Applicability and calibration of the TEX86 paleothermometer in lakes

We have conducted a global survey of archaeal glycerol dialkyl glycerol tetraether (GDGT) lipids in lake sediments in order to develop the TEX86 paleotemperature proxy for application in continental systems Surface sediments of 46 globally distributed lakes were analyzed for GDGT, but isoprenold GDGT derived from aquatic Crenarchaeota, were only unambiguously detected in 20 of the 46 lakes analyzed. Aquatic crenarchaeotal GDGT were detected mainly in sediments from large lakes (>4000 km(2)) and hydrothermal or volcanic lakes, suggesting that in some (mostly smaller) lakes either aquatic Crenarchaeota are present at low abundance resulting in sedimentary lipids below levels of detection, or they are absent Branched GDGT, thought to be derived primarily from soil bacteria, were identified in all lake sediments analyzed. Correlation of the TEX86 in those lakes with sufficient amounts of putative crenarchaeotal GDGT with annual mean lake temperature is reasonably good (r(2) = 0.68, N = 20) In order to reduce the influence of soil derived isoprenoid GDGT on the TEX86 lake temperatures, we have applied a filter based on relative soil derived Inputs (as determined by the BIT [Branched and Isoprenoid Tetraether] index) which results in a calibration relationship with r(2) = 0 86 (N = 12) and an estimated temperature error of 3.6 degrees C Our results suggest that the TEX86 should be applied only in lakes with sufficient production of GDGT by aquatic Crenarchaeota relative to isoprenoid GDGT derived from soil in the watershed or other aquatic sources. (C) 2009 Elsevier Ltd. All rights reserved.