The isotopic composition and mass balances of sources and sinks of sulfur are used to constrain the limnological-hydrological evolution of the last glacial Lake Lisan (70-14 ka BP) and the Holocene Dead Sea. Lake Lisan deposited large amounts of primary gypsum during discrete episodes of lake level decline. This gypsum, which appears in massive or laminated forms, displays δ34S values in the range of 14-28‰. In addition, Lake Lisan's deposits (the Lisan Formation) contain thinly laminated and disseminated gypsum as well as native sulfur which display significantly lower δ34S values (-26 to 1‰ and -20 to -10‰, respectively). The calculated bulk isotopic compositions of sulfur in the sources and sinks of Lake Lisan lacustrine system are similar (δ34S ≈ 10‰), indicating that freshwater sulfate was the main source of sulfur to the lake. The large range in δ34S found within the Lisan Formation (-26 to +28‰) is the result of bacterial sulfate reduction (BSR) within the anoxic lower water body (the monimolimnion) and bottom sediments of the lake. Precipitation of primary gypsum from the Ca-chloride solution of Lake Lisan is limited by sulfate concentration, which could not exceed ∼3000 mg/l. The Upper Gypsum Unit, deposited before ca. 17-15 ka, is the thickest gypsum unit in the section and displays the highest δ34S values (25-28‰). Yet, our calculations indicate that no more than a third of this Unit could have precipitated directly from the water column. This implies that during the lake level decline that instigated the precipitation of the Upper Gypsum Unit, significant amounts of dissolved sulfate had to reach the lake from external sources. We propose a mechanism that operated during cycles of high-low stands of the lakes that occupied the Dead Sea basin during the late Pleistocene. During high-stand intervals (i.e., Marine Isotopic Stages 2 and 4), lake brine underwent BSR and infiltrated the lake's margins and adjacent strata. As lake level dropped, these brines, carrying 34S-enriched sulfate, were flushed back to the shrinking lake and replenished the water column with sulfate, thereby promoting massive gypsum precipitation. The Holocene Dead Sea precipitated relatively small amounts of primary gypsum, mainly in the form of thin laminae. δ34S values of these laminae and disseminated gypsum are relatively constant (15 ± 0.7‰) and are close to present-day lake composition. This reflects the lower supply of freshwater to the lake and the limited BSR activity during the arid Holocene time and possibly during former arid interglacials in the Levant.
Bibliographical noteFunding Information:
We thank A. Amrani and Z. Aizenshtat for advice and support in laboratory work and E. Shalev for valuable discussions. We also thank Blair Jones and two anonymous reviewers for helpful and constructive remarks. The study was supported by BSF 2000271 (to I.G. and M.S.) and GIF 717.129.8/2001 (to M.S. and Y.K.).