Use of a submersible spectrophotometer probe to fingerprint spatial suspended sediment sources at catchment scale.

Sediment fingerprinting is used to identify catchment sediment sources. Traditionally, it has been based on the collection and analysis of potential soil sources and target sediment. Differences between soil source properties (i.e., fingerprints) are then used to discriminate between sources, allowing the quantification of the relative source contributions to the target sediment. The traditional approach generally requires substantial resources for sampling and fingerprint analysis, when using conventional laboratory procedures. In pursuit of reducing the resources required, several new fingerprints have been tested and applied. However, despite the lower resource demands for analysis, most recently proposed fingerprints still require resource intensive sampling and laboratory analysis. Against this background, this study describes the use of UV-VIS absorbance spectra for sediment fingerprinting, which can be directly measured by submersible spectrophotometers on water samples in a rapid and non-destructive manner. To test the use of absorbance to estimate spatial source contributions to the target suspended sediment (SS), water samples were collected from a series of confluences during three sampling campaigns in which a confluence-based approach to source fingerprinting was undertaken. Water samples were measured in the laboratory and, after compensation for absorbance influenced by dissolved components and SS concentration, absorbance readings were used in combination with the MixSIAR Bayesian mixing model to quantify spatial source contributions. The contributions were compared with the sediment budget, to evaluate the potential use of absorbance for sediment fingerprinting at catchment scale. Overall deviations between the spatial source contributions using source fingerprinting and sediment budgeting were 18 % for all confluences (n = 11), for all events (n = 3). However, some confluences showed much higher deviations (up to 52 %), indicating the need for careful evaluation of the results using the spectrophotometer probe. Overall, this study shows the potential of using absorbance, directly obtained from grab water samples, for sediment fingerprinting in natural environments.