Mainly summarized from Heiken and McCoy (1984), Bond and Sparks (1976), Druitt and others (1989) and Friedrich (1994). Part of the masters' thesis by Tom Pfeiffer (1999), geology department Aarhus, DK, under W. Friedrich.

Introduction

The Minoan eruption happened around 1645 BC in the Late Bronze Age. It was one of the largest plinian eruptions in younger time. It erupted ca. 30-40 km³ rhyodacitic magma and is ranked VEI=6 (Volcanic Explosivity Index after Simkin and others, 1981). The eruption was followed by collapse of the magma chamber that enlarged an existing caldera.
The height of the plinian eruption column is estimated 36-39 km (Pyle, 1990). It dispersed tephra throughout the Eastern Mediterranean and might have led to global climatic impacts. Its deposits on Santorini consist of up to 50 m thick layers of white pumice and ash.

The eruption destroyed an inhabited and culturally high-developed island which perhaps might be the origin of the Atlantis legend as many scientists believe. Since 1969 excavations near Akrotiri have brought to light an important marine Cycladic town famous for its well-preserved and magnificent wall-paintings.

The Minoan eruption has been studied in detail and described by many authors. Among the most important works are Fouqué (1879), Reck (1936), Bond and Sparks (1976), Pichler and Kussmaul (1980), Pichler and Friedrich (1980), Heiken and McCoy (1984) and Druitt and others (1989) (Reference list).

Eruption and tephra sequence

Reck (1936) described 4 major units as BO₁, BO₂, BO₃ and BO₄, while Druitt and others (1989) call these Minoan A, Minoan B, Minoan C and Minoan D. Heiken and McCoy (1984) report thin basal units as BO₀ that represent precursory volcanic activities before the main eruptive sequence.
Every unit corresponds to a distinct phase of the eruption with its characteristic stile.

Precursory tephra fall unit BO₀

Heiken and McCoy (1984, 1990) describe up to 4 very fine-grained yellow, orange-brown and/or light gray layers of fine ashes, lapilli-sized rounded pumice and lithic fragments that are present in southern Thera. Their thickness is in the range of 1 to 4 centimeters.
They interpret them as air-fall deposits of phreatic and phreatomagmatic activity from a vent near the present-day Nea Kameni island that preceded the eruption with a short time interval in the range of some months. Thus, they possibly provided a warning to the inhabitants.

The fine-grained yellowish precursory ash deposit (BO-0) overlying the pre-Minoan soil as it appears in the quarry near Megalochori. A thin veneer of the first pumice fall layer has been preserved from intense quarrying.	In places, only the cemented yellow ash layer BO-0 has survived erosion and/or quarrying. Here, it overlies directly a block of the red Cape Riva ignimbrite (21 ka). Quarry of Megalochori.	The yellowish, 2-3 cm thick ash layer BO-0 here seen in the left of the image with a rest of in situ pumice fall deposit in the quarry of Megalochori.

First major eruption phase: Plinian pumice fall BO₁

Minoan ash is found throughout the Eastern Mediterranean. The easterly disperal axis of the tephra blanket indicates the main wind direction at the time of the eruption. After Friedrich (1994).	The first pumice fall deposit near Athinios, here about 5 m thick. Walter Friedrich points to holes in the deposit that once were the trunk and branches of a tree. In some of the holes partly charred wood remnants are found, showing that the pumice was still hot enough to burn the tree when it was falling.	Detail of a larger partly charred branch in the big hole visible on the left photo. C-14 dating as well as attempts to use the tree for dentrochronology are underway; preliminary results of the C-14 anaylses confirm an age around 1700-1600 BC. For more information please contact Tom Pfeiffer.

The first phase of the eruption is a typical pumice fall-out deposit from an estimated 36 km high eruption column. It ranges in thickness on Thera from 50-500 cm (Pichler and Friedrich, 1980) or 10 to 600 cm (Heiken and McCoy, 1984). It mantles uniformly the pre-Minoan surface which proofs its origin as fall-out. A clear southeasterly trend in greatest accumulations reflects the dispersal of the tephra-bearing eruption column by strong atmospheric winds (Bond and Sparks, 1976).
The greatest thickness and the largest pumice clasts up to 30 cm are found directly south of Fira in the pumice quarries. This and the position of the isopach lines allow to define the vent. Its position probably was somewhere west of Fira between Cape Katofira and the present-day Nea Kameni island (Pichler and Friedrich, 1980).

The deposit is widespread throughout the Eastern Mediterranean region. Finer particles were transported to great heights and could be transported to long distances. The deposit is clearly present in many deep-sea cores of the eastern Mediterranean Sea and was found in locations on other islands and in western Turkey (Watkins and others, 1978; Sigurdsson and others, 1990).

On Santorini the deposit consists mainly of massive pumice and is only poorly sorted. In most sections, however, it shows a slight reverse grading in clast size which suggests that the eruption was increasing in violence with time (Sparks and Wilson, 1990).

More than 90% of the deposit is coarse white-pink rhyodacitic pumice (sizes typically 0,5-30 cm, 70,5-71,4% SiO₂), rare crystal-rich pumice (52,5-63,6% SiO₂) and few cauliform-shaped clasts of gray andesitic scoria (58% SiO₂). Less than 10% are fine ash and lithic fragments (Druitt and others, 1989).
No signs of magma-water interaction or breaks and changes in eruption style occur; the eruption was a continuous dry event driven only by magmatic gas (Sparks and Wilson, 1990). This adds support to the view that the vent was subaerial and lay on the supposed Pre-Kameni island (Druitt and Francaviglia, 1992).

In its uppermost part a thin (2-40 cm) layer of very fine white ash occurs that Heiken and McCoy (1984) called the "phreatomagmatic break". Its interpretation is the first significant interaction of magma with sea-water that intruded into the enlarged vent and announced the end of this phase.

Second major eruption phase: Base surge deposit BO₂

The deposits of this phase consist of numerous individual beds, mostly white layers of pumice lapilli bearing ash, with abundant lithic blocks and fragments of up to 1-2 m size. Up to 90% of the total volume is fine ash and well-rounded pumice, the latter more present in proximal sections (Heiken and McCoy, 1984). Strong variations in structure and composition are present within the deposit. Lithic fragments and blocks reach up to 20% of the volume and mark prominent beds (Heiken and McCoy, 1984).
Eroding contacts to the underlying phase 1 deposits, cross-stratification, ripple-, dune- and antidune-structures and bomb-sags implying ballistic transport for the larger blocks are characteristic.

Large ballistic block from the second phase on top of Mt. Profitis Ilias, ca. 7 km away from the vent. Note that the block has split in two on impact and penetrated into the first pumice fall layer BO-1.

Pichler (1973) was the first to interpret the second phase deposits as pyroclastic base-surges produced by phreatomagmatic explosions. As the result of cracks, fissures and vent-erosion seawater entered the crater and produced violent explosions that pulverized the magma and ejected large lithic blocks. Similar explosions were observed during atom-bomb tests. These produced fast-travelling, ring-shaped clouds that expand and spread horizontally away from the center (Friedrich, 1994).
The bedding structures show that the bulk of the material was transported laterally within individual flows moving at high velocity (15-50 m/s) (Pichler and Friedrich, 1980). The deposits are controlled by topography. Close to the vent they climb up slopes of 10-30E and reach heights of 200-400 m (Heiken and McCoy, 1984), whereas further away they thin out considerably and are lacking on most of Mt. Profitis Ilias.
In the lowest parts of phase 2 up to 2 pumice air-fall beds are present which are similar to the phase 1 fall-out unit. They show that phase 1 activity continued for a while after the beginning of phase 2, or that there was a period of fluctuation in erupting style on the transition from dry to phreatomagmatic activity (Heiken and McCoy, 1984).
Measuring the degree of pumice vesiculation, Wilson and Houghton (1990) found that during the whole eruption magma had vesiculated before the level of fragmentation, thus concluding that water-magma interaction occurred above this level, in a shallow depth of possibly a few hundreds of meters.

Literature data for phase 2 thickness differ considerably. Pichler and Friedrich (1980) report 0,5-7 m, Bond and Sparks (1976) and Heiken and McCoy (1984) give 0,1-12 m. The difference could be explained by the gradual transition of BO₂ into the phase 3 deposits as evident in many sections.
From the point of their greatest thickness between Fira and Cape Athinios and from ballistic impact sags, Pichler and Friedrich (1980) conclude a vent position similar to that of phase 1. Heiken and McCoy (1984; in press) measured a number of flow directions within that phase and found a more or less radial pattern of flow-directions pointing from a central area south of Kameni Island.

View of a 25m-high exposure of Minoan tephra in the quarry near Megalochori at the caldera cliff. The first pumice fall phase BO-1 overlies the dark Minoan paleosoil (lower middle of photo), followed by a thick sucession of BO-2 cross-bedded dune-forming surges and lithic-rich ash-flows of BO-3 (upper part of wall).

Third major eruption phase: Ash-flows BO₃

At most localities the base surge deposits grade into a chaotic, unsorted, massive lithic-rich flow deposit that is most prominent along the caldera cliffs.
It consists of fine ashes, pumice and 25-30% lithic fragments and blocks up to 10 m or more in diameter. It thickens strongly into topographic lows (Friedrich, 1994). Along the caldera rim it reaches maximum thickness of 40 m (Pichler and Friedrich, 1980), 55 m on southern Thera (Heiken and McCoy, 1984) and thins out rapidly at greater distances. On the Profitis Ilias Mountain it is missing (Heiken and McCoy, 1984).
Within the matrix of usually very fine vesiculated ash occasional degassing pipes occur. This implies emplacement with 3 components: pyroclasts, liquid water and gas (steam) (Sparks and Wilson 1990). Lithics and larger pumice blocks sometimes show concentrations that can be traced, horizontally and some few hundreds of m downflow. They define individual flow units (Sparks and Wilson, 1990).

Only a few of the larger lithic blocks show clearly ballistic impact sags; most of the lithics were obviously transported by flows (Sparks and Wilson, 1990). The flows must have been very energetic, as they are present on slopes as steep as 30E on the upper flanks of Micro Profitis and Megalo Vouno Mountain and because many intraclasts are sheared in the flow direction (Heiken and McCoy, 1984).
Studies from Wright (1978) and McCleeland and Thomas (1990) report different results concerning emplacement temperatures. Some of the deposits seem to have been emplaced cold, others at temperatures up to 400E C.

The origin of the phase 3 deposits is controversial:
Bond and Sparks (1976) interpret the deposits as mud-flows due to apse of a giant tuff-ring built up by the surge flows.
Pichler and Friedrich (1980) consider them rather ash-flows that were produced from an extremely enlarged vent and like "boiling milk" poured over the caldera-rim. The large amount of lithic fragments is seen as the result from beginning caldera collapse.
Heiken and McCoy (1984, p.8454) recognize a multiple emplacement facies comprising:
- ballistic (proximal, on the caldera rim),
- base surge (proximal, in the south-western corner of the caldera complex),
- mud-flows (distal, on the outer flanks of most of Thera and all of Therasia) and
- slumps (distal, on steep slopes of the Profitis Ilias Mountain).

Fourth major eruption phase: Ignimbrite or reworked (?) BO₄

Phase 4 deposits differ from phase 3 by a subtle color change from white to cream colored; they are finer grained with smaller lithic blocks and pumice clasts; the total lithic concentration, however, is significantly higher (34-50%) (Bond and Sparks, 1976). Phase 4 contains abundant different flow units usually marked by concentrations of normal-graded lithics or erosional contacts (Bond and Sparks, 1976). On the caldera rim the deposits are thin (0,7-2 m, Heiken and McCoy, 1984), if present at all, but form thick fan-shaped units up to 40 m thick on the coastal plains (Bond and Sparks, 1976).

Their origin is controversial:
Pichler and Kussmaul (1980), and Friedrich (1994; in press) believe that they are reworked phase 3 deposits. Processes like intensive flooding during and after the caldera collapse, rainfall, tsunamis, wind, agriculture and other factors could have eroded and washed away large parts of the Minoan tuff. Especially where it originally was deposited on steeper slopes, it would be easily removed by such processes and deposited in the coastal plains. On the outer flanks of Profitis Ilias, Mikro Profitis Ilias and Megalo Vouno Mountain the effects of erosion are visible: the Minoan tuff has virtually disappeared.
Bond and Sparks (1976), Heiken and McCoy (1984) and Sparks and Wilson (1990) argue that the deposits are ignimbrites, hot gas-rich and highly fluid pyroclastic flows that only came to rest "on low gradient (1-2° ) slopes" (Bond and Sparks, 1976, p. 7). The eruption-sequence from Plinian fall-out over phreatomagmatic deposits to ignimbrite-forming styles is a frequent observed feature.
They support their interpretation by the presence of typical lag-breccia, gas-segregation pipes, the existence of many individual flow units which follow the topographic valleys, and occasional co-ignimbrite ash-fall deposits interbedded within the flow units described by Sparks and Walker (1977). Furthermore, studies of Wright (1978), Downey and Tarling (1984) and McCleeland and Thomas (1990) indicate hot emplacement temperatures from 200 to 400E C.
The primary magmatic origin of the phase 4 deposits is questioned by some authors (e.g. Friedrich, 1994, 2000).