Branched GDGT-based temperature calibrations from Central European lakes

Abstract

Branched glycerol dialkyl glycerol tetraethers (brGDGTs) are common in lake sediments, where they are frequently employed to infer mean annual air temperatures (MAAT) or air temperatures of months above freezing (MAF) using the MBT'5Me lipid paleothermometer. The reliable reconstruction of such temperatures, however, requires robust calibration functions. Here, we investigated brGDGT distributions in surface sediments from 41 freshwater lakes located along an altitudinal gradient across the Alps (Central Europe) and spanning a MAAT range from 1.3 to 12.9 °C. Linear regression analysis demonstrates that fractional abundances of brGDGTs are strongly correlated with MAAT and MAF allowing to establish regional MBT'5Me-based transfer functions: MAAT (°C) = −2.19 + 31.91 × MBT'5Me (r2 = 0.72; RMSE = 1.51 °C) and MAF (°C) = 4.81 + 15.64 × MBT'5Me (r2 = 0.64, RMSE = 0.92 °C). Stepwise forward selection yielded the following, alternate temperature calibrations: MAAT (°C) = 7.11 + 67.66 × Ib – 13.54 × IIIa (r2 = 0.72; RMSE = 1.47 °C) and MAF (°C) = 5.19 + 16.22 × Ia (r2 = 0.66; RMSE = 0.98 °C). Our results demonstrate that the above calibration functions allow precise temperature reconstructions using lacustrine sediment records. However, our data also show that high-altitude lakes are more prone to warm bias and that lakes characterized by comparatively high abundances of 6-methyl brGDGTs show aberrant behaviors with temperature offsets up to ∼7 °C. Determining the IR6Me, as an independent control complementary to the MBT'5Me, is thus essential to validate the robustness of brGDGT-based temperature reconstructions of past climates. A cut-off value of 0.50 for the IR6Me is here proposed, after which MBT'5Me-reconstructed MAATs from lacustrine archives should be regarded as unreliable.