Understanding the transfer of contemporary temperature signals into lake sediments via paired oxygen isotope ratios in carbonates and diatom silica: Problems and potential

Abstract

Although the oxygen isotope composition (δ18O) of calcite (δ18Ocalcite) and, to a lesser extent, diatom silica (δ18Odiatom) are widely used tracers of past hydroclimates (especially temperature and surface water hydrology), the degree to which these two hosts simultaneously acquire their isotope signals in modern lacustrine environments, or how these are altered during initial sedimentation, is poorly understood. Here, we present a unique dataset from a natural limnological laboratory to explore these issues. This study compares oxygen and hydrogen isotope data (δ18O, δ2H) of contemporary lake water samples at ~2-weekly intervals over a 2-year period (2010–12) with matching collections of diatoms (δ18Odiatom) and calcite (δ18Ocalcite) from sediment traps (at 10 m and 25 m) at Rostherne Mere (maximum depth 30 m), a well-monitored, eutrophic, seasonally stratified monomictic lake in the UK. The epilimnion shows a seasonal pattern of rising temperature and summer evaporative enrichment in 18O, and while there is a temperature imprint in both δ18Odiatom and δ18Ocalcite, there is significant inter-annual variability in both of these signals. The interpretation of δ18Odiatom and δ18Ocalcite values is complicated due to in-lake processes (e.g. non-equilibrium calcite precipitation, especially in spring, leading to significant 18Ocalcite depletion), and for δ18Odiatom, by post-mortem, depositional and possibly dissolution or diagenetic effects. For 2010 and 2011 respectively, there is a strong temperature dependence of δ18Ocalcite and δ18Odiatom in fresh trap material, with the fractionation slope for δ18Odiatom of ca. −0.2‰/°C, in agreement with several other studies. The δ18Odiatom data indicate the initiation of rapid post-mortem secondary alteration of fresh diatom silica (within ~6 months), with some trap material undergoing partial maturation in situ. Diatom δ18O of the trap material is also influenced by resuspension of diatom frustules from surface sediments (notably in summer 2011), with the net effect seen as an enrichment of deep-trap 18Odiatom by about +0.7‰ relative to shallow-trap values. Contact with anoxic water and anaerobic bacteria are potentially key to initiating this silica maturation process, as deep-trap samples that were removed prior to anoxia developing do not show enrichment. Dissolution (perhaps enhanced by anaerobic bacterial communities) may also be responsible for changes to δ18Odiatom that lead to increasing, but potentially predictable, error in inferred temperatures using this proxy. High resolution, multi-year monitoring can shed light on the complex dynamics affecting δ18Odiatom and δ18Ocalcite and supports the careful use of sedimentary δ18Odiatom and δ18Ocalcite as containing valuable hydroclimatic signals especially at a multi-annual resolution, although there remain substantial challenges to developing a reliable geothermometer on paired δ18Odiatom and δ18Ocalcite. In particular, δ18Odiatom needs cautious interpretation where silica post-mortem secondary alteration is incomplete and diatom preservation is not perfect, and we recommend dissolution be routinely assessed on diatom samples used for isotopic analyses.