Fingerprinting sub-basin spatial suspended sediment sources by combining geochemical tracers and weathering indices

Abstract

Transport and deposition of fine-grained sediment, a pervasive nonpoint source pollutant, cause deleterious off-site impacts for water quality and aquatic ecosystems. Sediment fingerprinting provides one means of identifying the spatial sources of mobilised sediment delivered to fluvial systems in order to help target sediment control strategies and uptake of such source tracing procedures has been steadily increasing. Nonetheless, there remains a need to continue testing and comparing different composite signatures for source discrimination including the incorporation of physically grounded information relevant to erosion patterns. Accordingly, the objective of this study was to compare the discrimination and apportionment of sub-basin spatial suspended sediment sources in a mountainous basin in northern Tehran, Iran, using composite signatures comprising conventional geochemical tracers combined with lithological weathering indices or only the former. The list of conventional geochemical properties comprised Al, Ca, Cu, Fe, K, Mg, Mn, Na, Ni, Sr, Ti, and Zn whilst three weathering indices were included: the chemical index of alteration (CIA), the weathering index of Parker (WIP), and the indicator of recycling (IR) which were all calculated based on elemental oxides. Using a composite signature combining conventional geochemical tracers and one weathering index (IR), the relative contributions from the sub-basin spatial sources were estimated at 1 (Imamzadeh Davood; 1.4%), 2 (Taloon; 13.4%), 3 (Soleghan; 35.9%), and 4 (Keshar; 48.4%) compared with corresponding respective estimates of 0.7%, 45.5%, 40.2%, and 13.3% using conventional geochemical tracers alone. Wald-Wolfowitz Runs test pairwise comparisons of the posterior distributions of predicted source proportions generated using the two different composite signatures confirmed statistically significant differences. These differing proportions demonstrated the sensitivity of predicted source apportionment to the inclusion or exclusion of a weathering index providing information reflecting the relative coverage of more erodible lithological formations in each of the sub-basins (32.7% sub-basin 1, 53.6% sub-basin 2, 58.5% sub-basin 3, and 63.2% sub-basin 4). The outputs of this study will be used to target sediment mitigation strategies.

Keywords: Data mining; Geochemical tracers; Modified MixSIR Bayesian model; Sediment tracing; Sub-basin; Weathering indices.