni terminated at a depth of 201.5 m and recovered 155 m of dacitic lavas and 46 m of an underlying consolidated gravel horizon. The andesites are identical to those from the Kameni islands described by Pichler and Kussmaul (1972). The unconsolidated deposits are rich in pumice and reddish brown to brown horizons, containing iron-oxide rich exhalative-sedimentary components, resembling recent ones at Nea Kameni. These results demonstrate that the exhalative-sedimentary processes have been proceeding for a considerable time, starting well before Palaea Kameni surfaced (Arvanitides et al. 1990).
3. The drill hole did not hit the central ascending limb of the hydrothermal cell, but rather a part of the mushroom-like warm water plume welling out to the side. In the deepest waters the contents of dissolved carbon dioxide, iron, barium and silica are high, the content of oxygen is low, and the temperatures are considerably higher than in the caldera waters at the same depths. These relations indicate that the solutions are strongly affected by leaching processes at depth. Furthermore, the calcium-sodium relations are higher than in sea waters, suggesting that spilitization takes place at depth. The solutions strongly resemble those found in the most active springs on Nea Kameni and at Banu Wuhu (Boström et al. 1990b, c).
4. The lavas do not exhibit distinct signs of hydrothermal alterations, possibly because the alteration processes have been going on for a short time only, and because alteration temperatures are low, slowing down reaction rates (Paritsis et al. 1990).
5. Late vein coatings, particularly in the upper part of the drill hole, show a distinct similarity to the exhalative-sedimentary deposits that now form at the most active Nea Kameni hot springs (Arvanitides et al. 1990; Boström et al. 1990c). This relation and the fact that the hot-spring water chemistry seems to show small regional variations between Nea Kameni and Palaea Kameni suggests that the leaching zone is located at a considerable depth and that all solutions may have the same main source.

A PROPOSAL FOR FURTHER DRILLING AND RESEARCH ON NEA KAMENI

The lack of information about the structure at depth of the hydrothermal system under the Kameni islands precluded the choice of an optimal site from a scientific point of view. Nevertheless, the selected drill site was probably the best that could be obtained for a project of short duration, operating on a restricted budget. Furthermore, the results have significantly improved our understanding of how to attack these problems with a second drill hole.

Our results suggest that the drill hole must be considerably deeper than in this project, probably at least 500 m, and must be located closer to the suspected genetic centre of the hydrothermal solutions. In view of the intense exhalative-sedimentary deposition on Nea Kameni, we conclude that the next deep hole should be set on Nea Kameni, preferably in conjunction with those lavas that are near the most active springs.

Fig. 1 in Boström et al. 1990a shows that most of the very active hot springs occur near the Georgios lavas, whereas much younger lavas show less hydrothermal activity. This distribution pattern is probably partly related to the massive outpouring of lava that the Georgios flows of 1867 represent; no later effusive event on Nea Kameni matches the Georgios flows in volume. As a consequence these lavas may be cooling more slowly and are the only ones that can entertain hydrothermal leaching processes for a longer time, provided that these processes primarily take place in shallow sections; another interpretation is that the coinciding distributions for very active springs and Georgios lavas means that lavas and hot-spring solutions of a deep origin ascend in the same weakness zones, which are discussed by Fouqué (1879) and Heiken and McCoy (1984). Geophysical site surveys could solve some of these problems and make it possible to find an optimal drilling site.

The selection of a drill site on Nea Kameni will present severe problems, however. From a strictly scientific point of view it seems probable at present that one favourable drill site could be south of the most active bay on Nea Kameni, but logistically this is a difficult area, as is also most of Nea Kameni. The ground there is covered with large lava boulders, many 20-40 cm in diameter, generally with sharp edges. Drilling in this area would probably demand construction of suitable platforms of, e.g., cement, concrete or wood planking, and a simple road would have to be made to the site. Airlift by helicopter may also be the only possible way to bring in all heavy drilling gear, and possible also the materials for the construction of a simple shed for rest, shade and breaks in the work, as well as for some initial scientific work, including core description and sampling on site.

The costs of this project could therefore be 2-3 times higher than that for the drilling on Palaea Kameni, but without the experience gained in the latter project, the drilling on Nea Kameni would have been still more expensive. Consequently, this project should probably be launched as an EEC project, with supplementary funds deriving from IGME and possibly also the Thera Foundation, as well as from interested scientists in Sweden and the United Kingdom, for example.

We think that this project is well worth following up, and implies a scientific challenge to which the international scientific community should respond.

----------------------------------------------

--------------------------------------------------