The modern Aegean Sea is an important source of deep water for the eastern Mediterranean. Its contribution to deep water ventilation is known to fluctuate in response to climatic variation on a decadal timescale. This study uses marine micropalaeontological and stable isotope data to investigate longer-term variability during the late glacial and Holocene, in particular that associated with the deposition of the early Holocene dysoxic/anoxic sapropel S1. Concentrating on the onset of sapropel-forming conditions, we identify the start of 'seasonal' stratification and highlight a lag in d18O response of the planktonic foraminifer N. pachyderma to termination T1b as identified in the d18O record of G. ruber. By use of a simple model we determine that this offset cannot be a function of bioturbation effects. The lag is of the order of 1 kyr and suggests that isolation of intermediate/deep water preceded the start of sapropel formation by up to 1.5 kyr. Using this discovery, we propose an explanation for the major unresolved problem in sapropel studies, namely, the source of nutrient supply required for export productivity to reach levels needed for sustained sapropel deposition. We suggest that nutrients had been accumulating in a stagnant basin for 1-1.5 kyr and that these accumulated resources were utilized during the deposition of S1. In addition, we provide a first quantitative estimate of the diffusive (1/e) mixing timescale for the eastern Mediterranean in its "stratified" sapropel mode, which is of the order of 450 years.

Dates are show as radiocarbon convention ages (conventional age), calibrated radiocarbon years B.C. (calibrated years B.C.) and as calibrated radiocarbon years B.P. (calibrated years B.P.). Analytical errors are given as years (± errors), and calibration fitting errors for a 1 sigma spread are shown in years (±1 sigma). Dating on LC-21 is after Mercone et al. (2000). Samples with codes starting CAM were prepared as graphite targets at the NERC radiocarbon laboratory and analyzed at the Lawrence Livermore National Laboratory AMS facility. Sample codes AA were prepared at Scottish Universities Reactor Research Centre at East Kilbride and analyzed at the Arizona Radiocarbon Facility. KIA sample codes indicate the Leibniz AMS Laboratory at Kiel. Radiocarbon dating was calibrated using CALIB 4.2 after Stuiver and Reimer (1993) and using the marine data set (Stuiver et al., 1998). A reservoir age correction (Delta R) of 149 ± 30 years was used (Facorellis et al., 1998).

Supplement to: Casford, James SL; Rohling, Eelco J; Abu-Zied, Ramadan; Cooke, Steve; Fontanier, Christophe; Leng, M; Lykousis, Vasilios (2002): Circulation changes and nutrient concentrations in the late Quaternary Aegean Sea: A nonsteady state concept for sapropel formation. Paleoceanography, 17(2), 14-1-14-11