EVALUATING LONG CHAIN ALKENONES AND VISIBLE REFLECTANCE SPECTROSCOPY FOR PALEOTEMPERATURE RECONSTRUCTION FROM HARDING LAKE, ALASKA OVER THE LAST 15,000 YEARS

Long-chain alkenones (LCAs) from lake sediments have proven effective for paleoclimate reconstruction. Published calibration studies and genetic analyses suggest that the temperature sensitivity of the alkenone unsaturation index (U^k₃₇) differs among three phylogenetic groups of LCA-producing haptophyte algae. Furthermore, recent studies have shown that estimates of total sedimentary chlorin based on scanning visible reflectance spectroscopy (specifically, relative absorption band depth 660 and 670 nm; RABD_660:670) can provide a rapid and non-destructive approach to paleotemperature reconstruction and can increase the temporal resolution of down-core U^k₃₇-inferred temperatures. Here, we present U^k₃₇ and RABD_660:670 results from Harding Lake (64.421° N, 146.854° W; 217 m ASL) in the interior of Alaska spanning the last ~ 15,000 years. Harding Lake sediments contain isomers of tri-unsaturated alkenones, which indicates alkenone production from Group 1 (Greenland phylotype) haptophytes. Downcore U^k₃₇ values vary between -0.55 and 0.14 over the duration of the record, which produces an unreasonably large range in temperatures that cannot be reconciled with other independent temperature reconstructions from the Alaskan interior. Comparison of the U^k₃₇ and sedimentary chlorin records from Harding Lake shows contrasting patterns on both short (century to multi-century) and relatively long (millennial) time scales. The reason for this discrepancy is unclear, but one possibility is the presence of multiple LCA-producing species in the lake with distinct U^k₃₇-temperature relationships. We examine whether modeling the proportional contributions from different alkenone producers (“unmixing” the signal) can help reconcile the apparent discrepancies between the sedimentary U^k₃₇ and RABD_660:670 records. Results from this analysis provide insight to the feasibility of deconvolving a mixed LCA-signal to provide quantitative paleotemperature estimates.