A probabilistic approach to calculate lags between benthic d18O signals across Termination 1
Changes in the δ18O of benthic foraminiferal calcite (δ18Ob) from ocean sediment cores have been assumed to be a globally synchronous signal and are often used as a stratigraphic marker. However, previous studies have identified temporal offsets (“lags”) between δ18Ob signals that both introduce age errors during stratigraphic alignment and offer an opportunity to better understand the ocean circulation changes that occurred during Termination 1 (T1). We present a novel method to quantify benthic δ18Ob lags by subtracting a core’s radiocarbon age model from its δ18Ob age model. Bayesian software probabilistically constructs radiocarbon age models, δ18Ob aligned age models, and a stack that serves as the target for δ18Ob alignment. Age models and lags are reported with median values and 95% confidence bands. To evaluate the effectiveness of this technique, we calculate lags for a depth transect of 12 cores in the Brazil Margin and compare our results with the temporal offsets found by Lund et al. (2015). The technique yields a statistically significant lag between lower intermediate (1802 – 2296 m) and deep (2500 – 2296 m) cores from 20 – 10 ka BP with median values ranging from a maximum lag of 3.44 kyr (95% confidence interval (CI): 2.43-4.24 kyr) at 20 ka BP to a minimum lag of 1.51 kyr (95% CI: 0.66-2.42 kyr) at 12 ka BP. These results agree with the 2-3 kyr offset described by Lund et al. (2015). Additionally, our method identifies a previously undescribed lag between upper intermediate (1105-1627 m) and lower intermediate (1802-2296 m) cores that is statistically significant from 20 – 15 ka BP with a maximum value of 2 kyr (95% CI: 0.97 – 2.98 kyr) at 20 ka BP.