Distribution of tetraether lipids in the 25-ka sedimentary record of Lake Challa: extracting reliable TEX86 and MBT/CBT paleotemperatures from an equatorial African lake

Abstract

The distribution of isoprenoid and branched glycerol dialkyl glycerol tetraether (GDGT) lipids was studied in the sedimentary record of Lake Challa, a permanently stratified, partly anoxic crater lake on the southeastern slope of Mt. Kilimanjaro (Kenya/Tanzania), to examine if the GDGTs could be used to reconstruct past variation in regional temperature. The study material comprised 230 samples from a continuous sediment sequence spanning the last 25 ka with excellent age control based on highresolution AMS 14C dating. The distribution of GDGTs showed large variation through time. In some time intervals (i.e., from 20.4 to 15.9 ka BP and during the Younger Dryas, 12.9e11.7 ka BP) crenarchaeol was the most abundant GDGT, whereas at other times (i.e., during the Early Holocene) branched GDGTs and GDGT-0 were the major GDGTconstituents. In some intervals of the sequence the relative abundance of GDGT-0 and GDGT-2 was too high to be derived exclusively from lacustrine Thaumarchaeota, suggesting a sizable contribution from methanogens and other archaea. This severely complicated application of TEX86 palaeothermometry in this lake, and limited reliable reconstruction of lake water temperature to the time interval 25e13 ka BP, i.e. the Last Glacial Maximum and the period of postglacial warming. The TEX86-inferred timing of this warming is similar to that recorded previously in two of the large African rift lakes, while its magnitude is slightly or much higher than that recorded at these other sites, depending on which lake-based TEX86 calibration is used. Application of calibration models based on distributions of branched GDGTs developed for lakes inferred temperatures of 15e18 C for the Last Glacial Maximum and 19e22 C for the Holocene. However, the MBT/CBT palaeothermometer reconstructs temperatures as low as 12 C for a Lateglacial period centred on 15 ka BP. Variation in down-core values of the BIT index are mainly determined by the varying production rate of crenarchaeol relative to in-situ produced branched GDGTs. The apparent relationship of the BIT index with climatic moisture balance can be explained either by the direct influence of lake level and wind strength on nutrient recycling, or by influx of soil nutrients promoting aquatic productivity and nitrification. This study shows that GDGTs can aid in obtaining climatic information from lake records but that the obtained data should be interpreted with care