AmericanChemicalSociety.com

The principal goals of this award are to: 1) Investigate the environmental controls on distributional variations in a suite of nine branched glycerol dialkyl glycerol tetraethers (GDGTs) preserved in lake sediments; 2) Test and develop models to relate GDGT variations in lake sediments to environmental parameters, principally temperature and lake pH, and 3) Test the applicability of these models to reconstructing past environmental variations by applying the new GDGT proxy to lake sediment cores.

Funds from this award were received June 1, 2009.  Since that time we have made enormous progress in field and laboratory aspects of the proposed work, as well as in presenting and publishing our results.

Thus far we have measured the relative abundances of branched GDGTs in surface sediments from a suite of 46 lakes in tropical East Africa.  These lakes span large gradients in temperature, pH, and other physical and chemical characteristics. Except in the case of saline lakes, branched GDGTs are extremely abundant in these lakes.  More importantly, the relative abundances of the branched GDGTs clearly vary in response to mean annual air temperature (MAAT) and lake pH (See TOC figure plotting observed vs. GDGT-estimate MAAT and pH), and can indeed be used to predict mean annual air temperature with a high degree of accuracy.

That said, previously published calibrations relating branched GDGT variations to air temperature, based upon GDGTs in soils, are not applicable to our lake dataset, indicating different microbial sources of branched GDGTs in lakes and soils.  This finding is supported by that fact that limnological factors such as lake morphometry, pH, and near-bottom dissolved oxygen concentration appear to influence sedimentary GDGT distribution in lakes.  These results were just published in Geochimica et Cosmochimica Acta (Tierney et al., 2010), lead authored by my graduate student (now a post-doc at Lamont-Doherty Earth Observatory), Dr. Jessica Tierney.

To further demonstrate that branched GDGTs in lake sediments have different sources than in soils, in July 2009, we collected a suite of soil samples along an elevation gradient in the Rwenzori Mountains in East Africa, as well as numerous soil samples from within the catchment areas of East African lakes in our sample set using funds from this award.  Analysis of these samples has been completed, and the results support our contention for in-situ production of branched GDGTs in lacustrine environments.  There are statistically significant differences in GDGT distributions between soils and adjacent lake sediments at all elevations.  Interesting, in highly water-saturated soils, branched GDGTs begin to resemble the GDGTs in nearby lakes (although they still retain significant differences).  This suggests that the presence of water is likely a key control on the relationship between microbial membrane lipids and environmental parameters, and/or the microbial consortium producing GDGTs.  This work was presented at the American Geophysical Union meeting in December, 2009 (Loomis et al., 2009), and will be submitted for publication by the end of 2010, with my current graduate student, Ms. Shannon Loomis, as lead author.

Our current and future work is proceeding along three separate paths.

First, we are now beginning analyses of branched GDGTs in temperate-zone lake sediments.  In May, 2009, we installed a sediment trap equipped with thermistors in South King Pond, Vermont, a small, deep, alkaline lake in northern Vermont.  We collected near-monthly measurements of water column chemistry and temperature, and have filtered particulate organic matter from the surface waters and hypolimnion to test for the presence and relative abundances of branched GDGTs in each.  This work will continue through this summer, by which time we will have 1.5 years of sediment trap samples collected at bi-weekly intervals to test the extent to which water column chemistry and temperature influence branched GDGTs in temperate lakes, and whether these relationships are similar to those we observe in tropical East Africa.  Samples for this work have been prepped, and are awaiting analysis by LC-MS.  This work will form the basis for a senior thesis and publication by an undergraduate in my laboratory, Ms. Ana Hereux.

Second, we have added 30 new East African lakes to our initial set of 46, with the aim of more fully examining the influence of secondary variables, such as water column nutrient and dissolved oxygen concentrations on branched GDGTs in these lakes.  Expansion of our calibration dataset will also allow us to examine the extent to which we can detect the influence of small temperature variations within short segments of our calibration dataset.  Our initial calibration data consisted of lakes spanning nearly 4000 m elevation and 25 ^oC.  However, temperature changes in East Africa during the Plio-Pleistocene likely much smaller than this; for instance, temperatures rose ~4 ^oC during the last glacial termination.  Expanding our calibration sample set will provide us with enough samples to test the response of, and perhaps model, branched GDGTs across short, 5-10 ^oC segments of our data- a more realistic approach to reconstructing past temperature variations.  Samples for this work are in preparation, and will be analyzed during fall, 2010.

Finally, we have applied branched GDGTs to reconstruct temperature variations in the sediments of Lake Challa, East Africa, during the past 20,000 years BP.  We assume that the magnitude and patterns of temperature variations during the last glacial terminations (~20-10 kyr BP) at this site are similar to those established from nearby lakes and fossil groundwater deposits, providing checks on the reliability of our new proxy.  Based upon our published and new data we are experimenting with different statistical methods to model the response of branched GDGTs to temperature variations, and applying these models to our Sacred Lake dataset.  Our preliminary data suggest that statistical methods that utilize all nine branched GDGT compounds, as opposed to various models based upon ratios (e.g. the so-called MBT and CBT ratios, which charcterize the methylation and cyclization of these GDGTs) provide the most robust temperature reconstructions.  This work is ongoing; publication will depend upon the finalization of our other lake surface sediment data.  Ms. Loomis is spearheading both of these efforts.