Dating Quaternary lacustrine sediments in the McMurdo Dry Valleys, Antarctica

Abstract Reports of erroneously old 14 C dates for modern Antarctic materials have thrown doubt into 14 C chronologies. The carbon reservoir effect purported to exist in Quaternary lacustrine sediments of the McMurdo Dry Valleys was investigated by studying 14 C distribution, and testing alternate dating techniques. Our results show that the carbon reservoir effect is not pervasive. Stream and near-shore microbial mats and dissolved inorganic carbon (DIC) in the surface waters of Lake Fryxell are in equilibrium with modern 14 CO 2 . The surface waters of Lake Hoare and Lake Bonney, however, have DIC 14 C ages of 1650 and 2080 yr B.P., respectively. These older age estimates are suggested to be due to the direct input of large amounts of glacial melt with relict DIC. On the other hand, Lake Fryxell receives only a minor component of its inflow directly from a glacier, while a large component must travel long distances in numerous shallow ephemeral streams after leaving local valley glaciers. This mode of melt-water input allows the water to equilibrate with modern CO 2 before entering the lake. Bottom-water 14 C ages for Lake Hoare closely match surface sediment ages, supporting the widely published period ∼1200 yr B.P. (after a 1650 yr reservoir correction) when most dry valley lakes apparently evaporated to small brine ponds and/or disappeared completely. Lake Bonney bottom-water is ∼8000 yr B.P. Carbon dating is shown to be a viable technique for lake edge deposits, and possibly lake bottom deposits where a correction to the sediment surface age can be obtained. However, we conclude that deep-water paleolake deposits can not be reliably dated using 14 C alone because of an inability to determine the age of the reservoir correction (i.e. accounting for the initial carbon reservoir, plus the age of the bottom water). A suite of alternative and complimentary dating techniques were tested on modern and late Holocene lacustrine deposits. These include thermoluminescence (TL), 210 Pb, 137 Cs, and paleomagnetism. Of these techniques, TL (or the more sensitive optically stimulated luminescence) holds the most promise for correcting lake sediment 14 C ages. TL dating of the modern Lake Hoare sediments showed an ∼1000 yr relict signal. This signal is unaffected by the age of the lake water. Both 210 Pb and 137 Cs occurred in very low levels in the sediments and do not appear to be viable dating techniques for the perennially ice-covered lakes of the McMurdo Dry Valleys. Paleomagnetism was not suited to the coarse-grained nature of Lake Hoare bottom sediments, but could be a useful alternative given a finer sediment type.