Mapping Atlantic Benthic δ18O Lags During Termination 1

Abstract:

Recent studies have identified lags between Atlantic benthic δ¹⁸O records during Termination 1 (T1), which can introduce age errors during stratigraphic alignment. However, benthic δ¹⁸O lags also offer an opportunity to improve ocean circulation reconstructions if lags and their uncertainties are mapped on a basin-wide scale. Here, we present timeseries of δ¹⁸O lags and their uncertainties across T1 for 35 Atlantic cores between latitudes of 38 N and 30 S. A lag is defined by the difference between a core’s radiocarbon age model and an age model based on synchronous δ¹⁸O change (i.e., δ¹⁸O stratigraphic alignment to an Atlantic stack). Age models are constructed using Bayesian software that draws Monte Carlo samples in proportion to their probabilities. Timeseries of lags and their uncertainties are calculated by subtracting radiocarbon age model samples from δ¹⁸O age model samples. Lags are reported relative to the Atlantic stack, which represents the average Atlantic signal. Throughout the termination, statistically significant negative lags (leads) of up to -3 kyr are observed in the tropical West Atlantic at intermediate depths. In contrast, wide-spread positive lags are observed in the West Atlantic below 2500 m during the second half of the termination. Statistically significant lags are less common in the East Atlantic. Potential factors that contribute to lags include: (1) changes in the rate of overturning circulation, (2) asynchronous changes in end-member water mass properties, and (3) water mass boundary shifts. We compare the observed Atlantic lags with idealized experiments using an ocean circulation inverse model to better understand the circulation changes that occurred during T1 (Gebbie et al., AGU Fall Meeting 2020).