Statistical evaluation of stream sediment geochemistry in interpreting the river catchment of high-grade metamorphic terrains

Abstract Stream sediment geochemistry is extensively used in mineral exploration and environmental studies. However, quantitative assessments of the effectiveness of stream sediment geochemistry for describing upstream lithologies are rare, especially in high-grade metamorphic terrains. This study statistically evaluates whether stream sediment geochemistry can aid in recognizing variations in upstream geology in several high-grade metamorphic lithotectonic units having different metamorphic and tectonic histories, including the Highland Complex, Vijayan Complex, Wanni Complex, and Kadugannawa Complex of Sri Lanka. For this study, concentrations of 21 elements were measured in 2080 stream sediment samples collected from the Walawe River, Maha Oya, Gala Oya basins located on above lithotectonic units and Uma Oya, Belihul Oya, Badulu Oya basins situated adjacent to each other on the Highland Complex. These rivers flow across dry, intermediate and wet zones of Sri Lanka, with river courses having both slope (> 20°) and flat ( H tests show that both major and trace element levels in sediments from the four different lithotectonic units, as well as from adjacent Uma Oya, Badulu Oya and Belihul Oya basins, are significantly different. Discriminant function analysis appropriately classifies sediments into the four different lithotectonic units with an accuracy of 91.9%. This method also classifies sediments into river basins which share common lithology being situated adjacent to each other in the same lithotectonic unit with an accuracy of 89.5%. This strongly suggests that stream sediment geochemistry is capable of describing the upstream regional scale as well as local scale lithological changes at a great accuracy in complex high-grade metamorphic terrains. In both cases use of channel slope and basin climatic zone as additional variables does not significantly increase overall or individual accuracy in classification.