Evaluation of bacterial glycerol dialkyl glycerol tetraether and ²H–¹⁸O biomarker proxies along a central European topsoil transect

Abstract

Molecular fossils, like bacterial branched glycerol dialkyl glycerol tetraethers (brGDGTs), and the stable isotopic composition of biomarkers, such as δ2H of leaf wax-derived n-alkanes (δ2Hn-alkane) or δ18O of hemicellulose-derived sugars (δ18Osugar), are increasingly used for the reconstruction of past climate and environmental conditions. Plant-derived δ2Hn-alkane and δ18Osugar values record the isotopic composition of plant source water (δ2Hsource-water and δ18Osource-water), which usually reflects mean annual precipitation (δ2Hprecipiation and δ18Oprecipiation), modulated by evapotranspirative leaf water enrichment and biosynthetic fractionation (εbio). Accuracy and precision of respective proxies should be ideally evaluated at a regional scale. For this study, we analysed topsoils below coniferous and deciduous forests as well as grassland soils along a central European transect in order to investigate the variability and robustness of various proxies and to identify effects related to vegetation. Soil pH values derived from brGDGTs correlate reasonably well with measured soil pH values but are systematically overestimated (ΔpH = 0.6±0.6). The branched vs. isoprenoid tetraether index (BIT) can give some indication whether the pH reconstruction is reliable. Temperatures derived from brGDGTs overestimate mean annual air temperatures slightly (ΔTMA=0.5 ∘C ± 2.4). Apparent isotopic fractionation (εn-alkane/precipitation and εsugar∕precipitation) is lower for grassland sites than for forest sites due to signal damping; i.e. grass biomarkers do not record the full evapotranspirative leaf water enrichment. Coupling δ2Hn-alkane with δ18Osugar allows us to reconstruct the stable isotopic composition of the source water more accurately than without the coupled approach (Δδ2H = ∼-21 ‰ ± 22 ‰ and Δδ18O = ∼-2.9 ‰ ± 2.8 ‰). Similarly, relative humidity during daytime and the vegetation period (RHMDV) can be reconstructed using the coupled isotope approach (ΔRHMDV=∼-17±12). Especially for coniferous sites, reconstructed RHMDV values as well as source water isotope composition underestimate the measured values. This can likely be explained by understorey grass vegetation at the coniferous sites contributing significantly to the n-alkane pool but only marginally to the sugar pool in the topsoils. Vegetation-dependent variable signal damping and εbio (regarding 2H between n-alkanes and leaf water) along our European transect are difficult to quantify but likely contribute to the observed underestimation in the source water isotope composition and RH reconstructions. Microclimate variability could cause the rather large uncertainties. Vegetation-related effects do, by contrast, not affect the brGDGT-derived reconstructions. Overall, GDGTs and the coupled δ2Hn-alkane–δ18Osugar approach have great potential for more quantitative paleoclimate reconstructions.