VARIATIONS IN THE DISTRIBUTION OF NON-CLAY AND CLAY MINERALS IN SURFACE SAMPLES FROM THE WESTERN NORDIC SEAS (58° TO 75°N): ASSOCIATION WITH BEDROCK GEOLOGY, SURFACE CURRENTS, AND ICE-RAFTING

In order to understand past changes and interactions in cryosphere and ocean dynamics we need to describe and understand present-day conditions. Our interest is in the export of glacially eroded sediments from present and past ice sheets that covered land areas around the Nordic Seas. Our proxy is the quantitative X-ray diffraction (qXRD) estimates of the non-clay and clay mineralogy of the bulk sea floor sediment. Our study is based on 156 surface samples obtained during from several research cruises. We have good coverage for the NE Greenland shelf (70°-75°N, n = 26), E Greenland Shelf (66°-70°N, n = 30)), the Iceland shelf (n = 49), and the Irminger Basin (n = 38). We have very limited samples from off SE Greenland (n = 3), and modest coverage for the Greenland Sea slope and basin (n = 11) The qXRD intensity data were processed in the “Rockjock v6” software. A total of 24 non-clay and 10 clay minerals were initially included, but this number was reduced to 9 non-clay and 6 clay minerals by combining minerals, such as alkali and plagioclase feldspars, smectites, and illites. Based on a combination of 1) nearness to a glacial source, and 2) appropriate ocean currents and drift-ice trajectories, we initially hypothesized that we should be able to distinguish 6 mineral compositional assemblages. This hypothesis was tested directly by step-wise Discriminant Function Analysis (DFA) and indirectly by application of Fuzme, a k-mean” clustering program. A complicating factor is the influence of calcite (which can be detrital or in situ biogenic), which in this study varied between 0 and 87 wt% thus having a potentially significant impact on other mineral wt% estimates. To test for the impact of calcite on our understanding of regional compositional patterns the data were recalculated with calcite outside of the sum. Stepwise DFA of our 6 original clusters indicated that 70% of the samples were correctly identified. Five compositional groupings were obtained by using Fuzme and there is a highly significant association (p < 0.001) between our original 6 clusters and the 5 obtained by the k-means clustering algorithm. The coherence of the regionally based clusters indicates that the supply of sediment from the adjacent bedrock outcrop is the strongest control on sediment mineralogy, but this is modulated by ice and current transport.