Building A Library of Source Samples For Sediment Fingerprinting – Potential and Proof of Concept

Purpose: Sediment fingerprinting of fluvial targets has proven useful to guide conservation management. However, the collection and analysis of source samples can make these studies unaffordable. We investigate the potential use of source samples from a basin with similar physiography (using samples from one of a “pair” to evaluate samples from the other) and combining samples from multiple basins (a “library”) to evaluate samples from an additional basin. Methods: Source samples from eight studies were evaluated. The ability to accurately identify samples was evaluated both as part of source verification and by reclassifying source samples as “targets”. Results: Approximately half of cropland samples were identified as targets, both as pairs and with the multi-basin source dataset, indicating that cropland samples could be shared for basins in similar ecoregions and be combined for larger stream systems. Streambank samples were better identified with the multi-basin analysis relative to the pairs; in basins with relatively uniform land use, local streambank samples were critical, but samples from other basins provided key information in those with mixed land use. Inconsistent identification of pasture samples highlighted the need for local samples. Inconsistent identification of forest samples indicated that upland- and riparian-forest samples were distinct. Road samples were identified as both sources and targets and other source types were rarely apportioned as road: these may have the best potential to supplement local source samples. Conclusion: Ultimately, the source verification process that is already used in individual basin studies to evaluate the accuracy of sediment-fingerprinting apportionments was useful for determining how to supplement local source samples with those from other basins.