Uncertainty Assessment of Sediment Contribution in Tributaries to the Upper Esopus Creek, New York

Abstract

This study highlights the interplay between spatial scale, sub-catchment characteristics, and sediment dynamics in order to address the significant challenge of quantifying sediment source uncertainties across various spatial scales in the Upper Esopus Creek watershed, New York State, where sediment reduces water quality for more than nine million residents of New York City. The sediment samples collected from the Stony Clove and Woodland Creek sub-catchments, from 2017 to 2020, were analyzed with the Bayesian chemical mass balance (CMB) approach to measure source contributions and related uncertainties; discriminant power analysis and analysis of variance (ANOVA) tests helped identify key tracers with low p-values to distinguish sources clearly and lower uncertainty. The four primary sediment sources previously identified were forest, glacial till, lower and upper lacustrine, and terrace and bank alluvium. Findings indicate that analyzing each sub-catchment separately uncovered distinct source patterns and greatly reduced uncertainty compared to using combined methods. In contrast to Woodland Creek, which showed stable dominance of terrace and bank alluvium (60–95%), Stony Clove Creek was dominated by lower and upper lacustrine sources (60–80% contribution), with contributions changing dynamically by storm event magnitude and shifting across early versus later hydrograph stages. Storm event analysis in Stony Clove revealed flow-triggered source changes, with clear spikes in lower and upper lacustrine contributions. This research assists watershed managers by demonstrating how storm timing, scale-focused analysis, and careful tracer choices enhance sediment source tracking and reduce uncertainty in fingerprinting studies, supporting precise erosion control strategies and effective resource use.