Potential Biases in TEX86-Reconstructed Temperatures from Lake Malawi, Africa

Several plausible paleoclimate reconstructions have been produced from Lake Malawi sedimentary records using the isoGDGT-based paleotemperature proxy TEX86. Current interpretations of these records assume that the distribution of sedimentary isoGDGTs reflects mean annual lake surface temperatures (LST). In lacustrine environments, however, there are a number of factors in addition to LST that may also contribute to variability in TEX86 values (e.g. subsurface temperatures, seasonal bias, etc.). Proper interpretation of TEX86-based temperature records thus requires a thorough understanding of isoGDGT production in the modern water column of Lake Malawi. This study utilized a combination of lipid and genomic analyses to understand the ecology of isoGDGT-producers and to characterize the timing of isoGDGT production in Lake Malawi. Profiles of POM collected from the northern, central, and southern basins provide a snapshot of where isoGDGT-producers reside in the lake's water column and how much they contribute to overall GDGT production. To test for seasonal biases in the proxy, a multi-year time series of isoGDGT production was generated with settling particulate matter (SPM) collected in sediment traps placed in the northern and southern basin. Results of the POM analyses indicate that there are two distinct archaeal communities that produce isoGDGTs in the water column, though maximum GDGT production occurs just below the thermocline. There are two dominant periods of isoGDGT production in the lake: Jan-Apr and Jun-Sept. Despite this, TEX86-reconstructed temperatures from sediment trap samples appear to be biased towards surface water summer temperatures. This bias seems only to have developed over the last couple decades, however, as a comparison of measured lake water temperatures from 1940-2000 and TEX86 temperatures reveals that prior to 1990 TEX86 temperatures closely matched mean annual temperature of the upper 50 m of the water column.