Mechanism of the Minoan Eruption of Santorini

Previous calculations of the volume of erupted tephra are shown to be far too high. The discrepancy between data from the surrounding islands, including Crete and from deep-sea cores, is explained as due to the drift of pumice with ocean surface currents.

MORPHOLOGY OF STRONGHYLE BEFORE THE MINOAN ERUPTION

The present ring-islands of the Santorini group, namely Thera, Therasia, and Aspronisi are the remnants of a former single island, which, according to its originally circular shape, has been called Stronghyle (that means round). This island was mainly built up by the products of at least ten volcanoes among which the Thera volcano was the most prominent. The complicated volcanic history of the Santorini islands, which commenced about one million years ago in the Lower Pleistocene (Seward et al., this volume), is presented in greater detail in a separate contribution (Pichler, this volume).

The great Late Minoan eruption of the Thera volcano was preceded by a long period of volcanic quiescence, which lasted about 15,000 years (Pichler & Friedrich 1976; Friedrich et al. 1977). During this period Stronghyle island reached a height of at most 500 - 600 m. in its central part. Here the remnant of an old crater most probably formed a flat and wide depression covered by bushes and small trees forming a rather patchy forest (Rackham 1978; Friedrich, this volume). The central region as a whole was destined for wild-life, which is shown by the findings of the bones of red deer and hares (Gamble 1978). Because Stronghyle resulted from the agglomeration of at least ten different volcanoes, the surface of the island was not regular and approximately conical as in the case of a single strato-volcano, but irregular and divided into ridges, depressions, and ravines.

Field studies of the volcanic sequence underlying the products of the Minoan eruption give an idea of the pre-Minoan morphology of the island. The height of the pre-Minoan surface above sea-level and the dipping of the underlying strata show that a bay must have existed in the southwestern part of Stronghyle between the present islet of Aspronísi and the facing Cape Aspronísi. This bay extended about two kilometres in an eastnortheasterly direction and may have served the Minoans as a harbour. A second, more protected but smaller, harbour existed in a bay directly west of the Minoan town of Akrotiri. The bay, which was at most 200 metres wide, was situated between two parallel north-south ridges. The western of these two ridges, the Mavrorachidi hill, reached further to the south than the eastern ridge and thus formed an excellent natural mole for this Minoan harbour.

Stronghyle island was smaller than the present Thera and Therasia in the northwest, north, east, and south. The Monólithos cliff, now surrounded by Late Minoan pumice, was a rocky islet.

THE MINOAN ERUPTION

• Air-Fall Pumice Phase (Bo 1)  (Plate 1, part B)

Sometime between 1700 and 1500 B.C. there occurred a catastrophic earthquake with its hypocentre below Stronghyle island, destroying the Minoan settlements there.

This earthquake was caused by volcano-tectonic events indicating that rising magma was intruding an old blocked vent or a fissure. After a short pause of some months to two years at the utmost, there followed the volcanic disaster. It started with a paroxysmal eruption by which a conduit was opened. Immediately after blowing out the plug of the vent, gas-rich rhyodacitic magma was fragmentized into pumice and ashes all of which were ejected within a few hours. The whole of the conduit was cleared, as is indicated by the presence of xenoliths from the walls of the vent and/or the roof of the magma chamber. The bulk of such fragments of older lavas naturally occurs in the lowest parts of the Minoan volcanic sequence. In the Akrotiri excavations, for instance, these small lava fragments, mingled with the air-fall pumice, are particularly frequent in the basal parts of the Minoan volcanic deposits.

The eruptive pine-tree shaped cloud must have risen to a height of many kilometres. A rain of pumice lumps fell upon Stronghyle island and the sea around it, while the ashes were carried by the wind in southeasterly directions. Most of it fell into the sea, but small amounts fell on other islands some distance away.

The whole Minoan volcanic sequence is called the Upper Pumice Series (abbreviated Bo). The product of the first phase of the eruption consists almost entirely of typical air-fall pumice (fig. 1) reaching a thickness on Santorini of 0.5 - 5 m. This layer is called Bo 1 and consists mainly of slightly reddish pumice lumps mixed with some ashes and lithic fragments of older lavas and non-volcanic basement rocks. On account of its pink colour this layer was also named "Rosa Bimsstein" by Reck (1936) and "Rose Pumice" by Vitaliano et al. (1978). It was also recognized by Reck (1936) and later on confirmed by Gunther (1972), Gunther and Pichler (1973), and Bond and Sparks (1976). The Bo 1 layer reaches its greatest thickness directly south of the town of Fira (in the pumice quarries). Here are also found the coarsest lumps of pumice and the greatest lithic blocks, both up to a maximum size of 30 cm (Gunther 1972; Bond & Sparks 1976). These facts indicate that the vent must have been situated west of the present-day town of Fira between Cape Katofira and Nea Kameni island.

The Bo 1 pumice is also found on the top of the Profitis Ilias limestone massif (568 m), thus confirming its air fall origin.

• Base Surge Phase (Bo 2)  (Plate 1, part C)

At the end of the first (Bo 1) phase which had lasted a few hours the mode of the Minoan eruption changed dramatically. This occurred as a result of the opening of cracks or of similar volcano-tectonic events which allowed seawater to enter the crater or, along radial fractures, into the vent (pichler 1973). Thus water came into contact with the melt and the eruption rapidly increased in violence and intensity. A series of strong phreatomagmatic outbursts pulverized the magma into ashes and caused so-called base surges, i.e. ring-shaped basal clouds which swept outward from the crater with high velocity (15 - 50 m/sec). In this type of eruption, which on Thera was first recognised by Pichler (1973), the bulk of the volcanic material is laterally and not vertically ejected. In short, the second phase of the Minoan eruption was marked by the predominance of horizontally expanding eruptions, the transport of a great deal of the volcanic material by way of density flows near the surface, and by the fact that a part of the material was in wet condition when it was hurled out. Thus, since large amounts of ash were not whirled up to great heights, heavy ash falls did not occur on Crete and the neighbouring islands.

During this second eruption phase, forming the middle part (Bo 2) of the Upper Pumice Series great blocks of the walls of the vent were thrown out. These penetrated as bombs into the ashes or, as can be seen in the Akrotiri excavations, they struck and demolished Minoan houses (fig. 2).

The base surge phase (Bo 2) probably lasted only a few days to some weeks. A series of strong explosions, similar to the eruptions of Surtsey/Iceland in 1964 or of Capeliňhos, Azores, in 1957, produced dark cock's-tail-shaped eruptive clouds heavily laden with ash, pumice lapilli and a few lithic blocks. White clouds, consisting of seawater converted into steam, stood some kilometers high above the conduit.

The base surge deposits of the Santorini Islands are characterized by poorly sorted, but roughly inclined-stratified ashes and pumice. Grain size often varies strongly from stratum to stratum. Typical are cross-stratifications, antidunes and mega-ripple structures (fig. 3). Some discordances within the Bo 2 sequence were caused by single blasts which blew away some of the formerly deposited strata. An obvious bomb sag horizon (fig. 3) and some small strata of air-fall pumice lapilli and ashes are intercalated into the base surge deposits.

The thickness of Bo 2 ranges between about 7 and 0.5 m; we have nowhere found thicknesses of 12 m, as indicated by Bond and Sparks (1976). The greatest thickness exists between the town of Fira and Cape Athinios (Gunther 1972), proving again that the vent was situated between the present-day Nea Kameni island and the caldera walls west of Cape Alonáki. Parts of the Bo 2 sequence also occur, in strongly reduced thickness (up to 30 cm), on the Profitis Ilias limestone massif. However, on Profitis Ilias are the Bo 2 strata of air-fall origin, while typical base surge deposits do not occur, supporting the view that the bulk of the Bo 2 sequence was produced by a laterally directed mechanism which transported the material outward from the crater near the surface. The temperature of the Bo 2 ashes and pumice at the moment of its deposition must have been very low, probably near that of the seawater (around 20 - 25° C).

• Ash Flow Phase (Bo 3) (Plate 2, part D)

The base surge deposits (Bo 2) pass gradually upward into the bulk of the Upper Pumice Series, i.e. into the ash flow deposits (Bo 3; fig. 3, 4). This means that the base surge eruptions evolved - without a break of time - into a mode of eruption which resembles milk boiling over. Gas-rich suspension-clouds, heavily laden with ash, pumice and lithic blocks, poured out from the elongated crater, which became more and more enlarged due to the beginning of the caldera collapse. The erupted material was spread laterally from the vent, as in the Bo 2 phase, "flowing" from the crater down the slopes of the island into the sea. Vast quantities of ashes, pumice and lithic blocks, produced by at least five big ash flow eruptions, were carried into the water. The great quantities of huge lithic blocks, which are embedded in the chaotic ash flow deposits (fig. 5, 6) clearly demonstrate that the eruptions and the beginning of collapse of the island (caldera formation) occurred simultaneously. Sometimes collapsing solid material from the subsiding volcano was fragmented on falling down into the crater. If such material fell into a just emerging ash flow cloud, it could happen that it was so thoroughly encapsulated by the ashes that the block, though brecciated, was not dispersed (fig. 7). About 25 to 30 per cent by volume of the total Bo 3 material consists of lithic blocks of fragmentized older lavas and pyroclastics.

This great quantity of lithic blocks within the ash flow material led Bond and Sparks (1976) to the erroneous conclusion that the bulk of the Bo 3 sequence represents mud-flow deposits.

The ash flow deposits reach their greatest thickness south of the town of Fira (up to 40 m; fig. 4). At a distance from this area and on the higher parts of Thera the thickness diminished rapidly. For example, around Merovigli (directly north of Fira at a height of about 320 m) there are only 4 m, while on Mount Profitis Ilias the Bo 3 deposits are totally lacking. This again supports the view that also the Bo 3 material emerged laterally but not vertically. Thus, as in the preceding Bo 2 phase only a very limited quantity of ashes was ejected to great heights. Consequently the rate of ash fall at greater distances from the volcano must have been significantly smaller than supposed by many authors.

After the end of the ash flow eruptions the collapse of the caldera along radial and concentric fractures continued, thus in time forming the present-day Santorini caldera (plate 2, part E). The upper parts of the Bo 3 deposits were eroded, reworked and transported by eruptive and post-eruptive rains at the end of the Bo 3 phase, a process which continued during the post-eruptional stage. Due to these exogenic events the seaward parts of what are today Thera and Therasia, became enlarged, especially in the north, east, and southeast.

As a whole, however, the old Stronghyle island was greatly reduced in size, since the collapse of the centre during the caldera formation left only three remnants: Thera, Therasia and Aspronisi.

Some hundred years later, the Minoan eruption volcanic activity started again. In the central part of the caldera the post-Minoan Kameni islands grew-up. During several eruptive phases between ?197 B.C. and 1950 A.D. these achieved their present shapes (plate 2, part F). This filling up of the caldera is still going on.

EFFECTS OF THE MINOAN ERUPTION

• Quantities of Erupted and Collapsed Material

Watkins et al. (1978) calculated that the Minoan eruption produced at least 28 km³ of tephra (13 km³ dense rock equivalent). This calculation is based on the thickness of Minoan tephra in east-Mediterranean deep-sea cores. According to Ninkovich and Heezen (1965, p. 433) the Santorini caldera which covers 83 km³ "was formed by the sinking of over 60 km³ of the island". If 60 km³ of rocks have disappeared from the surface, it seems logical that this material cannot have disappeared without trace, but must have fallen into the empty magma chamber. If, on the other hand, only 28 km³ of tephra were produced by the Minoan eruption, the question arises how it is possible that a magma chamber with a total volume of about 30 km³ can have held 60 km³ of subsiding rocks.

To resolve these discrepancies new calculations on the quantities of the erupted and collapsed material were performed. The approximate calculations are mainly based on the results of field work on Santorini:

The surface areas are as follows:

Field studies have shown that Stronghyle island reached a maximum height of 500 - 600 m. However, large areas, such as all of the southern part, excluding mount Profitis Ilias, did not exceed an average height of about 120 m. Thus the average height for the whole of Stronghyle can be estimated at 200 m. If we add the average depth of the whole caldera, which also amounts to about 200 m. a total of 400 m.  results. A collapse height of about 400 m referring to an area of 85 km² gives a rock volume of 34 km³. This is about half of the amount calculated by Ninkovich and Heezen (1965).

Our calculation of the volume of ejected material is based on the average thickness of the Upper Pumice Series on the Santorini islands and its submarine surroundings. If the maximum pumice average thickness of 30 m and an area of 250 km² are used in the calculation, then the tephra produced by the Minoan eruption totals at most 7.5 km³. Even if we double this figure it is still only half of the calculated tephra volume of Watkins et al. (1978). Even if we assume the area of the tephra fan in be 180,000 km² as calculated by Watkins et al. (1978) then an average thickness of 10 cm, gives a total volume of only 18 km³.

So large a fall of tephra is unrealistic, as has already been pointed out by Pichler and Schiering (1977). If we assume a tephra fall of 2 cm over an area of 180,000 km², the volume of the fallen ash would be 3.6 km³.

Taking into account the thickness of the Minoan tephra both on Santorini. and inthe deep-sea cores, we get a total volume of ejected material of only about 10 km³ (4.5 km³ dense rock equivalent). The calculated 10 km³ of Minoan tephra can be divided as follows:

.     Bo 3 (av. thickness 20 m) about 6.7 km³

.     Bo 2 (av. thickness 6 m)  about 2.0 km³

.     Bo 1 (av. thickness 4 m)  about 1.3 km³

• Ash Fall or Ash Drift?

There is also disagreement about the amount of Minoan ash which fell on Crete, Karpathos, Rhodes, and other islands and on the Aegean Sea in general. Watkins et al. (1978) calculated from the thickness of Minoan tephra in the deep-sea cores that a significant amount of ash (up to 5 cm) must have fallen on these islands. Thorarinsson (1978) calculates an ash fall of 7 to 11 cm for Kato Zakros on Crete and between 0.5 and 4 cm for Knossos. Even if as little as 1 - 2 cm fell on these islands, remnants of this layer of white ash should have been preserved and should be visible to the naked eye in recent soil profiles. A special search for such remnants of Minoan ash on Crete and Karpathos was undertaken by Pichler and Schiering in 1976 and 1978, and by other colleagues. This field work showed that there are no visible traces of a Minoan tephra layer. Pichler and Schiering (1977), and Schiering (1978), therefore argued, that the ash fall on Crete, Karpathos, Rhodes, etc., and on the Aegean Sea must have been much less than calculated by Watkins et al. (1978), Thorarinsson (1978) and others. On the basis of the fieldwork it can be estimated that the ash fall did not exceed a thickness of 5 mm.

Traces of Minoan ash on Crete have only been found by microscopic research (Boekshoten 1971; C. & D. Vitaliano 1974; Pichler & Schiering 1977; D. & C. Vitaliano 1978 and others). The fact that these pumice particles occur mainly in archaeological sites and inside houses (Cadogan & Harrison 1978), suggests that many of the tiny glass particles do not represent air-fall material, but have resulted from the use of pumice lumps as abrasive tools (Pichler & Schiering 1977).

The fact that the Minoan tephra fall on Crete and the surrounding islands probably only achieved a thickness of millimeters leads to the conclusion that the amount of ash falling on the Aegean Sea must have been correspondingly quite small. On the other hand the thickness of Minoan tephra in deep-sea cores is in the order of centimeters, though it varies strongly from one sample locality to another, as indicated in plate 3.

What is the reason for this marked difference of the tephra in the deep-sea cores and on the islands?

We pointed out above that of the total mass of the Upper Pumice Series, only about one tenth - represented by the Bo 1 layer - was ejected more or less vertically. Such vertical ejection is essential if substantial amounts of ash are to fall at considerable distances from the volcano. But nine tenths of the Minoan tephra, represented by ash-flow and base surge deposits, was produced by laterally directed eruptive mechanisms.

From these lateral eruptions huge quantities of pumice and ash "flowed" down the slopes of the volcano and were transported into the sea. Pumice floats on water (fig. 8) and can therefore drift away with wind-directed surface currents. During the drift the ash slowly sinks down to the sea bottom. The amount of such tephra sedimentation varies greatly within short distances because it depends on many circumstances, such as the quantity of floating and sinking material, the action of waves, the direction and the velocity of surface and sub-surface currents, and the nature of slumping or reworking phenomena at the bottom. As the surface currents in the Aegean are mostly directed to the southeast (plate 3), they coincide with the axis of the tephra fall distribution fan of Watkins et al. (1978).

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

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