The Influence of Sampling Density on Bayesian Age-Depth Models and Paleoclimatic Reconstructions - Lessons Learned from Lake Titicaca - Bolivia/Peru

Abstract:

Lake Titicaca is one of the most important archives of paleoclimate in tropical South America, and prior studies have elucidated patterns of climate variation at varied temporal scales over the past 0.5 Ma. Yet, slow sediment accumulation rates in the main deeper basin of the lake have precluded analysis of the lake's most recent history at high resolution. To obtain a paleoclimate record of the last few millennia at multi-decadal resolution, we obtained five short cores, ranging from 139 to 181 cm in length, from the shallower Wiñaymarka sub-basin of of Lake Titicaca, where sedimentation rates are higher than in the lake's main basin. Selected cores have been analyzed for their geochemical signature by scanning XRF, diatom stratigraphy, sedimentology, and for ¹⁴C age dating. A total of 72 samples were ¹⁴C-dated using a Gas Ion Source automated high-throughput method for carbonate samples (mainly Littoridina sp. and Taphius montanus gastropod shells) at NOSAMS (Woods Hole Oceanographic Institute) with an analytical precision higher than 2%. The method has lower analytical precision compared with traditional AMS radiocarbon dating, but the lower cost enables analysis of a larger number of samples, and the error associated with the lower precision is relatively small for younger samples (< ~8,000 years). A 172-cm-long core was divided into centimeter long sections, and 47 ¹⁴C dates were obtained from 1-cm intervals, averaging one date every 3-4 cm. The other cores were radiocarbon dated with a sparser sampling density that focused on visual unconformities and shell beds. The high-resolution radiocarbon analysis reveals complex sedimentation patterns in visually continuous sections, with abundant indicators of bioturbated or reworked sediments and periods of very rapid sediment accumulation. These features are not evident in the sparser sampling strategy but have significant implications for reconstructing past lake level and paleoclimatic history.