BRANCHED GDGT DISTRIBUTIONS IN MAINE LAKES AND WATERSHED SOILS: DISENTANGLING LACUSTRINE AND SOIL INPUTS TO THE SEDIMENTARY RECORD

Branched glycerol dialkyl glycerol tetraethers (brGDGTs) are biomarkers that are becoming increasingly popular for reconstructing past continental temperature via the MBT’_5Me proxy. brGDGTs are produced in soils, peats, and in-situ in lakes. While many paleoclimate reconstructions have applied MBT’_5Me to lacustrine sediments, which often contain long and continuous archives and are preferable for paleoclimate studies, inputs of brGDGTs from watershed soils have the potential to complicate paleotemperature reconstructions because available calibrations of MBT’_5Me to temperature based on soils versus lake sediments differ significantly. Here we examine brGDGT distributions along transects from the lake edge to the depocenter at two adjacent sites that differ in size: Sebago Lake and Perley Pond (Maine, USA). Sebago Lake is one of the larger (117 km² surface area) and deeper (>100 m) lakes in New England. At this site, we determined that watershed soils differ in brGDGT composition from Sebago Lake surface sediments, and that lake sediments collected from shallow water depths and in front of major river inflows exhibit characteristics of mixed soil and lacustrine brGDGT inputs. However, in water depths greater than ~40 m, brGDGT distributions of surface sediments across the entire lake were remarkably similar and indicate a dominant lacustrine source. At nearby Perley Pond, a very small lake (surface area is 0.32 km²; 8.2 m max depth) situated within the larger Sebago Lake watershed, we also find that lake surface sediments and surrounding soils can be readily distinguished by their brGDGT distributions. However, brGDGT distributions of Perely Pond surface sediments are more similar to those of the watershed soils and differ significantly from Sebago Lake surface sediments. This result suggests that smaller lakes, such as Perley Pond, are more susceptible to containing mixed inputs of soil and lacustrine-derived brGDGTs. This result is important because it has implications for choosing sites for brGDGT-based paleoclimate studies.