The application of sediment fingerprinting to floodplain and lake sediment cores: assumptions and uncertainties evaluated through case studies in the Nene Basin, UK

Fine sediment has been shown to be a major cause of the degradation of lakes and rivers and, as a result, research has been directed towards the understanding of fine sediment dynamics and the minimisation of sediment inputs. The use of tracers within a sediment fingerprinting framework has become a heavily used technique to investigate the sources of fine sediment pressures. When combined with the use of historically deposited sediment, the technique provides the opportunity to reconstruct past changes to the environment. However, alterations to tracer signatures during sediment transport and storage are a major potential source of uncertainty associated with tracer use. At present, few studies have quantified the uncertainties associated with tracer use. This paper investigated uncertainty by determining the differences between sediment provenance predictions obtained using lithogenic radionuclide, geochemical and mineral magnetic signatures when fingerprinting lake and floodplain sedimentary deposits. It also investigated the potential causes of the observed differences. A reservoir core was fingerprinted with the least uncertainty, with tracer group predictions 28 % apart and a consistent down-core trend in changing sediment provenance produced. When fingerprinting an on-line lake core and four floodplain cores, differences between tracer group predictions were as large as 100 %; the down-core trends in changing sediment provenance were also different. The differences between tracer group predictions could be attributed to the organic matter content and particle size of the sediment. There was also evidence of the in-growth of bacterially derived magnetite and chemical dissolution affecting the preservation of tracer signatures. Simple data corrections for sediment organic matter content and particle size did not result in significantly greater agreement between the predictions of the different tracer groups. Likewise, the inclusions of weightings for tracer discriminatory efficiency and within-source variability had minimal effects on the fingerprinting results. This paper highlights the importance of tracer selection and the consideration of recognising tracer non-conservatism when using lake and floodplain sediment deposits to reconstruct anthropogenic changes to the environment and changing sediment dynamics. It was recommended that future research focus on the assessment of uncertainty using the artificial mixing of sediment source samples, the limitation of the fingerprinting to narrow particle size fractions and the development of specific particle size and organic matter correction factors for each tracer.