Statistical constraints on El Niño Southern Oscillation reconstructions using individual foraminiferal analyses
Recent scientific investigations of sub-millennial paleoceanographic variability have attempted to use the population statistics of single planktic foraminiferal δ18O in an attempt to characterize the variability of high-frequency signals such as the El Niño Southern Oscillation (ENSO). However, this approach is complicated by the relatively short lifespan of individual foraminifera (~2-4 weeks) compared to the time represented by a sediment sample of a marine core (decades to millennia). The resolving ability of individual foraminiferal analyses (IFA) is investigated through simulations on an idealized virtual sediment sample. We focus on ENSO-related sea-surface temperatures (SST) anomalies in the tropical Pacific Ocean (Niño3.4 region). We constrain uncertainties on the range and standard deviation associated with the IFA technique using a bootstrap Monte Carlo approach. Sensitivity to changes in ENSO amplitude and frequency and the influence of the seasonal cycle on IFA are investigated through the construction of synthetic time series containing different characteristics of variability. We find that the standard deviation and range of the population of individual foraminiferal δ18O may be used to detect ENSO amplitude changes at particular thresholds (though the uncertainty in range is much larger than in standard deviation); however, it is highly improbable that IFA can resolve changes in ENSO frequency. We also determine that the main driver of the IFA signal is ENSO amplitude as opposed to changes in the seasonal cycle although this is specific to Niño3.4 where the SST response to ENSO is maximal. Our results suggest that rigorous uncertainty analysis is crucial to the proper interpretation of IFA data and should become a standard in individual foraminiferal studies.