Characterizing Provenance of the Late Wisconsinan Rainy Lobe Using Fine-Fraction Geochemistry and Detrital Zircon Geochronology

Abstract

Till of the Late Wisconsin Rainy lobe, which emanated from the Labradoran sector of the Laurentide ice sheet (LIS), is exposed at the surface from SW Minnesota to the extreme NE part of the State. The Rainy lobe advanced to its maximum limit in southwestern Minnesota well prior to the Last Glacial Maximum (ca. 27-30 ka BP) and retreated into Ontario by 17.9 ka BP. This till exhibits dramatic spatial and temporal changes in provenance from the Hewitt till of SW Minnesota to the Independence till in the NE. Two models have been proposed to explain the lithological differences (particularly carbonate) in Rainy Lobe tills. Goldstein (1989) postulated that the downglacier increase in carbonate in the Hewitt till was the result of progressive incorporation, by regelation or deformation, of older underlying till that was rich in carbonate. Larson (2008) concluded that the changes in sedimentology and landforms record systematic changes in provenance related to changing basal boundary conditions in the interior of the LIS. Early in this phase of glaciation, the sediments reflect long-distance transport from Hudson Bay, and later phases reflect increased proportions of felsic shield lithologies and Duluth Complex rocks with a corresponding decrease in carbonate. These two models of Rainy lobe till sedimentology are evaluated using mixing models, till matrix geochemistry, and detrital zircon geochronology. The multicomponent mixing model is developed to examine sedimentological variability by incorporation of older, underlying tills (e.g. Goldstein, 1989). The mixing model shows that the Hewitt till does not lie on the mixing curve, suggesting that mixing is not a viable model for the origin of the sedimentary variability in the Hewitt till. To evaluate the model of Larson (2008), which implies long vs. short transport distances, twenty-eight samples collected along a transect from SW to NE Minnesota, and eight samples collected from the Hudson Bay lowlands, were processed and sent for geochemical analysis. Fifteen of these samples were processed and analyzed for detrital zircon geochronology using laser-ablation, ICPMS. Results of a 48-element analytical suite were run through a principal component analysis. Factors 1 and 3 distinguished mafic vs felsic igneous rock geochemical signatures and carbonate content, respectively. Results show that Core SLL (Independence) plots positively on factor 1 indicating a short mean transport length. Core CSS (Hewitt, north Wadena drumlin field) in central MN represents an intermediate mean transport length, while core TG (Hewitt, south Wadena drumlin field) in far SW MN has the longest mean transport length. In addition, the samples with the longest transport length plot in high carbonate space with the calcareous Hudson Bay lowland samples, positive on Factor 3. A Kolmogorov-Smirnoff (K-S) and degree of likeness test were used to statistically compare detrital zircon age populations. Results from these statistical tests reveal that high carbonate Hudson Bay lowland ages are statistically similar to samples from central Minnesota (core CSS). Geochemistry and detrital zircon analyses support the model of Larson (2008). Early deposits of the Rainy lobe in SW Minnesota are geochemically similar to the high-carbonate Hudson Bay lowland samples, indicating a distal provenance. This similarity is also observed in the detrital zircon results from statistical analyses. Subsequently younger deposits lose the Hudson Bay lowland signature and start to incorporate more felsic craton and eventually mafic signatures of the Mid-Continent rift system of NE MN.