Palaeotopographic and Palaeogeologic Reconstruction of Minoan Thera

A reconstruction was done by first 'stripping' the Minoan Tuff, wherever the tuff thickness is known or can be inferred, and noting the elevation and lithology of underlying rock and soil units. Rock units exposed by this stripping process include the pre-volcanic metamorphic rocks, older volcanic rocks of the Akrotiri peninsula, pyroclastic rocks of the Middle Tuff Sequence (MTS), and lava shields of the northern Thera volcanic field.

Additional evidence for palaeotopographic reconstructions comes from field relationships for both the MTS and Minoan Tuff. In southern Thera, the MTS originated largely from phreatomagmatic eruptions and forms wedge-shaped deposits plastered onto steeply sloping, inward-facing cliffs. Inward dips of the Minoan Tuff also document the presence of these palaeocliffs. These criteria, in addition to the distribution of the MTS, imply that a depression 6 km in diameter, flooded by the sea, was present before the Minoan eruption. This caldera most likely formed during eruption of the Lower Pumice Series, a tuff unit very similar in size and genesis to the Minoan Tuff.

Northern Thera was composed of three overlapping lava shields 300 m high, the Megalo Vouno, Therasia, and Skaros volcanoes, which are aligned along north-east-trending fissures. There also may have been a small caldera or calderas on one of these shields.

These observations conclusively demonstrate that Late Bronze Age Thera, before the Minoan eruption, was a complex volcanic field comprising overlapping shield volcanoes, cinder ridges and cones, tuff rings, and a flooded caldera 6 km in diameter. This caldera was bounded on the north by the lava shields and on the south by the present-day Akrotiri peninsula. Volcano summits were not more than 300 or 400 m above sea-level.

There is no evidence for a large composite cone or cones present before the Minoan eruption. The present-day flooded depression of Thera consists of two calderas and not one. The calderas were formed during the Lower Pumice Series and Minoan Tuff eruptions. They overlap approximately along a line coincident with the post-caldera domes of the Kameni islands. Our interpretation resolves the discrepancy between the volume of erupted Minoan Tuff and the volume of the caldera left by that eruption.

A flooded caldera adjacent to the Akrotiri peninsula, present during Minoan times, would have been an excellent natural harbour, well protected from most Aegean storms. The reconstruction, when viewed from the west, is similar to the island and harbour portrayed in the Flotilla fresco from Akrotiri.

INTRODUCTION

Calderas are usually formed late in the history of a volcanic field. Caldera collapse accompanies eruptions of large volumes of silicic (dacitic to rhyolitic) tephra, during which much of the pyroclastic ejecta is deposited within the collapsing crater. A caldera is usually superimposed upon a complex volcanic field that has evolved over several million years and consists of a variety of volcanoes, which range in chemical composition from mafic to silicic. Rarely, if ever, do large calderas form on the summits of simple composite cones (stratovolcanoes); even the well-known Crater Lake caldera of Oregon, USA was formed within a complex volcanic field and not at the summit of the mythical stratovolcano of Mount Mazama (Bacon 1983). At first glance, it is evident that the Thera volcanic field before the Minoan eruption was also composed of a large variety of volcanic landforms rather than one or two stratovolcanoes; this observation was made as early as 1879 by Fouqué.

Many calderas around the world have been formed during multiple eruptions - usually a few thousand to a few hundred thousand years apart. These calderas may overlap and form elongate depressions, as is the case for most of the caldera complexes of Central America, or exhibit multiple collapses of the same depression, as is the case for the Valles-Toledo calderas of New Mexico. Each period of collapse is recorded outside the caldera complex as a voluminous sequence of pyroclastic rocks.

The original purpose of our study was to establish the sequence of caldera collapse at Thera and to determine if there were multiple calderas, of which the Minoan eruption formed but one.

Marinatos (1972) and Pichler and Friedrich (1980) proposed that the Minoan eruption was solely responsible for the present caldera and that before the eruption, the island of Thera was composed of several large stratovolcanoes. They proposed that Minoan Thera consisted of a single island, which they called Stronghyle and had an elevation of 500-600 m. Because calderas are formed by collapse, the volume of material erupted should be approximately equal to that of the caldera; on Thera, therefore, the volume of magma erupted should have equalled the volume difference between Stronghyle and the present caldera. The volume estimate for tephra from the Minoan eruption of between 13 and 18 km³ (dense rock equivalent), was made by Watkins et al. (1978), who used measurements from the island and from piston-core samples of the Minoan ash from the eastern Mediterranean. Pyle (1990) has calculated a volume for the Minoan Tuff of ~ 31 km³. If one uses the palaeogeologic reconstruction of Minoan Thera by Pichler and Friedrich (1980), the caldera volume is nearly 60 km³, off by a factor of two or three if the Minoan eruption were responsible for the entire caldera at Thera. This discrepancy in volume between caldera and tephra can be explained if a caldera or calderas had been present in the volcanic field before the Minoan eruption.

CONSTRUCTION OF PALAEOTOPOGRAPHIC AND PALAEOGEOLOGIC MAPS OF MINOAN THERA

As part of their study of caldera development during the Minoan eruption, Heiken and McCoy (1984) prepared palaeotopographic and palaeogeologic maps of Minoan Thera. The maps presented here (Fig. 1) are unchanged from that publication; our continuing but unpublished work on tuff deposits of the Middle Tuff Sequence has not substantially changed our interpretation. This paper includes three-dimensional views of the palaeotopographic reconstruction, prepared using image analysis techniques being developed at Arizona State University; these images allow us to view the reconstruction from various directions and angles, including how the island would have looked from sea-level.

Where the thickness of the Minoan Tuff was known or could be inferred, it was 'stripped' and the elevation and lithology of the underlying rock units were noted. Units mapped for the palaeogeological reconstruction included the pre-volcanic metamorphic rocks, older volcanic rocks of the Akrotiri peninsula, the distinctive Lower Pumice (unter Bimstein), pyroclastic rocks of the Middle Tuff Sequence, and rocks of the lava shields, tuff rings and scoria cones of the northern part of the volcanic field. A key marker bed was the Lower Pumice, dated at 100,000 years by Seward et al. (1980); this massive tuff sequence is similar in composition, thickness and eruption types to the Minoan Tuff.

MIDDLE TUFF SEQUENCE

Overlying the Lower Pumice is a complex sequence of volcanic rocks, which were erupted between 100,000 years BP and the time of the Minoan eruption (~ 1600 BC). Exposed across the southern half of present-day Thera are overlapping deposits of tuff rings, scoria cones and small-volume ignimbrites. These are interbedded with the lava shields that make up the northern half of Thera. Most of these pyroclastic deposits appear to have been erupted from NE-SW-trending vents.

A key to the reconstruction of the southern half of the volcanic field lay in the distribution and origin of the Middle Tuff Sequence. Most of these tuffs are of phreatomagmatic origin and are present as onlapping wedges that are plastered onto steeply sloping walls of a depression not truncated by collapse of the Minoan caldera (Fig. 1). The distribution of these deposits, which slope into the southern half of the current caldera, implies that a depression 6 km in diameter was located here before the Minoan eruption. Eruptions of phreatomagmatic tephra of the Middle Tuff Sequence were from vents located within this depression, which was most likely flooded by the sea.

The presence of abundant phreatomagmatic tuffs implies that this depression was flooded during the 100,000 years before the Minoan eruption. The size of the depression appears to have been about one-half of the present caldera. This depression may have been, in part, a NE-trending Graben, that cut the pre-volcanic metamorphic rocks. Druitt et al. (1989) believe that this depression is mostly tectonic and does not have a volcanic origin.

We propose that the polygonal, flooded depression, 6 km in diameter, shown in Fig. 1 is a pre-Minoan caldera; its shape is controlled, in part, by pre-caldera normal faults; it is located south of the present-day Kameni islands, where water depth is 280 m. What we believe to be the Minoan caldera is located north of the Kameni islands where the average water depth is 380 m; the bathymetry alone can be used to identify two different depressions (Heiken and McCoy 1984).

The older, pre-Minoan caldera could have been formed during eruption of the Lower Pumice series, which is similar to the Minoan Tuff in relative magnitude and eruption history. Formation of an older caldera during eruption of the Lower Pumice series was first proposed by Van Padang (1936); he also suggested that pyroclastic rocks of the Middle Tuff Sequence partly filled this caldera.

     Pre-Minoan lava shields:      The highest of the lava shields, Megalo Vouno volcano, with an elevation of 329 m above sea-level, is composed of interbedded lava flows, scoria cones and one tuff ring. The easternmost lava shield is the Skaros volcano, upon which the village of Merovigli is located. No feeder dikes for these lavas are exposed; lavas of this shield are flat lying at the caldera wall and dip steeply to the east on the well-preserved eastern slope, implying that the summit of this volcano was near the present caldera wall; this reconstruction is consistent with the location and trend of exposed dikes that fed Akrotiri and Megalo Vouno volcanoes (Fig. 2; Heiken and McCoy 1984). The reconstructed palaeosummit of the Skaros volcano has an elevation of 270 m. The island of Therasia, 6 km long and 2.5 km wide, is the largest remnant of the Therasia volcano. Some flows from this volcano are visible below the town of Oia, on Thera, where they interfinger with lavas from Megalo Vouno volcano. Much of this volcano was cut by caldera collapse and now forms the western wall of the Minoan caldera. The highest remnants of Therasia volcano are at an elevation of 260 m above sea-level. Above the lowest lavas of Therasia volcano is a distinctive red scoria unit that appears to have erupted from NE-SW-trending fissures (Fig. 2); it crops out below the village of Oia, where it is 80 m thick, and at Cape Simandiri on Therasia.

The pre-Minoan Tuff lava shields and composite cones were developed within a NE-trending Graben. The orientations of palaeovalleys, the flow directions, and the thickness of scoria deposits support the hypothesis that lava flows and scoria were erupted from NE-trending fissures (now mostly hidden by caldera collapse). Megalo Vouno, whose southern slopes are partly preserved, appears to have comprised a line of cinder cones, associated lava flows and one tuff ring (Cape Kolombos). ) The vents for Therasia lavas appear to have been coincident with fissure vents responsible for a narrow, NE-trending massive scoria deposit. Profiles of the reconstructed lava shields were prepared on the basis of field observations and by analogy with the post-caldera shield of the Kameni islands, which are now similar in form to the three pre-Minoan shield volcanoes. The Kameni shield was formed by eruptions from SW-NE-trending fissures during 11 observed periods of activity over the last 2000 years (Georgalas 1962). The Kameni shield is 4 km wide at the base and about 400 m high (above the caldera floor) and has a profile very similar to that of the remnants of the Skaros volcano.

Our reconstruction of the topography and geology of pre-Minoan Thera is very different from that of Pichler and Kussmaul (1980). Their reconstruction consists of a round island made up of a pair of 600-m-high composite cones. We feel that their reconstruction is not consistent with the observed geology and stratigraphy (Fouqué 1879; Van Padang 1936; Heiken and McCoy 1984).

SUMMARY

Pre-Minoan Thera was a complex volcanic field, consisting of a flooded depression 6 km in diameter in the south, immediately north of the older volcanic rocks of the Akrotiri peninsula (Fig. 1, 3). The northern half of the field consisted of overlapping shield volcanoes and composite cones. Tuffs that make up deposits surrounding the flooded depression are similar in composition to lavas of the north. All appear to have been erupted from NE-SW-trending fissures. Vents in the northern part of the field were subaerial, erupting scoria and lava flows, whereas those in the south, within the flooded depression, erupted mostly phreatomagmatic tephra. Small-volume ignimbrites appear to have erupted from subaerial vents on the shield volcanoes (Druitt et al. 1989) and may have left small calderas or craters, that are now hidden by Minoan caldera collapse.

COMPARISON OF THE RECONTRUCTION OF MINOAN THERA WITH THE FLOTILLA FRESCO

When viewed from the north-west, at sea-level, our reconstruction of the southern half of Minoan Thera is remarkably close to the island pictured on the left side of the Flotilla fresco, from Room 5 of the West House, Akrotiri (Fig. 4). The fresco and reconstruction have the following features in common:

1. The nearly circular bay (older caldera).
2. A wedge-shaped peninsula with low hills on the right, connected to the rest of the island by a narrow, low ridge, is similar to the Akrotiri peninsula.
3. The two high peaks behind the bay fit the high pre-volcanic ridges of Mount Profitis Ilias and Exo Gonia.
4. A low point in the ridge behind the bay and rising slope to the left corresponds to the region between Profitis Ilias and Skaros volcano.
5. The low peninsula in the foreground (with buildings) is similar to the low-lying, reconstructed western peninsula extending from Therasia to the present-day island of Aspronisi.

If the island and bay on the left end of the Flotilla fresco represent Thera before the Minoan eruption, as viewed from the west, then the mountains on the right end of the fresco may represent part of the northern coast of Crete.

Comparisons of the Heiken and McCoy (1984) reconstruction of Minoan Thera with the Flotilla fresco were made independently by Scandone (1987) and McCoy and Heiken (in prep.). If the comparison is correct, then it is possible that ruins of a village may be present below the Minoan Tuff on the island of Aspronisi. The authors have briefly examined the Minoan Tuff and rocks underlying it on the eastern shore of Aspronisi for the purpose of measuring stratigraphic sections. We did not, however, observe any artefacts at the base of the Minoan Tuff. This contact should be examined by archaeologists on all sides of the island.

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

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