Assessing n-alkane and neutral lipid biomarkers as tracers for land-use specific sediment sources

Abstract

Sediment fingerprinting (SF) methods using taxonomic-specific biomarkers such as n-alkanes have been successfully used to distinguish sediment sources originating from different land uses at a catchment scale. In this study, we hypothesise that using a combination of soil biomarkers of plant, fungal and bacterial origin may allow greater discrimination between land uses in SF studies. Furthermore, we assess if the inclusion of short chain (shorter than C22) neutral lipid fatty acids (SC-NLFA) improves land use discrimination, considering the Loch Davan catchment (34 km2) in Scotland as a case study. Fatty acids are commonly used to measure abundance and diversity of soil microbial and fungal communities. The spatial distribution of these soil communities has been shown to depend mainly on soil properties and, therefore, soil types and land management practices. The n-alkane and SC-NLFA concentrations and their compound specific stable isotope signatures (CSSI) in four land cover classes (crop land, pasture, forest, and moorland) were determined and their contribution to six virtual sediment mixture samples was modelled. Using a Bayesian un-mixing model, the performance of the combined n-alkane and SC-NLFA biomarkers in distinguishing sediment sources was assessed. The collection of new empirical data and novel combinations of biomarkers in this study found that land use can be distinguished more accurately in organic sediment fingerprinting when combining n-alkanes and SC-NLFA or using SC-NLFA and their CSSI alone. These results suggest that fingerprinting methods using the output of unmixing models could be improved by the use of multiple tracer sets if there is a commensurate way to determine which tracer set provides the “best” capacity for land use source discrimination. This new contribution to the organic sediment fingerprinting field highlights that different combinations of biomarkers may be required to optimise discrimination between soils from certain land use sources (e.g., arable-pasture). The use of virtual mixtures, as presented in this study, provides a method to determine if addition or removal of tracers can improve relative error in source discrimination. Combining biomarkers from different soil communities could have a significant impact on the identification of recent sources of sediment within catchments and therefore on the development of effective management strategies.