Modelling the provenance of detritus flushed through the Strait of Bosphorus, Turkey, during early Holocene outflow from the Black Sea to the world ocean

Abstract

Abstract Source-to-sink tracing of very fine sand supplied to a controversial lower Holocene deltaic lobe in the northeastern Marmara Sea was undertaken using SEM backscatter mapping and quantitative ‘mineral liberation analysis’ of very fine sand fractions. Forty-two samples from the Kurbagalidere, Goksu, Golden Horn and Riva watersheds, and from Oligo-Miocene outcrops along the southwestern Black Sea coast, show that each of these potential sources has a distinct mineralogical fingerprint. Comparison with 11 samples from the lower Holocene delta and 19 samples from the overlying mud drape show that no single source can account for the mineral proportions in the sand fractions of these depositional sinks, but a mixed provenance provides good matches. For the deltaic lobe, quantitative mixing based on 250,000 Monte Carlo simulations suggests ~ 50% contribution from Oligo-Miocene successions of the southwestern Black Sea coast and inner shelf, ~ 20% contribution from the Goksu stream, and minor contributions from other sources including the Kurbagalidere stream. This balance of source contributions continued to supply the Holocene mud drape until ~ 6 cal ka, after which local sources became more important. It is hypothesized that unconsolidated sediment from a number of watersheds was parked along the coast of the Neoeuxine Lake and in the paleo-Bosphorus valley during the MIS 2 lowstand. A rising Neoeuxine Lake, possibly swollen by glacial outburst floods from the Altay region of Central Asia, flushed ~ 0.5 km 3 of this sediment through the Strait of Bosphorus to the deltaic lobe and permitted it to rapidly advance into the Marmara Sea, which at that time was rising ~ 1 m each 100 yr. Provenance of the lower portion of the Holocene mud drape is consistent with continued Black Sea outflow throughout the deposition of sapropel M1 in the Marmara Sea.