Sedimentary cycles and volcanic ash beds in the Lower Pliocene lacustrine succession of Ptolemais (NW Greece): discrepancy between 40Ar/ 39Ar and astronomical ages

Abstract

A high-resolution cyclostratigraphy for the rhythmically bedded lignite–marl sequences of the Lower Pliocene Ptolemais Formation is combined with 40Ar/ 39Ar dating results of intercalated volcanic ash beds. Detailed field reconnaissance in three open-pit lignite mines reveals three end-member sediment types: lignites, composed primarily of organic material; grey marls, a mixture of carbonate and organic material; and beige marls, almost exclusively composed of carbonate. These lithologies are arranged in two basic types of sedimentary cycles: lignite–grey marl and lignite–beige marl cycles. A cyclostratigraphic composite section comprising 56 lignite–marl cycles is constructed which combines the consistent cycle patterns from three parallel sections. The concordant positions of 20 volcanic ash beds in these sections confirm the cyclostratigraphic correlations and indicate that the lignite–marl cycles result from regional, basin-wide forcing rather than lateral facies migrations. 40Ar/ 39Ar ages on sanidine and biotite separates from nine volcanic ash beds were obtained by multiple total fusion and incremental-heating experiments. The 40Ar/ 39Ar ages range between 5.00±0.05 and 4.04±0.04 Ma and are, in general, consistent with the stratigraphic order. A least-square linear regression using the measured 40Ar/ 39Ar ages gives an average duration of 21.8±0.8 kyr per lignite–marl cycle. Evidently, the lignite–marl cycles in the Ptolemais Formation are linked to the precessional variation in the Earth's orbit through its influence on Mediterranean climate. For the first time, 40Ar/ 39Ar dating results, totally independent from any other dating and or tuning technique, confirm the astronomical theory of climate change. The 40Ar/ 39Ar ages of the volcanic ash beds show a constant ∼200 kyr (∼4.5%) age discrepancy with the astronomical ages of the same ash beds. This inconsistency remains difficult to explain. The discrepancy is unlikely to have resulted from erroneous astronomical ages, through incorrectness in the astronomical tuning, inaccuracies of the magnetostratigraphic data or the orbital time-series used, and/or errors in the APTS. The 40Ar/ 39Ar dating results neither give clear indications for a possible source of error. From the excellent data set it is evident that neither loss of radiogenic 40Ar, nor an underestimation of the contribution of Ca- and K-derived Ar isotopes could have caused the discrepancy. Moreover, the discrepancy is also beyond the errors in the systematic variables, like the decay constants of 40K or the ages for the neutron-fluence monitors.