A PRELIMINARY ASSESSMENT OF THE CONSISTENCY OF NEOGENE SEA SURFACE TEMPERATURE ESTIMATES FROM THE UK’37 AND TEX86 PALEOTHERMOMETERS

A variety of different proxies have been employed to reconstruct past sea surface temperature (SST). Several studies applying different paleothermometers to sediments from the same site have yielded dissimilar estimates of past SST, leading to conflicting interpretations of past climate system behavior. These discrepancies have revealed a need to explore the consistency of climate estimates from different SST proxies. The U^k’₃₇ index, is a well-vetted and widely employed paleothermometer. Contrastingly, TEX₈₆ is a relatively new technique for reconstructing past temperature, which is potentially particularly useful in high temperature environments (>29ºC) where the U^k’₃₇ index reaches saturation. Previous studies that have undertaken paired measurements show temperature offsets and/or reveal differing down core patterns for records derived from these two proxies. Here, we explore long-term (>1 Myr), orbital resolution, paired U^k’₃₇ and TEX₈₆ SST reconstructions from the late Miocene and Pliocene at Site ODP 1088 (41ºS) in the South Atlantic Ocean and Site ODP 846 (3ºS) in the Eastern Equatorial Pacific Ocean. At Site ODP 1088, our records which span ~2.4 to ~4 Ma reveal some intervals in which there is remarkable consistency in structure, absolute SST, and variability between estimates based on U^k’₃₇ and TEX₈₆ methods and others in which temperature estimates are offset by several degrees and variability of the TEX₈₆ estimate is twice that of the U^k’₃₇ estimate. At Site ODP 846, companion U^k’₃₇ and TEX₈₆ records for the interval from ~5.1 to ~6 Ma show similar structure, but TEX₈₆ temperatures are consistently offset toward colder values by several degrees and variability of the TEX₈₆ record is almost twice that observed in the U^k’₃₇ record. These preliminary comparisons suggest additional research aimed at exploring factors such as diagenesis, ecology, seasonality, and calibration which might give rise to the differences observed between the two proxies should be undertaken. Until we better understand the factors that account for differences between U^k’₃₇ and TEX₈₆ estimates, caution should be used in treating records based on these proxies as interchangeable.