U–Th–Ra variations in Himalayan river sediments (Gandak river, India): Weathering fractionation and/or grain-size sorting?

Abstract

Understanding the origin of U–Th–Ra variations in the Ganga river sediments is a prerequisite for correctly using U-series nuclides to constrain the sediment transport times in Himalayan rivers. For this purpose, U, Th, and Ra concentrations, along with 238U–234U–230Th–226Ra radioactive disequilibria, were analyzed in bank, bedload and suspended sediments from the Gandak river, one of the main tributaries of the Ganga river. The data confirm that U and Th budgets of the Himalayan sediments are significantly influenced by minor resistant minerals, such as zircon, garnet and Ti-bearing minerals, the dissolution of which required the use of a high-pressure acid digestion process. Most importantly, the results indicate that the variations in (238U/232Th) and (230Th/232Th) activity ratios and 238U–234U–230Th–226Ra disequilibria in sediments along the river alluvial plain mainly reflect modifications in the mineralogical and grain-size compositions rather than the degree of weathering during transport. The (238U/232Th) and (230Th/232Th) activity ratios in the bank and bed sediments are related to variations in the minor primary minerals strongly enriched in U and Th (i.e., zircon, REE-bearing minerals and Ti-bearing minerals), whereas the activity ratios in the suspended load are related to variations in the proportions of clay, Fe-oxyhydroxides and the silt-sand fraction, which contains U- and Th-bearing minor minerals. The data also indicate that 238U–234U–230Th–226Ra disequilibria are strongly influenced by secondary mineral phases: the 230Th budget is likely mainly controlled by Fe-oxyhydroxides, and the 226Ra budget is likely mainly controlled by clay minerals. Therefore, the variations in the 238U–234U–230Th–232Th system in the sediments of the Gandak river cannot simply be interpreted as the result of fractionation due to chemical transformation of the bulk sediment during its transport within the alluvial plain and/or the result of radioactive decay. Consequently, they cannot be used to infer long sediment transport times within the Gandak plain (10–100ka), as previously proposed. Such analytical and interpretative artifacts are certainly not specific to the present study on the Gandak basin. These issues will certainly be encountered anytime this technique is applied to alluvial systems in which the U and Th budgets of the sediments are influenced by “heavy” minerals that can be sorted during the transport of sediments within the plain.