The Geochemistry of the Radial Dykes of the Santorini Caldera and its Implications

All rocks are light rare earth enriched, the more siliceous lithologies yielding the highest LREE values. Despite the frequent occurrence of plagioclase phenocrysts, only slight negative europium anomalies are apparent in chondrite-normalized patterns. Locally, iron is oxidized to the trivalent state. The Zr/Hf ratio is generally high (about 40) as is Rb/Cs. Rare elements such as Li, B, Be, Ga, Nb and Ta display clear trends with differentiation. Base metal elements (Cu, Zn, Pb) are measured and compared to the metalloids As, Sb, Se and Te. Several of these elements were analysed for the first time for Santorini rocks. In total, more than 50 components were determined for each sample, using XRF, INAA, ICP-MS, infra-red absorption and wet chemistry.

INTRODUCTION

In the north-east part of the Santorini caldera wall Reck (1936) described the occurrence of more than 60 dykes between Oia and Skaros. Most of them seem to be radial structures. The width of the dykes lies between 0.25 and 5.00 m. None of them penetrates to the surface of the volcano. Since they vanish sooner or later in the steep wall, they are only accessible from the sea by boat.

Thicker dykes are sometimes zoned, with a vesicular outer and a massive inner part. The dip of most of the dykes is close to perpendicular. Nineteen of them from the Megalovouno area have been sampled (Fig. 1) and measured (Table 1).

PETROGRAPHIC DESCRIPTION

From the thin sections it was learned that the textural development of the dykes is determined by their cooling history. A few of the samples have a completely glassy matrix. Most of the dyke rocks contain microlites of plagioclase and clinopyroxene in the glass matrix. Due to degassing of the magma most of the rocks are in part or completely vesicular. Plagioclase is the prevailing phenocryst. It shows polysynthetic twinning and is often zoned. In many samples the plagioclase is corroded and contains inclusions of pyroxene. Clinopyroxene is the second most frequent phenocryst. In a few dykes orthopyroxene has been found. Altered olivine occurs occasionally in more basic samples. Ferric minerals are proof of oxidation.

GEOCHEMISTRY

Major element analyses were performed on all of the samples using XRF and classical methods (Table 2). Rare earth and some trace elements were analysed by INAA on all of the samples (Table 3). ICP-MS analyses of a large number of additional trace elements were carried out on seven samples (Table 4).

Altogether 59 elements have been determined. Data from the different methods are in good agreement. In their major element composition the dykes are typical members of the calc-alkaline suite. When plotted on the potassium-silica diagram (Fig. 2) they range from high alumina basalts to dacites, using the classification of Taylor (1969). The broad compositional range of the dyke rocks argues against a derivation of the magmas from one homogeneous chamber. Several intrusive events either from different magma chambers or separated by a longer time span may have caused the variable compositions.

Regarding the trace elements (Fig. 3, 4, 5), Li, Be, Th and U show the normal increase of their concentrations with increasing silica content. Thorium is correlated to uranium with a rather constant ratio Th/U between 2 and 3.5.

With increasing silica content arsenic concentration does not change, barium increases and copper and strontium decrease.

A few samples with low SiO₂ are characterized by high Ni and Cr values. They are higher than most other Santorini rocks (Puchelt 1978) and indicate derivation from a deeper zone of the crust (Nicholls 1978).

The rare earth elements behave as in previous investigations of Santorini rocks (Puchelt 1978). All patterns are LREE enriched (Fig. 6, 7). There is no or little negative Eu anomaly and La/Lu values are between 20 and 40. The higher the SiO₂ of the sample the higher is the total REE content.

Selenium data indicate some losses of the volatile elements to the atmosphere. The only surface sample, 'Skaros dacite', contains selenium below detection limit (= 2 ppm) only. Very low sulphur contents, mostly below 50 ppm, as observed during preparation of the dyke rocks for sulphur isotope analyses (Hubberten 1984), support this assumption (Sakai et al. 1982). The sulphur isotope composition of the dyke rocks demonstrates an early or recent admixture of marine sulphate (Hubberten 1984).

DISCUSSION

Two observations have to be taken into consideration: (1) none of the dykes penetrates to the surface of the volcano, and many of them are only visible just above the present sea-level; (2) the major element geochemistry of the dyke rocks covers almost the full scale of Santorini volcanic rocks. The .pistribution of their chemistry is not a unidirectional development.

The first topic to be discussed is the mode of formation of the fissures where magma can intrude. Fig. 8 gives several possibilities: the open space for the intruding magma can be provided (I) by tensional forces; (II) by convex bending; (III) by concave bending.

While the first possibility is not possible for this restricted part of the Aegean Sea, convex bending requires either an intrusion of magma or compressional forces. In any case, fissures generated by possibility (II) would be wider at the top than at their lower end. At the same time, intrusion of the dyke material would be difficult.

The third model requires a concave bending of the layers in order to open the fissures. For fissures which do not penetrate to the surface only mode (III) seems possible. The fact that the dyke rocks do not show a homogeneous chemistry or at least a distinct development may be explained by the existence of various periods of magma intrusion and effluence from the magma chamber. Clarifying the sequence of dyke intrusion will be difficult. One possibility would be a kind of pulsation of the magma with different new lower portions added.

Subsidence of the surface of the volcano is also assumed by Huijsmans and Barton (1990). If such a process does not happen always in the same place and intensity, different fissures may open at different times.

Studying the profiles of the north-east caldera walls given by Reck (1936), it becomes clear that repeated formation of fissures happened, which formed with different intensity. From his sketches types of dykes can be grouped which penetrate into the same volcanic layer. One group could be nos. 20-24, and another would be nos. 25-33. A few others (nos. 34, 39a and another unnumbered dyke) penetrate almost to the top of Megalovouno. All these dykes must have been emplaced when Megalovouno was still a complete edifice.

Dyke emplacement must have happened prior to the development of the last Santorini caldera (cf. Friedrich et al. 1988).

CONCLUSIONS

Dykes in the north-east Santorini caldera are probably formed in fissures which generated prior to collapse of at least the last caldera due to subsidence of parts of the volcanic surface. The reason was perhaps the formation of an empty space above a magma in the charnber.

The fissures were not filled in one event, but with lava material of different origin or differentiation development.

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

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