Selecting fingerprint properties for discriminating potential suspended sediment sources in river basins

The absence of guidelines for pre-selecting the most effective combination of properties for inclusion in a composite fingerprint represents a key methodological uncertainty hampering the wider adoption of the fingerprinting approach for identifying suspended sediment sources. This contribution reports a preliminary attempt to address this issue by testing the discrimination of potential sediment sources within a number of contrasting river basins in the UK and Africa, afforded by a range of sediment properties. Statistical analysis confirms that there is no single diagnostic property capable of discriminating the range of potential suspended sediment sources in the study basins. The use of composite fingerprints based on several constituents drawn from a single group of properties consistently improves sediment source discrimination. However, the level of discrimination afforded by a particular combination of properties is not consistent between the study catchments. Composite fingerprints incorporating constituents selected from several groups of properties using a stepwise statistical selection procedure consistently provide the most robust discrimination of potential sediment sources. Whilst it is not possible to identify a universally applicable optimum composite fingerprint, the results suggest that at present, the most effective means of optimising sediment source discrimination is to identify a number of groups of properties for subsequent use in the establishment of a composite fingerprint. More specifically, this study suggests that measurements of a combination of acid and pyrophosphate-dithionite extractable metals, base cations and organic constituents should provide an effective basis for establishing composite fingerprints for discriminating individual sediment source types. Radiometric properties can also provide useful information for improving sediment source discrimination. The final choice of the groups of properties will necessarily reflect the laboratory facilities available.