The Stronghyle Cladera: Geological, Palaeontological and Stable Isotope Evidence from Radiocarbon Dated Stromatolites from Santorini

The results together with the fossil fauna indicate that a sea water filled caldera existed prior to the Minoan eruption. This caldera was even bigger than that proposed by Heiken and McCoy (1984). It extended into the northern part of Stronghyle and was partly separated from the southern part by an island, called pre-Kameni.

1.     GEOLOGICAL EVIDENCE 

     1.1.    Introduction:     

Santorini consists today of three old ring-islands and two younger islands. The latter were formed after the violent Minoan eruption which occurred around 1645 BC (Hammer et al. 1987; Friedrich et al. 1990) and destroyed flourishing Bronze Age settlements. One of the settlements is presently being excavated near the village of Akrotiri on Thera (Doumas 1983). The old ring-islands are the remnants of a former circular island, called Stronghyle, which means 'round'.

One of the hitherto unanswered questions which is of great interest for archaeologists and earth scientists alike is what shape the island had prior to this eruption. Since no written or painted information exists which could give us an idea of the shape of Stronghyle island, the only data available are those present on the island itself, namely the remnants of the old island and the products of the volcanic eruption.

Several attempts have been made to produce a reconstruction of this island, but only recently has it become clear that a caldera actually existed just prior to the Minoan eruption. In a paper about the mechanism of the Minoan eruption Pichler and Friedrich (1980) showed a reconstruction where, on the basis of observations on the caldera walls of the present day caldera, they deduced that a bay extending 3-4 km in a north-east direction must have existed in the southern part of Stronghyle. Heiken and McCoy (1984) gave evidence for a much bigger pre-Minoan caldera. Friedrich et al. (1988) presented geological evidence that there existed an even bigger caldera than that proposed by Heiken and McCoy. This caldera was filled with sea water and extended into the northern part. It was partly separated from the southern part by an island. This observation was based on the occurrence of xenolithic stromatolites collected in the third layer of eruption products of the Minoan eruption. It was concluded, on the basis of their radiocarbon ages and stable isotope investigations, that the stromatolites were formed in shallow marine water prior to the Minoan eruption.

In this paper we present further radiometric dates and palaeontological data to support this observation. During the Third International Congress on Thera, T. Druitt and V. Francaviglia presented compelling evidence for the extension of the caldera in the northern parts, where they demonstrated ancient cliff lines at Cape Phira, Cape Tourlos and north-north-west of Mikros Profitis Ilias (Druitt and Francaviglia 1990). In the region of Megalo Vouno, Heiken and McCoy (1984) had already observed ejected units of the Minoan eruption plastered in the upper quarter of the slope. These independent approaches support the caldera reconstruction given by Friedrich et al. (1988). In the reconstruction given in the present paper, the above mentioned observations concerning the ancient caldera cliffs are incorporated.

     1.2.    Xenoliths of the upper pumice series:     

The volcanic products of the Minoan eruption contain information not only about the mechanism of the eruption, but also about the origin of this material. We estimate that about 99% of the material consists of volcanic products, like pumice, ashes and different types of older dark xenolithic lava blocks and pyroclastics that were ejected during the eruption. Less than 1% consists of non-volcanic light-coloured xenoliths. The latter were ejected during the widening phase of the vent in the third phase of the eruption (Fig. 1).

There are two types of non-volcanic xenoliths, which can easily be distinguished from each other on the basis of their fossil content and their colour: the first type includes fragments from the non-volcanic basement, which consists of metamorphosed rocks like phyllites and schists, marbles and sandstones. Their age is Permo-Triassic, based on fossils found in the Mt. Ilias region on Thera (Papastamatiou 1958). However, fossils from these relatively seldom occurring xenoliths are not yet reported. Much more frequent is the second type of xenolith, which consists of calcareous blocks with stromatolitic structure. They are not metamorphosed and, therefore, must post-date the first group. Their shape is often globular, with a lamellar structure, and they are often fossiliferous. These stromatolites were found both 'in situ' in the third layer (Bo3 or C-layer in Druitt et al. 1989) of products of the Minoan eruption, and also lying on the surface of the fields on the islands of Thera and Therasia. The stromatolites found 'in situ' occurred mostly in the lowermost part of the third eruption phase (Fig. 2 and 3). A great deal of the finds occurred in the erosional channels on Thera and Therasia. Some stromatolites had been gathered together in stone heaps in the abandoned pumice quarries in the northern parts of Thera and Therasia where, because of their size, they were considered a waste product which, together with certain boulders and stones, could not be used for practical purposes. In a quarry in the northern part of Therasia 5‰ of those blocks were stromatolites. Blocks of considerable size were sometimes found in these stone-heaps (Fig. 4). In other cases the calcareous blocks are very easily recognized when they have been used in stone-walls, together with the darker volcanic material. Such stone-walls are very common, surrounding the fields on Santorini.

One of us (U.E.) collected information about more than 1400 specimens of stromatolitic blocks from the Santorini islands. The biggest of these had sizes up to 1.7 metres and a weight of 5 tons (Fig. 5). The weight distribution of the sampled blocks (Fig. 6) shows that the biggest blocks occur in the northern part of Thera and Therasia.

A systematic sampling of stromatolitic blocks was undertaken by means of a grid of observations on the islands. The individual observation areas added up to a total area of 16 km². In this area a total of 53 tons (1253 samples) were observed. In some areas, e.g. in the higher parts of the islands and in the shadow of the higher mountains, no stromatolitic blocks were found. The latter may be explained due to the flow character of the third layer of the Minoan eruption. The weight distribution (kg/km²) is shown in Fig. 7. It shows, like Fig. 6, that the highest concentration of stromatolitic material occurs in the northern parts of Thera and Therasia. Using this method it can be estimated that the total amount of stromatolittc material now deposited on the present-day surface of the ring-islands is about 250 tons. This amount must be even higher if we take the thickness of the Bo3-layer into consideration.

     1.3.    Dating of the stromatolites:

     1.3.1. Relative dating:      Fossiliferous carbonate blocks had been observed as early as 1871 by von Fritsch and later by Reck (1936). Such blocks were investigated by Quenstedt (1936), who dated them on the basis of their fossils to the Late Quaternary. Stromatolites, however, were not reported.

     1.3.2. Absolute dating:      Carbonate from a few stromatolites, gastropods and marine travertine was used for ¹⁴C dating using both Accelerator Mass Spectrometry (AMS) and the conventional dating method. Radiometric analysis has also been made of charcoal from trees, which were covered by the ignimbrites. The ignimbrites at the Akrotiri quarry have recently been dated to 18,150 ± 200 y. BP using the AMS dating facility (Table 1; Heinemeier et al. 1988). This value is in accordance with earlier measurements (Friedrich et al. 1977).

     1.4.    Stratigraphical position of the stromatolites

     1.4.1. Discussion of the validity of the ¹⁴C datings:      Two different types of material were used for the ¹⁴C dating: charcoal from small trees and branches covered by ignimbrites, and carbonates from stromatolites, gastropods and travertine formed in a marine environment. Whereas charcoal poses no special problem, the marine carbonates are suspicious as suggested by the extreme, positive δ¹³C values (+12.3 to +13,8 ‰ PDB). Such high δ¹³C values are not found in marine bicarbonate in equilibrium with atmospheric CO₂, A different origin is, therefore, suggested: an admixture of volcanogenic CO₂ (see section 3). Such CO₂ will be expected to have an infinite radiocarbon age and will therefore render the measured values maximum age limits rather than true ages.

The extreme δ¹³C values indicate that very large amounts of volcanogenic CO₂ must have been incorporated into the carbonates of the stromatolites (this is consistent with the finding that the highest δ¹³C value corresponds to the highest apparent ¹⁴C age). The real age of the dated stromatolites must, therefore, be considerably younger than the dated ignimbrite (c. 18,000 BP) and could well be of an age close to the time of the Minoan eruption.

Based upon their measured ages and the stratigraphical evidence it may be concluded that the stromatolites were formed after the formation of the ignimbrites but before the Minoan eruption.

2.       PALAEONTOLOGICAL EVIDENCE

     2.1.     Stromatolites:

Stromatolites are laminated sedimentary structures formed by dense mats of primary blue-green algae and bacteria, which selectively trap and bind sediment particles on their mucilaginous filaments or precipitate carbonate. Autochthonous precipitation can occur in the following modes (Cohen et al. 1977): 1) Precipitation of carbonate via photosynthetic CO₂ uptake and consequent carbonate precipitation by HCO₃^- shifts, 2) Precipitation by chemo-organotrophic and chemolithotrophic bacterial activities during degradation and decay of blue-green algae material, 3) Inorganic precipitation from sea water supersaturated for Ca, Mg, and bicarbonate.

Algal mats exist in an enormous range of environments and they have a remarkable ecological tolerance. These mats grow today under ecological conditions that are controlled mainly by light, water depth, salinity, temperature, sedimentation rate and invertebrate grazers. Stromatolites, however, are a product of extreme environmental conditions. Under normal circumstances the growth of the algal mats is limited by grazing snails. However, if the environmental conditions of the grazers are changed dramatically, e.g. due to increasing or decreasing salinity, the balance between algal growth and intensity of grazers is disturbed. In this case the algal mats start to create structures which can be preserved as stromatolites.

The stromatolites from Santorini were mainly collected and studied by one of us (U.E.). They contained several thousands of marine macro- and micro-fossils, which were found in the loose calcareous sediment trapped in cavities in some of the stromatolitic blocks (Fig. 8). Small hydrobia snails were very common in these cavities, which also included several specimens of the snail pirenella conica, together with foraminifera, ostracods and diatoms (see section 2.3).

     2.2.    Description of the stromatolites:

The stromatolitic blocks from Santorini can be divided into two groups. The first is massive calcareous with typical stromatolitic structure and the second is more spongeous with travertine structure, with or without dark lamellas. However, both forms have been observed on the same calcareous block.

The blocks were cut and studied in thin sections using standard light microscopy. They showed the typical layered stromatolitic structure, where calcareous material and thin layers of organic material alternate (Fig. 9). This organic material was so well preserved in some thin sections that cellular structures of algae are visible (Fig. 10). The calcareous material consists of aragonite needles which are arranged perpendicular to the surface of the lamella. They must have been precipitated in the interface of the solid material and water or inside the mat. The travertine type has a vesicular appearance with long tubiform structures. The calcareous material has been precipitated on the inner walls of the tubes; it might indicate that water saturated with dissolved carbonates had circulated in the tubes.

     2.3.    Fossils from the stromatolitic blocks

     Gastropods

Hydrobia neglecta Muus

Several thousand small snails were found in the loose sediment trapped in the cavities of the stromatolites. Some of them were incorporated in the lamellar structure of the stromatolites. All ages from juvenile to adult were observed which indicates that they had lived in this biotope. They normally live at a water depth of 1 metre to extremely shallow conditions. They can even survive shorter dry periods. They reproduce in salinity conditions of 10-30‰ but can also exist in extreme hypersaline conditions up to 70‰ (Table 3). The larvae of H. neglecta can survive in a temperature interval between 10-30° C (Fish and Fish 1981).

Pirenella conica (Blainville) Lozouet

This gastropod, which formerly was named Potamides conicus, is common in the above mentioned loose sediment. Pirenella is known as a limiting factor for the growth of stromatolites since it grazes on the algal mats. It demonstrates tolerance to extreme temperatures (5-45° C) and salinities (15-90‰) (Taraschewski and Paperna 1981). Pirenella conica cannot survive in marine inter-tidal or littoral habitats exposed to waves as observed in the Persian Gulf by Evans et al. in 1973. According to Lozouet (1986) Pirenella conica is very common on the bottom of lagoons between Ermioni and Plepi (Argolis) in the present day Aegean Sea.

     Ostracod

Cyprideis torosa Jones

The ostracod Cyprideis torosa is very common in all samples of loose sediment. This species is preferably observed in shallow waters such as lagoons and river deltas. Its geographical distribution ranges from Scandinavia to lakes in Central Africa. It has even been reported from thermal springs in Iceland (Elofson 1941). C. torosa has been reported to exist at water-depths ranging from a few centimetres to some decimeters, but it is also found at depths down to 30 metres (Elofson 1941). In the Gavish Sabkha (Sinai Peninsula) C. torosa occurs in 'sea water springs' and seasonally filled shallow water areas where the salinity is 65 to 130‰. C. torosa has been observed to live together with 130‰ Pirenella conica and stromatolites in the same area (Gerdes and Krumbein 1984).

     Foraminifera

Ammonia tepida Cushman

This species was only found in a few samples in the material from Santorini. On the basis of laboratory experiments it is known that A. tepida can exist in temperatures ranging from near freezing point up to 42° C. However, the temperature where it reproduces ranges from 23-32° C. It tolerates low salinities, even fresh water over shorter periods and fluctuating salinities ranging from brackish water to hypersaline conditions (8-67‰; Bradshaw 1961).

Agglutinated foraminifera were observed in several samples from Santorini, but they have not yet been studied in detail.

3.      STABLE ISOTOPES

The stable carbon and oxygen isotope compositions have been determined for the calcareous phase of stromatolites, travertine crusts, and gastropods. Results are shown in Table 4 and Fig. 11. Oxygen isotope data are all compatible with a marine origin. Eastern Mediterranean sea water is known to be slightly enriched in ¹⁸O due to evaporation, typical δ¹⁸O_water- values being from 1.5 to 2.0‰ SMOW (Buchardt 1974). Carbonate precipitated from such a water body at temperatures about 16°C will have δ¹⁸O_carb- values comparable to those observed here. Carbon isotope compositions are, on the other hand, atypically enriched in ¹³C. Carbonate precipitated in equilibrium with dissolved marine bicarbonate will have δ¹³C values between 0 and +2‰ PDB, and no marine processes are known to cause ¹³C enrichments of the magnitude observed here. It is, therefore, necessary to find other origins for the carbonate carbon which has been incorporated into the stromatolitic carbonates. Two sources are possible, both of volcanic origin: CO₂ formed by dissolution at elevated temperatures of Mesozoic carbonate rocks or magmatic CO₂ equilibrated with CH₄ at low temperatures. The first process has been suggested by Craig (1963) for extremely ¹³C-enriched travertines (up to +12‰) from the Tivoli geothermal area in Italy; the second process is well known in many geothermal areas, where CO₂ may be enriched by up to 50‰ as compared to CH₄ (Hoefs 1987).

It is interesting to note that the majority of both stromatolites and travertinic crusts probably formed from the same water phase. As seen from Fig. 11, all these samples except one (6536) have δ¹⁸O values between +1.5 and +2.5‰ and δ¹³C values between +12 and +13‰. These values suggest the precipitation environment to be marine, but affected by strong influx of isotopically heavy, volcanogenic carbon.

Two of the gastropods (6540 and 6541) and one stromatolite (6536) seem to have grown in water formed by mixing between the ¹³C-enriched sea water and isotopically light meteoric water as indicated by the slope of the mixing line. The two gastropod samples (6542 and 6543) have secondary overgrowth of carbonate, probably formed under normal marine conditions.

The anomalous carbon-isotope composition of stromatolites, travertinic crusts and gastropods indicates that influence of probably ¹⁴C-free volcanogenic CO₂ have been considerable, rendering the radiocarbon dates maximum age limits (as mentioned above).

It may be noted that if the stromatolites were formed shortly before the Minoan eruption, an apparent ¹⁴C age of approximately 20,000 years indicates that c. 85% of the carbonate were of volcanic origin. In this case the bicarbonate in the caldera water must have originated almost entirely from volcanogenic CO₂. This suggests extreme ecological conditions which may have contributed to the high growth rates of the stromatolites.

4.      DISCUSSION OF THE FOSSILS

Stromatolites are most commonly formed in hypersaline conditions where other organisms cannot exist. However, they are sometimes observed living in ecological communities together with other marine animals, like gastropods, ostracods, foraminifera and diatoms, etc. Blue-green algal mats are common in Gavish Sabkha (Sinai Peninsula) but high population densities of the gastropod of Pirenella conica inhibit potential stromatolites from developing. They first occur in areas above the salinity boundary of 70‰, which restricts the snails (Gerdes and Krumbein 1984).

On the basis of the above observations it can be concluded that the stromatolites and the associated animals coexisted in the same biotope. This biotope must have been a very shallow bay where the water depth most probably did not exceed a few metres. It was protected from the open sea, thus resulting in poor water circulation and a minimum of water exchange. This caused periodically extreme conditions of salinity ranging from nearly brackish to hypersaline water. However, the mean conditions were around normal marine, as indicated by the stable isotope investigations and by the fact that the snails found in the stromatolites can only reproduce in normal marine water.

The stromatolites from Santorini were not of the sediment-trapping type but were precipitating. The water must have been saturated with carbonate as indicated by the findings of travertine and, therefore, precipitation did not depend on the growth of the algae alone.

We conclude that the grazer snail Pirenella, which usually restricts the development of stromatolites, had no strong influence on the growth of the stromatolites from Santorini.

5.      CONCLUSIONS

The geological, palaeontological and stable isotope results, combined with the radiometric data, indicate that Stronghyle island had a water-filled caldera in its central part prior to the Minoan eruption (Friedrich et al. 1988). This caldera was filled with sea water, as indicated by the stable isotopes and the fossil invertebrates. The formation of travertine and stromatolites with a high ¹³C content suggests that volcanogenic carbonaceous material from the metamorphic basement has contributed to their formation. CO₂ from the degassing magma chamber dissolved old calcareous material from the basement on its way to the surface. This material could then be precipitated in the sea water on the loss of carbon dioxide to the atmosphere. This phenomenon could explain both the existence of the travertine and the stromatolites and probably also the very positive δ¹³C values recorded. A similar observation has been made on travertine from Tivoli in Italy, where extremely high δ¹³C values were observed (Craig 1963).

Taking the possibility of old carbon dioxide in consideration, the radiocarbon ages of about 20,000 years BP obtained from the carbonate of the stromatolites represent maximum age limits. Their real age could, therefore, just as well be quite close to that of the Minoan eruption. It may be noted that if the stromatolites were formed shortly before the Minoan eruption, an apparent ¹⁴C age of approximately 20,000 years indicates that c. 85% of the carbonates were of volcanic origin. In this case the bicarbonate in the caldera water must have originated almost entirely from volcanogenic CO₂. This suggests extreme ecological conditions which may have contributed to the high growth rates of the stromatolites.

Before the Minoan eruption there must have existed two caldera basins: a deeper, older one in the southern part which was connected with a shallow one in the northern part. The latter was formed in connection with the 18-K-ignimbrites (Druitt 1985). The basins were open to the sea in the southern part.

Since it is generally accepted that the first phase of the Minoan eruption started with a strong plinian phase, with no contact with the sea water, we conclude that this phase started on a small island in the central part of the caldera. This island, which may be named pre-Kameni, must have been situated north of the present day Nea Kameni island in the so-called Kameni Line (Druitt et al. 1989) where, according to Bond and Sparks(1976) the vent of the Minoan eruption was located (Fig. 12). During subsequent phases sea water came into contact with the erupting magma and produced the base surges and pyroclastic flows (Bond and Sparks 1976; Pichler and Friedrich 1980; Heiken and McCoy 1984). The third phase led to the collapse of a greater caldera area, and during that phase the vent reached the northern area of the present day caldera where the stromatolites grew in a shallow basin. They were ejected and deposited on the flanks of the volcanic island. The biggest were concentrated in the northern part of Thera and Therasia proximal to the collapsing area and the smaller ones in more distal regions (Fig. 6 and 7). Additional evidence for this reconstruction of the Stronghyle caldera may be obtained from a reinterpretation of the archaeological evidence, e.g. the distribution of Bronze Age settlements along the edge of the present-day caldera.

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

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