Multi-indicator conductivity transfer functions for Quaternary palaeoclimate reconstruction

Abstract

Diatoms (Bacillariophyceae; single-celled algae) and ostracods (Ostracoda; shelled microcrustacea) are known for their sensitivity to salinity. In palaeolimnology, the potential has yet to be tested for quantifying past salinity, lake level, and by inference, climate change, by application of multiple-indicator transfer functions. We used weighted averaging techniques to derive diatom (n=91; r 2 =0.92) and ostracod (n=53; r 2 =0.83) conductivity transfer functions from modern diatom, ostracod and water chemistry data collected in lakes of central, western and northern Turkey. Diatoms were better represented across the full gradient than ostracods, at intermediate levels of conductivity in particular, but both transfer functions were statistically robust. Because transfer functions are not infallible, we further tested the strength and simplicity of salinity response and the potential for identifying characteristic associations of diatom and ostracod taxa in different parts of the salinity gradient, to improve palaeoclimate reconstruction. We identified a subset of 51 samples that contained both diatoms and ostracods, collected at the same time from the same habitat. We used Two-Way Indicator Species Analysis of a combined diatom-ostracod data set, transformed to achieve numerical equivalence, to explore distributions in more detail. A clear ecological threshold was apparent at ~3gl-1 salinity, rather than at 5gl-1, the boundary used by some workers, equating to the oligosaline-mesosaline boundary. Other salinity boundaries were poorly defined, indicating lack of a simple, well-defined salinity response. We did, however, define characteristic associations of taxa, to facilitate the distinction to be drawn between a hydrologically open, fresh lake and an oligosaline lake, in palaeoenvironmental reconstruction. Over the rest of the salinity gradient, we highlighted the potential for the multi-proxy approach to strengthen ostracod-based reconstruction in particular, to overcome the problem of broad apparent tolerance ranges in common halophilic taxa such as Limnocythere inopinata, which often dominate in low-diversity fossil assemblages. The combination of multi-proxy quantitative reconstruction, complemented by qualitative understanding of ecological responses generated by the analysis, remains a powerful tool in Quaternary palaeoclimate research.