Soil chemistry, temperature and bacterial community composition drive brGDGT distributions along a subarctic elevation gradient

Abstract

Branched glycerol dialkyl glycerol tetraethers (brGDGTs) are bacterial membrane lipids which can be used to reconstruct past terrestrial mean annual air temperature (MAAT) and soil pH values. To reconstruct these environmental conditions in geological archives, we make use of the brGDGT ratios MBTʹ5ME, CBTʹ and IR, which summarize the structural variation in brGDGT lipids. However, the most recent calibrations between the brGDGT-based temperature proxy MBTʹ5ME and temperature on a global scale are characterized by a residual error between 3.8 °C and 6.0 °C. This lack of accuracy of the MBTʹ5ME-MAAT calibration is often attributed to the difference between soil and atmospheric temperatures. Also, previous studies suggested that the variation of the MBTʹ5ME along chemistry and temperature gradients reflects a possible influence of bacterial community changes. Here, we analyzed the effect environmental variables have on brGDGT signatures collected along five elevation gradients in the north of Sweden and Norway, where MAAT changed between –4.7 °C and 2.7 °C and soil pH varies between 3.3 and 5.7. Specifically, we determined the impact of air and soil temperature, the bacterial community composition and soil chemical characteristics. The range of MBTʹ5ME values encountered (0.30–0.70) results in a wide range of reconstructed temperatures (1–13 °C). The use of in situ soil temperature data spanning one year did not improve the correlation with brGDGT MBTʹ5ME values, compared to using a long-term MAAT dataset. Although a temperature gradient was present, soil chemistry apparently determined brGDGTs concentration and distribution on this local scale. Specifically, soils with high cation exchange capacity (CEC) showed an increased concentration of brGDGT Ia, resulting in increased Community Index (CI) and MBTʹ5ME values, both brGDGT ratios that reflect the degree of methylation. Soils with increased pH (pH range 5–6) had a distinct brGDGT fingerprint with generally decreased MBTʹ5ME values, which resulted in the correlation between MBTʹ5ME and soil pH (r = –0.60, p < 0.01). Contrary to expectations, pH was a better predictor of MBTʹ5ME values than temperature (r = 0.47, r = 0. 44, p < 0.01, with MAAT and mean summer soil temperature (MSST), respectively). Soil pH also shaped the bacterial community composition, and a bio-indicator approach was used to narrow down the proposed bacterial producer of brGDGT lipids in high CI (Acidobacteria subgroups 1 and 3) and high pH soils (Acidobacteria subgroups 6 and 7). Building upon previous research, this confirms that brGDGTs respond to similar changes in bacterial community composition across sites. Because of the interplay between temperature and soil chemistry, the relationship between the MBTʹ5ME and soil temperature is clearly complex at the local scale. Further disentangling these environmental drivers is still essential in the development of the MBTʹ5ME proxy as paleothermometer.