Validation of the MBT-CBT paleotemperature proxy:Effects of environmental and seasonal variability in soils and lacustrine sediments.

Abstract

Branched glycerol dialkyl glycerol tetraethers (GDGTs) are bacterial derived membrane lipids found ubiquitously in soils and lacustrine sediments (Weijers et al. 2006b, Blaga et al. 2009). The degree of methylation and cyclization of these lipids have been shown to be dependent on the temperature and pH of the growth environment (Weijers et al. 2007a). These relationships are the basis of the MBT-CBT proxy, which has been used to reconstruct paleotemperature from marine and lacustrine sediment archives (Weijers et al. 2007b, Blaga et al. 2010). Here, we aim to test the validity of the MBT proxy in terrestrial soils and lake sediments to determine whether branched GDGT distributions do refl ect annual mean air temperature (MAT) of the watershed, as suggested by studies, and how other environmental factors, such as seasonality of growth and sub-environments within a watershed, might infl uence the preserved MBT-CBT temperature signal. GDGT-derived annual MAT was compared with instrumental temperature measurements at three sites in the continental United States. Watershed soils were collected monthly for one year under three different types of vegetative cover in Minnesota and Ohio. In Florida, soils were collected twice from an open fi eld environment. Sediment cores were collected from corresponding lakes in each of the three states. We observed no signifi cant differences in soil proxy-derived annual MAT or soil GDGT concentration with seasonal changes in temperature at any of the three sites. Concentrations of GDGTs in the soil were found to have a slight positive correlation with organic carbon content. The effects of vegetative cover on proxy estimates of annual MAT were minimal. Only under deciduous vegetation in Minnesota and Ohio, did proxy-derived annual MAT differ signifi cantly from instrumental measurements. Soil GDGT concentration was unaffected by vegetation type in Minnesota, and in Ohio, pine soils had consistently higher concentrations, usually by an order of magnitude, than other vegetation types. The CBT proxy was found to be an accurate estimate of soil pH in some sub-environments, but in Minnesota and Ohio deciduous soils and Ohio open fi eld soils, CBT-pH differed signifi cantly from measured values. In the sediments of all three lakes studied, the MBT-CBT proxy provided a good estimate of measured annual MAT within the error of the proxy. Proxy-derived temperatures from the surface sediments of all lakes studied were cooler than corresponding instrumental measurements. These cooler surface temperatures could be attributed to in-situ production of branched GDGTs, but surface sediment MBT-MAT only differs signifi cantly from instrumental measurements in Bath Pond, Ohio.