NOAA/WDS Paleoclimatology - Baffin Island Small Lakes Surface Sediment GDGT Data

STUDY NOTES: The GDGTs (glycerol dialkyl glycerol tetraethers) in sediments from 71 Arctic lakes were measured for quantitative concentration values using synthesized GDGT standards at known concentrations.
ABSTRACT SUPPLIED BY ORIGINATOR: The GDGTs (glycerol dialkyl glycerol tetraethers) in sediments from 71 Arctic lakes were measured for quantitative concentration values using synthesized GDGT standards at known concentrations. Each sample was measured in duplicate with uncertainties of only ±1ºC in the TEX86 temperature proxy and ±0.01 in the BIT terrestrial proxy. A comparison of calculated TEX86 proxy temperatures to instrumental data showed no correlation to lake surface temperature (LST) or January, July or mean annual air temperature. TEX86 versus instrumental LST yielded a correlation r2 of only .02, and TEX86 versus instrumental air temperature yielded no correlations as well (r2 of -.06, .09 and -.0006 for mean annual, Jan and July air temperature, respectively). These results were not entirely unanticipated, as TEX86 has been shown to be problematic when applied to nearshore marine sites or small lakes with significant terrestrial input. Therefore, in addition to TEX86, we investigated a number of alternate indices based on average number of ring structures, or ratios of highest-to-lowest ring number, within the suite of GDGT compounds. These alternate indices were calculated for various combinations of ring number, including isolation of the compounds found at highest or lowest concentrations. For each of these alternate 'ring number' indices, there was no significant correlation to instrumental lake surface or average air temperature. The different temperature indices show no correlation to LST for these small lakes, and suggest that reconstructing water temperature with aquatic GDGTs from small, shallow Arctic lakes will not be a practical approach. Non-isoprenoidal GDGTs, primarily produced by soil microbes, are particularly abundant in the Arctic lakes we sampled. BIT values for the 71 Arctic lakes we investigated averaged 0.93, indicating a very high contribution of terrestrial organic matter to these lakes. This is an expected result, as these samples came from relatively small and shallow lakes, and further supports the lack of a strong correlation between TEX temperature indices and LST. Although in preliminary data based on 10 samples the BIT values showed a strong anticorrelation to mean annual precipitation (r2 = -.76), the final results on all 71 lake samples showed a much weaker relationship (r2 = -.23). For mean July precipitation, the correlation is weakest (r2=-.01), whereas for January the relationship is similar to annual precipitation (r2 = -.21). The BIT index in these lakes appears to be dominated by amount of winter snowpack, with an inverse relationship such that lower snowpack equals greater runoff of terrestrial GDGTs into the lakes. To test whether snowpack melt is related to the BIT index, we compared BIT to July air temperature, but found no relationship (r2 = -.03). Comparing BIT values to other parameters potentially linked to runoff also failed to yield any significant correlations (BIT vs transmissivity, r2 = -.02; BIT vs salinity, r2 = -.00003; BIT vs max depth, r2 = .0003). These BIT results indicate some success in developing a precipitation proxy for use in small Arctic lakes. However, the anticorrelation of BIT with mean annual precipitation is striking, as it is opposite what has been found previously, where BIT in river deltas and near-shore marine settings is positively correlated with river runoff. In addition, the correlation with January rather than July precipitation precludes a simple interpretation such as the hypothesis that for small Arctic lake catchments, the flux of terrestrial matter is controlled by the effect of aridity on increased erodability of peaty soils and tundra. Instead, the anticorrelation with only winter precipitation implies a potential protective effect of the snowpack, preventing soil erosion and runoff when the snowpack is larger. Nevertheless, any relationship between BIT and precipitation in Arctic lakes is complex, and the low degree of variance explained (~23%) suggests that the relationship may not be robust.