Late Quaternary lake-level changes constrained by radiocarbon and stable isotope studies on sediment cores from Lake Titicaca, South America

We present and compare AMS-C geochronologies for sediment cores recovered from Lake Titicaca, South America. Radiocarbon dates from three core sites constrain the timing of late Quaternary paleoenvironmental changes in the Central Andes and highlight the site-specific factors that limit the radiocarbon geochronometer. With the exception of mid-Holocene sediments, all cores are generally devoid of macrophyte fragments, thus bulk organic fractions are used to build core chronologies. Comparisons of radiocarbon results for chemically defined fractions (bulk decalcified, humate, humin) suggest that ages derived from all fractions are generally coherent in the post-13,500 yr BP time interval. In the pre-13,500 yr BP time interval, ages derived from humate extracts are significantly younger (300–7000 years) than ages from paired humin residues. Gross age incoherencies between paired humate and humin sub-fractions in pre-13,500 yr BP sediments from all core sites probably reflect the net downward migration of humates. Ages derived from bulk decalcified fractions at our shallow water (90 m) and deep water (230 m) core sites consistently fall between ages derived from humate and humin subfractions in the pre-13,500 yr BP interval, reflecting that the bulk decalcified fraction is predominantly a mixture of humate and humin sub-fractions. Bulk decalcified ages from the pre-13,500 yr BP interval at our intermediate depth core site (150 m) are consistently older than humate (youngest) and humin sub-fractions. This uniform, reproducible pattern can 0921-8181/$ see front matter D 2003 Published by Elsevier B.V. doi:10.1016/S0921-8181(03)00031-6 * Corresponding author. Department of Geological Sciences, University of Kentucky, 101 Slone Research Bldg., Lexington, KY 405060053, USA. Tel.: +1-859-257-2770; fax: +859-257-2770. E-mail addresses: [email protected] (H.D. Rowe), [email protected] (T.P. Guilderson), [email protected] (R.B. Dunbar), [email protected] (J.R. Southon), [email protected] (G.O. Seltzer), [email protected] (D.A. Mucciarone), [email protected] (S.C. Fritz), [email protected] (P.A. Baker). www.elsevier.com/locate/gloplacha Global and Planetary Change 38 (2003) 273–290 be explained by the mobilization of a relatively older organic sub-fraction during and after the re-acidification step following the alkaline treatment of the bulk sediment. The inferred existence of this ‘alkali-mobile, acid-soluble’ subfraction implies a different depositional/post-depositional history that is potentially associated with a difference in source material. While internally consistent geochronologies can be developed for the Lake Titicaca sequence using different organic fractions, mobile organic sub-fractions and fractions containing mobile sub-fractions should generally be avoided in geochronology studies. Consequently, we believe humin and/or bulk decalcified ages provide the most consistent chronologies for the post-13,500 yr BP interval, and humin ages provide the most representative ages for sedimentation prior to 13,500 yr BP interval. Using the age model derived from the deep water core site and a previously published isotope-based lake-level reconstruction, we present a qualitative record of lake level in the context of several ice-core records from the western hemisphere. We find the latest Pleistocene lake-level response to changing insolation began during or just prior to the Bølling/Allerød period. Using the isotope-based lake-level reconstruction, we also find the 85-m drop in lake level that occurred during the mid-Holocene was synchronous with an increase in the variability of ice-core dO from a nearby icecap, but was not reflected in any of the polar ice-core records recovered from the interior of Antarctica and Greenland. D 2003 Published by Elsevier B.V.