Search

It is unclear whether or not the paleotemperature proxies based on the branched glycerol dialkyl glycerol tetraethers (brGDGTs) in soil exhibits any seasonal bias in temperature reconstruction. To test this, we examined the brGDGT distributions and compared the mean annual air temperature (MAAT) estimates inferred from the brGDGT-based paleothermometers for surface soils from Nanyang (NY) and Kunming (KM), two sites in China with similar MAAT, but with contrasting climate seasonality (i.e. different summer and winter temperatures). If a significant bias towards summer or winter temperature were to occur, the soils from the locality with stronger seasonality would be expected to yield different temperature estimates than those from the other locality. The results show that, at a regional isothermal scale, the most important environmental control on the soil brGDGT distribution is pH at NY but soil water content (SWC) at KM, suggesting a regional difference in the key environmental control on the brGDGT distributions. MBT′ and CBT are both controlled primarily by pH at NY but by SWC at KM. In contrast, the MBT′5ME index, known to be dependent on MAAT alone, displays no relationship with pH or SWC at both sites, allowing us to evaluate the impact of temperature seasonality. MBT′5ME is statistically higher at the NY soils than at the KM soils; a MBT′5ME difference of 0.04 units is equivalent to a difference of ca. 1.3 °C in MAAT. Likewise, the MAATmr estimates produced from multiple linear regression calibration for the two soil sites differ by only 0.8 °C. This suggests that there is no significant bias towards summer temperature in the brGDGT-based paleothermometer as the mean summer temperature differs by 6.4 °C at the two soil sites. By extension, temperature reconstructed from the brGDGT-based proxies for the loess-paleosol sequences in the southern Chinese Loess Plateau is unlikely to be biased towards summer.

The methylation index (MBT or MBT′) of branched glycerol dialkyl glycerol tetraethers (brGDGTs) in soils along altitudinal transects of some mountains, e.g. Mount (Mt.) Shennongjia and Mt. Gongga, exhibited no correlation with altitude or mean annual air temperature (MAT) alone. However, a significant correlation between MBT′, cyclization ratio (CBT) of brGDGTs and MAT for each transect could still be found. To investigate the reason, we reanalyzed the brGDGTs in soils from Mt. Shennongjia using an improved liquid chromatography method and examined whether the newly described 6-methyl brGDGTs were the main cause of the large scatter in the correlation between MBT′ and MAT. In contrast to MBT′, the MBT′5ME, based on 5-methyl brGDGTs alone, correlated significantly with MAT (R2 0.89, p