COMPOSITIONAL HOMOGENEITY OF LODGEMENT TILL MATRIX AND RELIABILITY OF A MODERATELY DISTRIBUTED SAMPLING METHOD

In an effort to establish the degree of compositional homogeneity in the matrix of lodgement till and to test a sampling plan for a study of till composition by the Vermont Geological Survey, we began a study addressing matrix composition and the reliability of a sample to reflect composition over a range of scales. Central Vermont glacial deposits include basal and readvance-related lodgement tills often separated by glacio-lacustrine deposits; however, where till deposits are especially thick (10+ m) we suspect the tills are stratigraphically juxtaposed. We were interested in whether matrix composition could discriminate these tills, vertically and laterally.

For the Survey’s study, a 10 km grid sampling system was adopted, where a sample represents ~100 km². Following USGS standards, analysis is performed on the < 2 mm fraction. To test the reliability of this sampling plan and in the hope of discriminating tills we analyzed the matrix composition at various scales. All sampling was collected in triplicate, where three samples were collected in a cluster about a midpoint with a 25 cm radius. Samples representing the vertical profile were collected at each site. We also sampled the same unit at widespread geographic positions, in the same and in different watersheds, but with similar underlying bedrock. For elemental analysis samples were analyzed by ICP spectrometry in our laboratory and at a commercial laboratory. We have conducted limited analysis of mineral composition by XRD and will eventually combine mineral and elemental composition to characterize the till matrix.

Results to date indicate that, over similar bedrock, nearly negligible variability exists in matrix composition, irrespective of scale of comparison. Variation that exists may grow with scale of comparison but there is no recognizable pattern and variation is within acceptability for identifying units. Basal and readvance tills are compositionally identical. Sampling in a 10 km grid to characterize composition is validated, but composition is a poor tool for discriminating basal and readvance lodgement tills. We have started a high-resolution study across a major lithologic boundary to determine the rate of bedrock uptake in the till matrix and will compare this to clast “renewal distance” to establish a predictive model of composition.