Document Type

Contribution to Book

Publication Date


Journal or Book Title

Proxies in Late Cenozoic Paleoceanography


Pt. 2: Biological tracers and biomarkers


C. Hillaire-Marcel and A. de Vernal



Place of Publication



The most popular proxies based on microfossil assemblage data produce a quantitative estimate of a physico-chemical target parameter, usually by applying a transfer function, calibrated on the basis of a large dataset of recent or core-top samples. Examples are planktonic foraminiferal estimates of sea surface temperature (Imbrie & Kipp. 1971) and reconstructions of sea ice coverage based on radiolarian (Lozano & Hays, 1976) or diatom assemblages (Crosta, Pichon & Burckle, 1998). These methods are easy to use, apply empirical relationships that do not require a precise knowledge of the ecology of the organisms, and produce quantitative estimates that can be directly applied to reconstruct paleo-environments, and to test and tune global climate models. Assemblage-based proxy methods that do not yield fully quantitative results have become less popular over the last decennia, mainly because semi-quantitative or qualitative proxy methods implicitly admit a considerable degree of uncertainty. In many allegedly quantitative proxies, however, the error may be as large. but is concealed by the numerical aspect of the estimates. commonly expressed with 1 or 2 numbers behind the decimal point, suggesting a highly precise and trustworthy reconstruction of the target parameter. The decreased popularity of assemblage-based proxy methods affects proxies based on benthic foraminifera, although they have been used to reconstruct a wide range of oceanographic parameters, including water depth, water mass properties, bottom water oxygen content, and the extent and/or seasonality of the organic flux to the ocean floor. Although the use of benthic foraminiferal assemblage compositions has become less common, the use of benthic foraminiferal tests as carriers of geochemical proxy methods (stable isotopes, Mg/Ca, Sr/Ca etc. ) has never been so widespread (e.g., Wefer, Berger, Bijma & Fischer, 1999; Lea, 2004; Lynch-Stieglitz, 2004; Ravizza & Zachos, 2004; Sigman & Haug, 2004).