Radio-Isotope Analyses of Aegean Tephras: Contribution to the Dating of Santorini Volcano

The aim was to establish the provenance of the tephras found overlying Minoan settlements. Distinct differences were found between different tephra layers in a profile at Yali which overlie a palaeosol whereupon Neolithic implements were found. Most interesting is the similar tephra deposit at Kos (Kefalos site) and at the top layer at Yali. Other deposits on the Minoan settlements were similar to the sampled sections on Santorini and Yali. The work presented here is ongoing and the results presented here form the first stage of the programme.

INTRODUCTION

It is well known from the excavations of the late Professor Marinatos that the Bronze Age town of Akrotiri on Thera (Santorini) came to an abrupt end about 1500 BC while the Minoan civilization of Crete suffered a catastrophic destruction somewhat later, at about 1450 BC (Marinatos 1939).

The role of the volcanic eruptions of Santorini in the decline of the then flourishing Minoan civilization is still enigmatic. The exact relationship between the paroxysmal volcanic eruption(s) of Santorini or other Aegean volcanic islands and the Late Minoan IB destructions on Crete and other Minoan centres is an ongoing debate (see Doumas 1978, 1980). The effect of earthquakes or even Mycenaean invaders have also been considered as possible causes of the decline.

The time-gap between the earthquakes, the volcanic eruption(s) as well as the complexities of the relationship between LM IA and LM IB pottery styles are still matters of question (Pichler and Schiering 1977, Sparks et al. 1978; Doumas 1974; Page 1970; Doumas and Papazoglou 1980; Keller 1980; Aitken 1987; Doumas 1980 and references therein).

In our research we will tackle the problem of multi-stage volcanic eruptions by radio-isotopically analysing tephras from various Aegean islands; in so doing we will try to place them in a chronological order of events.

THE PROBLEM IN PERSPECTIVE

Gamma-spectroscopy and beta-spectroscopy measurements of tephras from the Aegean islands of Santorini, Rhodes, Kos, Yali and Nisyros were carried out. For Santorini and Yali, in particular, samples were taken along the profiles of their respective tephra deposits in different parts of these islands and at closely spaced stratigraphic intervals.

Thermoluminescence (TL) dating will be applied to heated stones derived from the Yali volcano (work in progress).

The overall objective is to define similarities and differences amongst and between the whole set of samples from all the sources, based on their radio-isotopic concentrations and ratios (i.e. regarding U-238 series, Th-232 series and K-40).

Chemical analyses and refractive index measurements have previously been used to distinguish tephra layers as well as their stratigraphic position, colour and thickness (Vitaliano et al. 1980). As ash layers are very much alike, no attempt to distinguish them based on major element analysis or refractive index is necessarily effective since different eruptive phases can produce similar recycled material. This might differ in its radio-nuclidic content, which is also documented from studies of variations in radio-nuclidic concentration in volcanoes during their magmatic evolution.

Such variations would imply: 1) that the tephra may have been associated with lava flows fed from different sources; 2) the primary magma could not have reached isotopic homogenization; or 3) contamination of primary magma by marine sediments, sea water or by minerals formed by the interaction of lava and sea water.

Thus, whatever the reasons, any radio-isotopic differences in the tephras would indicate certainly a different volcano or several eruptive phases of the same volcano (Condomines et al. 1976, 1980, 1981; Allegré et al. 1968).

Taking these into account, we hope therefore to attribute the tephras in the Aegean islands either to respective ash layers on Santorini, Phira Quarry (reinforcing conventional theory) or to the eruption of the other volcanoes such as Yali (including adjacent islets).

The latter volcanic islands have not hitherto been given appropriate consideration. The reasons why we believe that the Yali volcano may have erupted contemporaneously with the Santorini event are as follows:

1. Both islands have tephra layers or deposits.
2. On a preliminary reconnaissance visit to the island of Yali by one of us (YL), with archaeologist Dr A. Sampson, who discovered the Neolithic site, and geologists, a tephra deposit was located above the Neolithic settlement in Section A. As excavations were ongoing in Section A, it was decided to take tephra samples from below and above the palaeosol which was clearly evident in an open section (Fig. 1). In order to see if this section A was indeed recent, samples of tephra were obtained from the north-east part of the island on a previous crater edge near Kamara, as Brousse (1986) believes that the tephra there is post-Neolithic in date at least (Fig. 2, 3). It is intended to apply TL dating to fired stones from this volcano and we will here seek to determine the date of the post-Neolithic eruption of the Yali volcano and the associated tephra deposit which might possibly be Late Minoan in date (Doumas 1980; Sampson 1988). Similarly, we will attempt to date fired stones and upper tephra deposits from Pergussa on Pachia (in progress at time of writing).
3. Both islands are sited on the same volcanically active zone in the Aegean.
4. The south-eastern distribution of tephras from Santorini volcanic eruptions may not be simply attributable to the effect of prevailing winds at the time but rather to the effects of proximal and simultaneous eruptions of Yali and Santorini.
5. Dating of pumices and obsidians from Yali have provided dates in the Upper Pleistocene (Wurm II a-c) but, to our knowledge, no attempt has been made to date later volcanic phases from the islands. All the above will test the conventional theory (see introduction).

The attribution of the various islands' tephras to the Santorini volcano might help to answer two questions:

 a) whether or not they all come from a single eruption? or

b) if they correspond to different eruptive phases of Santorini and to appropriate layers in its quarry tephra deposit.

The implications of the results of this work relate directly to the causes of the end of Minoan civilization. These preliminary results will help delineate, with somewhat greater resolution, the geological sequence of events synchronous with the collapse of the Minoan cultures in the Aegean (the work is ongoing at the time of writing).

The above mentioned provide a detailed précis of the problem we intend to tackle. The use of radioactivity on the provenance of tephras has been mentioned earlier (Liritzis et al. 1983). In this case the alpha-spectroscopy method was employed. Both beta and gamma spectroscopy are employed here for the first time. All radio-isotope measurements offer an alternative approach and may add to the analytical methods of provenancing volcanic tephras.

There are particular reasons for believing that radioactivity measurements can be effective determining tools for volcanic tephras and these are summarized below.

USE OF U, Th, IN MAGMATIC DIFFERENTIATION

The petrochemical evolution trends observed in a single igneous suite and the genetic relationships amongst different sequences are classically discussed in terms of the bulk chemical composition and the phase equilibria. Several determinations of elements contained in trace amounts have also been carried out in the last twenty years or so.

Many trace elements have been observed to have highly variable concentrations in different rock types. The study of their distribution in well-known suites has furthermore shown that they behave typically in different evolution series.

Sensitive variations of some trace elements (including U and Th) have also been observed in rocks having the same bulk composition but different magmatic histories. All these findings have shown that relative and absolute abundance of trace elements in rocks can reveal much useful petrologic information and increasing interest is being devoted to their application to the study of magmas and/or volcanic tephras.

YALI AND ITS TEPHRAS

Yali together with the small islands of St. Antonios, Strongili, Pergussa and Pachia, all belong to the volcanic complex of Nisyros. Yali, Nisyros, Kos and Santorini form part of the south-eastern section of the Aegean volcanic arc (subduction zone: plate convergence between African and Eurasian tectonic plates). So far, no detailed dating investigations have determined the chronological stratigraphy of tephra deposition in Yali. The tephra and pumice deposit in profile A could have been deposited in either of the following two ways.

1. Pentarakis and Markoulis (1974) attribute the deposits to the result of aeolic action during the volcanic activities which created Pergussa and Pachia islands (Fig. 2).
2. The French volcanologist R. Brousse (1986) believes that profile A is made from a recent geological rearrangement of Yali - that is, the tephra fall took place during minor volcanic eruptions of a recent age, i.e. at least post-Neolithic.

Concerning the now submerged centre (caldera), Pentarakis and Markoulis locate it between Yali and Nisyros, while Brousse believes that there were several minor centres which created volcanic domes in the north-east part of Yali island, and relates these to another centre between Yali and Kos (Pasteels et al. 1986).

On the other hand, Di Paola (1974) believes that the islands of Pergussa, Pachia, Strongili and Yali developed independently of one another (Fig. 2).

Perhaps it is not coincidental that the volcanological history of both Yali and Santorini were similar. The question remains: was it synchronous?

SAMPLING

     Santorini.      The tephras were taken from the Akrotiri trench, along a profile of around 2 m, which correspond to the layered tephra of the nearby Phira quarry (Fig. 4). A further pumice sample was given to us by Dr Francaviglia and comes from the Bo and Chania pumice series, and another from Chania, Crete (Francaviglia 1986).

     Yali.      The tephra samples derive from a profile, section A, which was separated by two palaeosols and from the slope of the crater-like mound in the entrance to the island's port on the east coast (Fig. 2). Above palaeosol 2, Neolithic ceramic sherds and obsidian tools or debris from workshop were found in abundance (Fig. 1). In an adjacent location on the same site, the palaeosols can be related directly to palaeosol 3 as in Fig. 3. This is section E.

     Rhodes.      The tephra samples were taken from tephra deposits layered along overlying Minoan levels at Trianda, to the north side of the town at the footsteps of Filerimou, at the airport of Rhodes (Paradisi), in the city of Rhodes (Daskalou site), and at the Koumelo site, Archangelou.

     Kos.      Tephra samples were obtained from the excavations at Kefalos site and from the eponymous town on the island.

   Nisyros.      Tephra samples came from a cross-section near the volcanic vent and on the road to Mandraki-Palli.

The sampies from the excavated sites were collected in plastic bags and care was taken to avoid contamination with soil. The thickness of the layers was in general no more than a few cm (up to 30 cm).

Most samples analysed are volcanic ashes (< 2.5 mm in diameter) and these range in texture from coarser (> 60 μm) to fine grain (< 60 μm).

MEASUREMENTS

• Gamma and beta spectroscopy:

Measurements were carried out by gamma spectroscopy employing a 5" diameter by 4" thick NaI (TI) detector and by beta spectroscopy employing a 2" diameter by 2 cm thick plastic detector NE102A. For the former, the activity of the following isotopes were measured: Bi-214; Ac-228; Tl-208 and K-40 (see Tables 1, 2). The efficiency calculation of the scintillator was made through the appropriate standard ores: uranium, thorium and feldspar for potassium. The isotopes were measured from their pulse height. In the calculation of the net activity, the samples (and standards) were, of course, normalized to a common thickness. The procedure will be described in full elsewhere (Galloway and Liritzis 1989, in preparation). For the beta dose-rate measurements the detailed procedure is described elsewhere (Liritzis and Galloway 1989, in preparation).

Table 1 shows the sample/site number and the U, Th and K contents derived from the NaI detector. The isotopic data are given in pCi and in their converted form as ppm. For the U, this conversion assumes secular equilibrium. However, in tephras U-disequilibrium has been observed (Ra/U < 1), (Liritzis and Danali-Cotsaki 1988).

Therefore, the U ppm values are equivalent-U values. Disequilibrium in the U-series will cause a difference between the beta dose-rate calculated from the gamma-ray measurements and the measured beta dose-rates. Only eight out of twenty-five comparisons agree within 2σ (random plus systematic errors), indicating that disequilibrium in the U-series is common.

The higher beta dose-rates measured by the plastic detector are compatible with a low Ra/U activity ratio, which provides a false low beta dose-rate contribution from the U series (Table 2).

Gamma-ray readings were also taken by a portable threshold scintillometer (miniscint UG130, Canada), operating in five windows for U, Th, K and in two modes of total counts.

DISCUSSION

For comparison purposes, five parameters are used, i.e. the activity of Bi-214, the average activity of Ac-228 and TI-208, the K₂O % content, the measured beta dose-rate (β^plastic), and the ratio of measured beta dose-rate to beta dose-rate calculated from the NaI data (β^plastic/β^calc.NaI).

Although five parameters are used, they are not all independent; thus the close dependency between dose-rate and K₂O % content is as would be expected. Each value of the five parameters was then compared with the average of all, taking as typical (T) all that fell within ± 2 x (the error attached to the measurements) of the mean, the outliers being the higher (H) and lower (L) ones.

The band for typical (T) values is 0.9-1.4, 0.9-1.5, 1.9-2.5, 3.05-3.75 and 1.1-1.45, respectively for the above five parameters.

Fig. 4 and 5 are schematic representations of the sample locations together with their corresponding beta dose-rates and isotope activities (Tables 1, 2).

The samples are first grouped according to their volcanic origin and the sites where they were found (islands of Rhodes and Kos).

A similarity is apparent between the following volcanic samples: a) YP2, YP3, YP5, (YP9), S1; b) YP8, NIS1, 2; c) YP1, YP6, S3; d) YP4, YP7, YCR, PBØ, CHAN 10; e) S5; and for the tephras found in the islands: f) RT4, RT6; g) RT3, RD7; h) K9; i) RTS, RA2A; j) K8.

Τable 3 shows most similar groupings amongst all samples based on the five parameters. The following groups and comments may derive from these.

• Groups:

     A.      The two samples from Rhodes at Trianda (RT4, 6) resemble Yali 2nd, 3rd and 5th layers and that of Santorini no. 1.

     B.      A similarity exists between Yali no. 8, a tephra layer below palaeosol 2, the two Nisyros samples and two samples from Rhodes at the Daskalou and Trianda sites. This may imply that the tephra fall at the two Rhodes sites represents an earlier eruptive event.

     C.      The next similarity group is Yali nos. 1 and 6 with Kos (at the Kefalos site) and Santorini no. 3.

     D.      Yali no. 4 layer is next in order of similarity, with Yali no. 7, Yali crater and pumices PBØ from Santorini and that from Chania (CHAN10). (A similar Thorium value for pumice deposits from Santorini was reported in the recent work by Druitt et al. 1989.)

     E.      Another group appears to be Santorini no. 5, Rhodes at Trianda no. 5 and Rhodes airport (RA2A).

• Comments:

1. The three Santorinian samples S1, S2 and S5 from the Akrotiri trench tephra imply different consecutive eruptions of short duration.
2. The Yali tephra layers nos. 2, 3, 5, 9 are most similar, the next similarity set being nos. 1 and 6, followed by nos. 4 and 7.
3. Yali 5 and 9 differ by two parameters (Th and K) out of five. Yali 9, found between two palaeosols, could have been slightly contaminated by soil and may have been deposited there by aeolic action from Yali no. 5 once being the top of the tephra profile.
4. Kos no. 8 seems to be a mixture, probably of tephra and soil, though it may belong to group A.
5. The sampled profiles in Yali and Santorini consist of tephra layers with different radioactivity, possibly implying different eruptive phases, recorded on the neighbouring islands. The latter may indicate occasionally simultaneous eruption(s) of both Yali and Santorini volcanoes, with a communicating volcano chamber carrying convective currents of similar mechanism.
6. The above differences and similarities in Yali tephras are also supported by the portable threshold scintillometer results.
7. Earlier measurements of beta dose-rates by a gas-flow counter on similar tephra samples and some other samples provided similar conclusions (Liritzis and Danali-Cotsaki 1988). The beta dose-rate, however, did not effectively discriminate between S1, S3 and S5 (THT, i.e. typical, high, typical, respectively) which in our present work exhibit some differences based on five parameters.

Certainly the tephra deposits on the islands of Kos and Rhodes resemble either those at Santorini or those at Yali or both. One way to differentiate between the two volcanic origins is to date by thermoluminescence the eruptive phases of the Yali volcano (including Pergussa on Pachia) presumably of post-Neolithic times, a project which is in progress.

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

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