An Ancient Caldera Cliff Line at Phira, and its Significance for the Topography and Geology of Pre-Minoan Santorini

This occurrence confirms that the Megalo Vouno, Skaros, and Therasia shield complexes had already partly collapsed well before the onset of the Minoan eruption, as deduced previously from independent evidence. Recent estimates of the volume of the Minoan ejecta are consistent with the former existence of the 18 ka caldera only if (i) significant Minoan collapse occurred in southern Santorini, deepening the already existing depression there, (ii) the 18 ka caldera was partly infilled by intra-caldera lava prior to the onset of the Minoan eruption, or (iii) both. The reconstruction of pre-Minoan Santorini presented is consistent with all available field evidence.

INTRODUCTION

Knowledge of the topography and geology of pre-Minoan Santorini is important in assessing the mechanism of the Bronze Age eruption, and in reconstructing the natural environment and lifestyle of the island's Cycladic inhabitants. Pichler and Friedrich (1980) envisaged a single volcanic edifice dominated by pyroclastic cones up to 500-600 m high. In contrast, Heiken and McCoy (1984) reconstructed pre-Minoan Santorini as a cluster of lava shields (Megalo Vouno, Skaros, and Therasia shields), bordered to the south by a 6-km-diameter depression. They found no evidence for the large pyroclastic cones which dominated Pichler and Friedrich's model. Recently, it has been proposed that the northern shields had already collapsed prior to the Minoan eruption, and that a caldera dominated the topography of northern pre-Minoan Santorini (Druitt 1985; Friedrich et al. 1988).

In the present paper we describe an occurrence of Minoan pumice where it lies plastered to the present-day caldera cliff below Phira. The occurrence demonstrates that part of this cliff is an ancient caldera wall which pre-dates Minoan eruption and collapse, and confirms the arguments of Druitt (1985) and Friedrich et al. (1988) for a pre-Minoan caldera in northern Santorini. We discuss the different lines of evidence for the topography and geology of pre-Minoan Santorini, and suggest a reconstruction.

VOLCANIC DEVELOPMENT OF SANTORINI

 The volcanic evolution of Santorini was described in detail by Druitt et al.(1989), and is summarized in Table 1. Santorini is a complex, multi-centre volcanic field constructed adjacent to a NE-SW-trending ridge of late Mesozoic to early Tertiary basement. The north-west margin of the ridge is marked by a steep cliff of schists, plastered by pyroclastic ejecta, which forms the present-day caldera wall at Athinios (Fig. 1). The oldest exposed volcanic rocks (Akrotiri Volcanics) have been dated 630-1590 ka (Ferrara et al. 1980), and occur on the Akrotiri peninsula (Fig. 1). The bulk of the volcanic field has been built up over the last ~ 200 ka, during which time twelve major ( ~ 1-10 km³) explosive eruptions occurred. During intervening periods, intermittent shield- and cone-building activity resulted in effusion of basaltic, andesitic, and dacitic lavas, as well as numerous minor explosions. Truncated remnants of the Megalo Vouno and Skaros shields, the Mikros Profitis Ilias cone and the Skaros-Therasia dome complex (Therasia shield of Heiken and McCoy 1984) comprise much of Therasia and northern Thera. A major conclusion of Druitt et al. (1989) was that the northern half of Santorini has been the main focus of magmatism throughout the last ~ 200 ka (Fig. 1). In contrast, southern Thera consists of a succession of pyroclastic deposits, up to 200 m thick, draped over the basement ridge and Akrotiri volcanoes.

Development of the volcanic field has been influenced strongly by the prominent NE-SW lineament which passes through the Kameni islands (Fig. 1). This lineament (Kameni line) splits the present caldera into two basins, and is believed to be the surface expression of a basement fracture zone. The fractures have acted as conduits for magma throughout Santorini's history. The summit of the Skaros shield and vents of at least four plinian eruptions lay on or near the Kameni line. Calderas elsewhere in the world are known to have developed astride basement fractures or faults (Walker 1984).

THE PUMICE DEPOSIT AT CAPE APANOFIRA

Francaviglia (1989) reported the discovery of a pumice-fall deposit which lies adhered to the present-day caldera cliffs of Cape Apanofira, below the town of Phira (Fig. 1, 2). The deposit is about 1 m thick, has an eroded top, and is typically plinian (Walker 1981) in character (Fig. 3). It occurs about 140 m above sea-level, where a prominent bench occurs in the cliff profile (Fig. 2). The deposit lies unconformably upon lithic breccias of the Middle Pumice pyroclastic member (Druitt et al. 1989) and upon scoria falls which are slightly younger than Middle Pumice. The presence of an underlying palaeosol and the lack of lenticular, inversely graded beds typical of avalanching (Duffield and Bacon 1979) suggest that the pumice is preserved in situ, and has not been reworked. Francaviglia (1989) tentatively concluded that the pumice was of post-caldera origin, although he did not rule out that it was Minoan. The following evidence now leads us to favour a Minoan origin: (i) maximum pumice and lithic diameters resemble those of the Minoan fall deposit near Phira (Bond and Sparks 1976); (ii) lithic clast types are similar to those in the Minoan pumice-fall deposit; (iii) like the Minoan pumice fall (Druitt et al. 1989), the caldera-wall deposit contains three juvenile clast types: a dominant rhyodacitic pumice which is chemically indistinguishable from Minoan rhyodacitic pumice (Table 2) (and unlike any other Santorini pumice sampled by Druitt et al.), a few per cent of a white to pink, microphenocryst-rich pumice type and <<1% of a poorly vesicular, cauliform-shaped grey scoria. The low abundance of grey scoria is typical of the base of the Minoan pumice fall near Phira (as opposed to the top, where it reaches several per cent), and may explain the good compositional sorting of the caldera-wall deposit noted by Francaviglia (1989). Examination of soil profiles on the caldera rim and on Palaea Kameni have failed to reveal any lithologically or chemically similar post-Minoan deposit.

This occurrence of Minoan pumice shows that the upper part of the caldera wall at Cape Apanofira was already exposed prior to the onset of the Minoan eruption. The cliff of this headland has a rounded and weathered surface which extends to within 100 m above sea-level (Fig. 2) and contrasts markedly with the steep, morphologically fresh pyroclastic cliffs of southern Thera. The best interpretation of this cliff surface is that it is part of a pre-Minoan caldera wall. The possibility that it is the wall of an ancient vent is not favoured because of the lack, below the pumice or interstratified with the underlying palaeosol, of poorly sorted tuffs typical of intravent ejecta (Fisher and Schmincke 1984).

If a caldera origin is accepted, stratigraphic relationships (Table 1) enable us to date the collapse event. The ancient cliff surface cuts lavas of the Skaros shield and the Skaros-Therasia dome complex (unit T6 of Pichler and Kussmaul 1980), which form the present caldera rim at Phira. It follows that the collapse must post-date the effusion of these rocks, and is best attributed to the 18 ka Cape Riva eruption (Table 1). This eruption discharged at least a few km³ of magma, and generated lithic-rich, pyroclastic-flow deposits similar to those from caldera-forming eruptions elsewhere (Druitt 1985).

IMPLICATIONS FOR PRE-MINOAN TOPOGRAPHY AND GEOLOGY

The existence of an 18 ka caldera has important implications for Santorini palaeotopography because it requires that the northern shield complex reconstructed by Heiken and McCoy (1984) had already collapsed before the onset of the Minoan eruption. The occurrence of fragments of marine stromatolites in the Minoan ejecta support the former existence of an 18 ka caldera (Friedrich et al. 1988).

     Size of the 18 ka caldera:      The area or depth of the 18 ka caldera cannot be estimated accurately. The upper third of the present caldera cliff at Megalo Vouno is probably pre-Minoan in origin (Heiken and McCoy 1984), but cannot be linked unambiguously to the Cape Riva eruption. Probably the best indication of collapse area is the extent of the Skaros-Therasia dome complex (Fig. 1). These dacites represent precursory leaks from the 18 ka magma chamber, a process documented at other calderas (Bacon 1985), and their wide distribution suggests that the 18 ka caldera may have covered a significant fraction of northern Santorini. The weathered cliff surface at Cape Apanofira suggests that the floor of the caldera must have lain within 100 m of present-day sea-level in the area west of Phira.

     Volume of the Minoan ejecta:      Another important constraint on the pre-Minoan topography of Santorini is the volume of magma discharged during the Minoan event. For collapse calderas like Santorini, the volume of magma erupted as airfall and outflow must approximately equal the observed collapse volume. Probably the most realistic estimate of Minoan ejecta volume is that of Pyle (1990), although his value of ~ 30 km³ (magma-equivalent) may, owing to the way it was calculated, include some intra-caldera ignimbrite. The error on this estimate is probably of the order of plus or minus a few km³. Heiken and McCoy (1984) calculated a collapse volume of about 19 km³ based on their reconstruction of the northern shields (and assuming no collapse south of the Kameni line). Our own calculation of the same volume is about 25 km³. The following question must therefore be addressed: How, if collapse of the entire pre-18 ka northern shield complex can hypothetically yield a volume of ~ 25 km³, could discharge of about 30 km³ of Minoan magma have resulted from collapse of a shield complex already dissected by a deep, and possibly extensive, caldera?

There are two possible solutions to this volume problem. First, significant Minoan collapse may have taken place in southern Santorini. Heiken and McCoy (1984) recognized that a long-lived depression had existed in southern Santorini before the Minoan eruption. Evidence for this depression is provided by the gentle inward dips of pyroclastic strata in the cliffs of south and south-east Thera. This is particularly striking at Athinios, where pyroclastic deposits plaster the schists which form the present caldera wall there. Heiken and McCoy proposed that the depression was an ancient caldera. Druitt et al. (1989) confirmed the former existence of the depression, but argued that it may in part have been a constructional feature defined by the Akrotiri volcanoes and the basement ridge to the south and south-east respectively. Heiken and McCoy (1984) assumed that the entire southern basin of the present caldera pre-dated the Minoan eruption, but field evidence does not preclude significant Minoan collapse in this area. The pyroclastic cliffs of south and south-east Thera are generally steep, morphologically fresh, and probably formed by subsidence and rotational slip during the Minoan eruption. Since the present depth of the southern basin is about 280 m below sea-level and the subaerial cliffs are in places 140 m high, Minoan collapse of at least 300 m (amounting to a volume of about 7 km³) in southern Santorini would not conflict with field evidence. The marked depth difference between the northern (~ 380 m) and southern (~ 280 m) basins of the present caldera may reflect decoupling of basement fault blocks along the Kameni line during Minoan collapse.

The second possible solution to the volume problem is that lava effusion during the 15,000-year period between the Cape Riva and Minoan eruptions may have partially infilled the 18 ka caldera. The only evidence for post-18 ka (but pre-Minoan) volcanism is provided by minor dacitic and olivine-bearing andesitic fall deposits which lie between the Cape Riva and Minoan members at Cape Tourlos (Mellors 1988, 25). However, it is relevant that during the 3,400 years since the Minoan eruption, about 2.5 km³ of dacitic magma has been erupted to form the dominantly submarine Kameni islands shield, yet no post-Minoan tephra layers are preserved outside the present caldera. We infer that eruption of several km³ of post-18 ka intra-caldera lavas does not conflict with the paucity of tephra of that age. Blocks of fresh, glassy dacite (including hyaloclastite) which occur abundantly in the third-phase deposits of the Minoan ejecta may be fragments of post-18 ka intra-caldera lava.

     Reconstruction of pre-Minoan Santorini:      Fig. 4 shows our best guess at a reconstruction of pre-Minoan Santorini, based on all available evidence. Immediately prior to the Minoan eruption, Santorini consisted of a northern shield complex dissected by an 18-ka-old caldera, probably partially infilled by intra-caldera lavas, bordered to the south by a long-lived, pyroclastic-filled depression. The volumes, diameters and shapes of the 18 ka caldera and intra-caldera shield(s) are conjectural, since nearly all evidence for these features was destroyed by Minoan collapse. Our schematic reconstruction of these features is based on the location of the 18 ka caldera cliff at Phira, the large area covered by the pre-18 ka Skaros-Therasia dacites, and the former existence of a low-lying depression south of the Kameni line. Reconstruction of post-18 ka shields astride the Kameni line seems reasonable in view of the prominent role as a conduit that this feature has played throughout the evolution of the volcanic field (Druitt et al. 1989). Our depiction of the 18 ka caldera as flooded is based on the discovery by Friedrich et al. (1988) of marine stromatolite fragments in the Minoan ejecta, and the reconstruction shown in Fig. 4 is essentially similar to theirs. It is also consistent with the Minoan eruption having begun with a plinian phase from a subaerial vent about 1 km west of Phira, and later generated phreatomagmatic explosions when fissures intersected the sea (Bond and Sparks 1976; Heiken and McCoy 1984). The point of entry of the sea into the 18 ka caldera is entirely speculative and based on the observation that the Minoan ejecta descends to near present sea-level in the south-west corner of Thera. However, the possibility that the west or north-west channels were already breached prior to the Minoan eruption cannot be disproved. Given the significant uncertainties involved in reconstruction, we consider the volume difference between the pre-Minoan model and post-Minoan (pre-Kameni) Santorini to be compatible with the observed ejecta volume.

CONCLUSIONS

The cliff at Cape Apanofira is in part an eroded and weathered surface of an ancient caldera wall, formed 18 ka ago during the Cape Riva eruption. This discovery confirms earlier speculation that the shield volcanoes of northern Santorini had already collapsed prior to the Minoan eruption. Judging from the distribution of pre-18 ka rhyodacitic lavas, the 18 ka caldera may have extended across much of the northern basin of the present-day caldera. The reconstruction of pre-Minoan Santorini broadly resembles that of Friedrich et al. (1988), but is different from those of Pichler and Friedrich (1980) and Heiken and McCoy (1984). Like Krakatau in 1883 (see Williams 1941; Self and Rampino 1981) the Minoan eruption deepened an already extant caldera complex, and may have done relatively little to change the fundamental subaerial physiography of the volcanic field. The crude physiographic similarity between pre-Minoan Santorini and Plato's metropolis of Atlantis will undoubtedly fuel speculation at the 1989 Thera Congress.

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

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

(For figures and tables please refer to book)

Source: "Thera and the Aegean World III, Vol. 2" (pp. 362 - 369)

Authors: T.H. Druitt (Department of Geology, University of Wales, United Kingdom), and V. Francaviglia (C.N.R. -Istituto per Le Tecnologie Applicate ai Beni Culturali, Italy)