Abstract
The mitigation of diffuse sediment pollution requires reliable provenance information so that measures can be targeted. Sediment source fingerprinting represents one approach for supporting these needs, but recent methodological developments have resulted in an increasing complexity of data processing methods rendering the approach less accessible to non-specialists. A comprehensive new software programme (SIFT; SedIment Fingerprinting Tool) has therefore been developed which guides the user through critical data analysis decisions and automates all calculations. Multiple source group configurations and composite fingerprints are identified and tested using multiple methods of uncertainty analysis. This aims to explore the sediment provenance information provided by the tracers more comprehensively than a single model, and allows for model configurations with high uncertainties to be rejected. This paper provides an overview of its application to an agricultural catchment in the UK to determine if the approach used can provide a reduction in uncertainty and increase in precision. Five source group classifications were used; three formed using a k-means cluster analysis containing 2, 3 and 4 clusters, and two a-priori groups based upon catchment geology. Three different composite fingerprints were used for each classification and bi-plots, range tests, tracer variability ratios and virtual mixtures tested the reliability of each model configuration. Some model configurations performed poorly when apportioning the composition of virtual mixtures, and different model configurations could produce different sediment provenance results despite using composite fingerprints able to discriminate robustly between the source groups. Despite this uncertainty, dominant sediment sources were identified, and those in close proximity to each sediment sampling location were found to be of greatest importance. This new software, by integrating recent methodological developments in tracer data processing, guides users through key steps. Critically, by applying multiple model configurations and uncertainty assessment, it delivers more robust solutions for informing catchment management of the sediment problem than many previously used approaches.
Highlights
Numerous studies have used sediment source fingerprinting to investigate specific catchment management problems (Collins et al, 2010a; Gellis and Walling, 2011; Miller et al, 2015; Owens et al, 2017; Collins et al, 2017), yet its application as a standard research tool remains limited
Tracer concentrations can be controlled by numerous environmental factors, such as geology (Laceby and Olley, 2015), soil type, hydrology and topography (Blundell et al, 2009; Jordanova et al, 2012), anthropogenic pollutants (Foster and Charlesworth, 1996) and land use (Walling et al, 1993), which will often result in high withinsource group variabilities if broad source groups such as those based on land use or surface/subsurface sources are used (Pulley et al, 2016)
A low between-source group variability in tracer concentrations will cause tracer non-conservatism to have a larger effect on unmixing model outputs, as the sediment provenance signal used for discrimination is small (Collins et al, 2010a, 2010b; Sheriff et al, 2015; Pulley et al, 2016)
Summary
Numerous studies have used sediment source fingerprinting to investigate specific catchment management problems (Collins et al, 2010a; Gellis and Walling, 2011; Miller et al, 2015; Owens et al, 2017; Collins et al, 2017), yet its application as a standard research tool remains limited. Since the publication by Mukundan et al (2012), several sediment fingerprinting papers have highlighted uncertainties associated with certain procedural steps, such as tracer conservativeness, tracer corrections, weightings and statistical operations (Koiter et al, 2013; Smith and Blake, 2014; Laceby and Olley, 2015; Pulley et al, 2015a; Laceby et al, 2017; Collins et al, 2017; Owens et al, 2017) This questioning of procedures in the sediment fingerprinting approach is necessary for the science to move forward but it is necessary to communicate to land managers a streamlined and robust procedure. A low between-source group variability in tracer concentrations will cause tracer non-conservatism to have a larger effect on unmixing model outputs, as the sediment provenance signal used for discrimination is small (Collins et al, 2010a, 2010b; Sheriff et al, 2015; Pulley et al, 2016)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
