Fission Track Ages of Santorini Volcanics (Greece)

The age of the uppermost part of the volcanic sequence, which forms the caldera walls of the ring-islands Thera and Therasia, belonging to the Santorini Group (Fig. 1) in the southern Aegean, has been determined previously with the radiocarbon method (Pichler & Friedrich 1976). Below the late-Minoan Upper Pumice Series (Bo; P1 in Fig. 2), three different plant-bearing palaeosols of about 13,000, 18,500 and 37,000 yr. B.P. have been dated and correlated with interstadials in eastern Macedonia, deep-sea cores from the Mediterranean and interstadials from the northern hemisphere (Friedrich et al. 1977). These dates indicate that the age of the Lower Pumice Series (Bu), which was originally put at 17,000 - 18,000 yr. B.P. (Keller & Ninkovich 1972), must be older than 37,000 yr.

In the central part of Thera, the Lower Pumice Series is underlain by lava flows and intercalated pyroclastics of the Thera volcano. To the south, this sequence thins out. In the profile north of Cape Thermia (P2 in Fig. 2), thick, slightly welded to non-welded red scorae and dark welded ignimbrites represent the lowermost part of the Thera volcano sequence. This sequence rests here on a thick palaeosol horizon which overlies white vitric lapilli tuffs. The horizontal extent of this 2 - 3 m thick rhyodacitic stratum is limited to some meters. It rests directly on outcropping Lower Tertiary basement rocks, i.e. phyllites with intercalations of semi-metamorphosed greywackes. Stratigraphically this lenticular tephra layer represents the oldest volcanic events on Santorini.

A second outcrop of this layer was found in the southern part of Thera, directly at the base of the caldera walls north of Akrotiri village. This part of Thera is mainly built up of pyroclastics and lavas of the Akrotiri volcanoes (Pichler & Kussmaul 1972). The andesitic to dacitic products of these volcanoes, which are limited to the southern part of Thera, are older than the Thera volcano sequence. At the Balos cliff, north of Akrotiri village, the Akrotiri volcanics are disconformably underlain by white vitric tuffs of rhyodacitic composition. These tuffs are lithologically identical with those in the profile north of Cape Thermia and identical ages for both outcrops should be expected.

In order to provide absolute age data for the older volcanics of the Santorini group, ranging stratigraphically from the white vitric tuffs up to the Middle Pumice Series (Bm; P2 in Fig. 2), seventeen tephra samples and some obsidian blocks were selected for fission track dating.

METHODS

a)     Zircons

Tephra samples, each about one kilogram in weight, were treated in 20% HF solution to extract zircons. Only two samples yielded enough zircons for the dating procedure. They were mounted in teflon sheet, polished and etched in molten NaOH-KOH (Gleadow et al. 1976). Total etching time was 70 hours at 205^o C. The spontaneous tracks were counted in the zircons themselves on prism faces. The induced tracks were counted in external mica detectors etched for 13 minutes in 48% HF at 22^o C.

The mica counts were doubled on the assumption that the geometry factor in this case is 0.5. The same area was counted on the mica as on the zircon for each grain, at a magnification of X 2500.

b) Obsidians

Two splits of up to one cubic centimetre were taken from three obsidian blocks. One of each pair was irradiated. After irradiation, they were mounted and polished, and each pair was etched together for 2½ minutes in 16% HF at 22^o C. Tracks were counted at a magnification of X 500, and track lengths were measured at X 2500.

Initial results indicated a 12 - 14 % size reduction of the spontaneous tracks. Therefore aliquots of both spontaneous and induced material were partially annealed together at 200^o C in order to determine a plateau age (Storzer & Poupeau 1973). Each pair required different annealing times (Table 1) until the size of the spontaneous and induced tracks was equal and the induced track density was reduced by approximately 50%. It was then assumed that the plateau had been reached.

It should be noted that normally a plateau age is older than the measured age, but in this study, only sample 085 (Table 1) followed the expected pattern. Sample 083, in particular, contained numerous inclusions which etched out like tracks, such that, with reflected light alone, it was impossible to distinguish track from inclusion. This fact was not obvious on the first count; but after the first annealing and the size of the fission tracks was reduced, the size of the false tracks formed from the inclusions remained the same. This phenomenon has previously been recorded in other glasses from the Mediterranean (Wagner et al. 1976). The number of inclusions is small and has no marked effect on the density of the induced tracks, but they have a crucial effect on the density of the spontaneous tracks in such young samples with a low uranium concentration.

Only a small portion of the polished glass surface, sometimes less than 10 mm², was suitable for observing and counting tracks and consequently the obsidian pairs had to be repolished, reetched, and tracks counted and measured several times. Because of the time consuming nature of this problem the final number of spontaneous tracks counted is low and thus the error on the dates large (Table 1). The error for all dates takes into account errors on the spontaneous and induced tracks as well as that on the dose determination.

c) Uranium analysis

Uranium content was determined in Makrofol track recorders placed over pieces of polished obsidian. A standard glass (NBS SRM 614 with 0.823 ppm uranium: Carpenter & Reimer 1974) was treated in the same manner. Samples and standard were irradiated together. The Makrofol was etched in 6N NaOH for 10 minutes at 70° C, and the track densities were counted, and ratios determined.

d) Dose determination and decay constant

Dose was calculated using NBS SRM 962 (37.38 ppm U) and NBS SRM 963 (0.823 ppm U: Carpenter & Reimer 1974) glasses with copper calibration, which for one of the authors (Seward, in prep.) has yielded geologically meaningful ages in conjunction with the decay constant 6.85 X 10^-17 y^-1 (Fleischer & Price 1964).

It was also calculated using the glass dosimeter (moldavite) of the Heidelberg laboratory in conjunction with the decay constant 8.46 X 10^-17 y^-1 (Galliker et al, 1970). This decay constant together with the moldavite also leads to geologically meaningful results (Wagner et al. 1975).

Doses calculated using the NBS glasses with gold calibration were similar to that of the moldavite.

The determination of dose and the choice of decay constant continue to be one of the foremost problems in fission track studies. It is interesting and comforting to note that the ages reported here, using, for the first time, the two different standards and decay constants, yield very similar ages (Tables 1 and 2). It seems that the controversy on the decay constant when determined from independently known ages, is essentially due to neutron dosimetry problems.

In tables 1 and 2 and throughout the text, the first age quoted is that with the decay constant 6.85 X 10^-17 y^-1 and the second with 8.46 x 10-¹⁷ y^-1.

RESULTS AND DISCUSSION

a) Zircon

The two horizons which contained sufficient zircons for dating crop out at Cape Plaka (S8) and at the base of the Balos cliff north of Akrotiri village (S45). These tuffs represent the oldest volcanics on Santorini. At Cape Plaka, these pyroclastics overlie basement phyllites of Lower Tertiary age. At the Balos cliff, the vitric tuffs are well-stratified and are disconformably overlain by the Akrotiri volcanics, which are older than the lavas and pyroclastics of the Thera volcano, including the Lower Pumice Series (Bu). The resulting ages were 1.05 ± 0.16 (0.94 ± 0.14) Myr and 1.02 ± 0.08 (0.94 ± 0.07) Myr (Table 2).

One of these samples (S8, Table 2) contained two populations of zircon crystals. The younger one (1.05 ± 0.16 Myr) tended to be subhedral to euhedral, and generally colourless, while the second was anhedral, generally rounded and slightly pink. This second group gave an age of 13.9 ± 1.2 (12.5 ± 1.1) Myr. The age of the younger population is statistically similar to that of the zircons in the other tephra horizon (S45) and it is assumed that this is the age of the tephra and of the earliest eruptions of the island. The older population of zircons may have been derived from basement material brought up during the eruption. Rb/Sr and K/A ages of basement rocks of the Cyclades of 8 to 15 Myr (Wendt et al. 1977) lend support to this conclusion. It could not be determined whether these zircons had suffered any annealing during the eruption but, as they are contained in air fall tephra it is assumed not.

b) Obsidians

The three analysed obsidian blocks (Table 1) are all from the base of the Lower Pumice Series (Eu; Fig 2). The expected ages for these blocks and therefore the Bu Series should be older than 37,000 yr. (Fig. 2). Samples 083 and 084 contain numerous small xenolithic inclusions. They are both from the base of the Lower Pumice Series near Akrotiri Village (Fig. 1). Sample 085 is from a thin ignimbrite layer in the northern part of Thera island and corresponds stratigraphically to the other two samples. It is also chemically equivalent to the pumice of the Bu series (Pichler & Kussmaul 1972; Kussmaul 1971).

The uranium content of the obsidians 5.64 ppm, 5.05 ppm and 6.08 ppm, samples 083, 084, 085 respectively is very similar and also confirms the correlations.

The plateau ages of the obsidians are 0.138 ± 0.030 (0.126 ± 0.030) Myr, 0.074 ± 0.016 (0.067 ± 0.016) Myr and 0.106 ± 0.021 (0.097 ± 0.020) Myr (Table 1), and are consistant with the stratigraphy. An average age of approximately 100,000 years would represent a more accurate estimate of the age of the Bu horizon.

The eruptions, which produced the Lower Pumice Series, started with the extrusion of overheated rhyodacitic magma. Thus obsidian plugs within one or several vents were formed. Later the obsidians were blown up by vent opening paroxysmal eruptions which preceded the pumice fall and ash flow eruptions of the Bu phase.

Assuming that the temperature of eruptions was of the order of 300 - 400° C and that this obsidian required 24 - 48 hours at 200° C to achieve 50% track loss (Table 1) it is not inconceivable that this vent opening activity would generate sufficient heat to anneal all the tracks in the glass. Any fission tracks that may have accumulated in the obsidian prior to these eruptions would have been completely annealed such that the obsidian then represents the age of the tephras of the Bu Series.

Since the tephra eruption all three obsidians from different sites have undergone an equivalent amount of annealing (12 - 14%) which after correction yields an age of approximately 100,000 years for the Lower Pumice Series. 

CONCLUSIONS

1. The age of the oldest subaerial volcanic activity on Santorini has been determined by the fission track method on zircons. This explosive activity, by which white vitric ashes and pumice lapilli were deposited on the strongly decomposed and reworked surface of Lower Tertiary phylitic rocks, occurred about 1 million years ago.
2. The age of the Lower Pumice Series is approximately 100,000 years. This age is consistent with the stratigraphy on Santorini and with the ¹⁴C ages of the younger strata.
3. Based on our dating of the Lower Pumice Series, which occurs also as a petrologically significant tephra layer in the Quaternary deep-sea cores of the eastern Mediterranean, a better correlation of the different layers in the cores should be possible.

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

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