Reconstruction of the last deglaciation: deconvolved records of $\delta^{18}$O profiles, micropaleontological variations and accelerator mass spectrometric $^{14}$C dating
Résumé
Bioturbation acts as a low-pass filter in displacing and reducing the amplitudes of stratigraphic signals. This often leads to a loss of high-frequency events in the
stratigraphic record. In addition, when considering an isotopic signal ($\delta^{18}$O, $^{14}$C) measured in stratigraphic "carriers", such as foraminifera, bioturbation and carrier abundance changes can create artifacts which may be falsely interpreted as leads or lags in the paleoclimati c record. We present here a model in which bioturbation is treated as a time-invariant filter whose impulse response function (IRF) is like that of a first-order system. The method involves first deconvoluting the abundance curves of the carriers and then the isotopic signals using the restored "carrier" abundances. This analysis was initially used toartificially generate ideal curves, with the aim of qualitatively modelling the effects ofbioturbation. Following this, deconvolved curves were obtained using data from the core CH73-139C using $\delta^{18}$O, A.M.S. C-14 ages, and abundances of two planktonic foraminifera: $G.bulloides$ and $N.pachyderma$ left-coiling. A comparison of the data with the unmixed curves enables separation of the bioturbation artifacts and the construction of a common deglaciationcurve based on the restored signals. Importantly, the model emphasizes some severe limitations of mathematical analysis of stratigraphic signals.