No abstract

Sediment composition in modern fluvial settings is commonly assessed regarding spatial but rarely temporal variability, potentially leading to a bias of unknown extent. Here, we present the grain‐size distribution, bulk chemical and mineralogical composition of a time‐series set of 36 suspended sediment samples from the Brahmaputra river, as well as clay and heavy mineral analysis of selected samples. Sampling covers the June–November 2021 period, which included two major flooding events. We show that the two flooding events are characterized by contrasting grain size, with the first event characterized by a grain‐size minimum and the second by a grain‐size maximum. Although grain sizes of the first flood and the period after the second are similar, their compositions differ significantly, highlighted by a factor‐two decrease of biotite largely compensated by an increase in quartz. By contrast, the content of garnet, clinopyroxene, sillimanite, and rutile increased compared to epidote and amphibole during the second flood event. By relating the results to spatio‐temporal rainfall and discharge patterns and basin morphology, we conclude that the first flooding primarily mobilized hydraulically pre‐sorted sediments from the exposed sandbars of the floodplains, while those sandbars are already submerged during the second flooding in a single‐channel system, resulting in higher sediment contributions from highland tributaries draining igneous and high‐grade metamorphic rocks. Such temporal variations pose constraints on the interpretation of compositional differences between individual samples regarding sediment provenance and dispersal and should be considered in studies of modern drainage basins as well as ancient sediment routing systems.

Superimposed signatures of grain size effect, lithology and chemical processes on fluvial sediments need to be resolved to answer the profound research questions related to sediment provenance and processes. Hydraulic forces sort the sediments into different grain size classes in the river water column. The finest fraction is transported as the suspended sediments which have different sediment composition than the bed sediments. Thus, suspended sediment may provide additional information about earth surface processes, such as chemical weathering. Hydraulically sorted river bed sediments may or may not provide this information, as bulk sample may be depleted in finer grains by hydraulic processes. Bed sediments that are easily sampled from the sand bars and river banks are often investigated to study the weathering intensity and sediment provenance. It is thus crucial to identify and quantify the specific grain size classes in the bulk sample to be investigated for the research question at hand. End Member Modelling Algorithms (EMMA) for grain size distribution is a useful tool to unmix the grain size population into geological meaningful end members. We applied Hierarchical alternating least squares nonnegative matrix factorization (HALS-NMF) algorithm to unmix the grain size data (62 samples) of river bed sediments collected from the freshly exposed sand bars of the Brahmaputra river over a stretch of 550km. The grain size distribution of the finest end member (mean=18μm) is closely approximated to be of the surface sediment grain size distribution reported previously for the Brahmaputra river. Thus, we were able to quantify the relative contribution of suspended sediment to the bed sediment of the Brahmaputra trunk.Results show that the contribution of the suspended sediments in the bed sediment is higher at the lower reaches of the river near floodplain outlet, possibly due to the reduced flow energy in downstream regions. The findings may also be used to select samples and grain size classes for additional geochemical and mineralogical study in order to interpret signals of weathering, provenance and physical processes in the Brahmaputra's large dynamic floodplains at a finer spatial scale.

No abstract