Insights into bulk stable isotope alteration during sediment redistribution to edge-of-field: impact on sediment source apportionment

Organic tracers (bulk δ13C and δ15N) are being increasingly used in sediment/particulate organic carbon source apportionment studies at the catchment scale to support sustainable land management decisions. Here, the use of these isotopic tracers in sediment fingerprinting depends on the critical assumption that δ13C and δ15N values remain conservative during the sediment delivery continuum. Such assumption, however, requires critical evaluation, especially since standard tracer conservation tests applied in conjunction with catchment scale studies are prone to masking potential issues at smaller spatial scales including the field scale. Against this background, our study evaluated the subtle isotopic shift associated with sediment redistribution to edge-of-field and assessed the impact on sediment source apportionment. In a C3-grass dominated study field, δ13C and δ15N values exhibited differences (i.e., Δ13C = 1.4 ± 0.7‰ and Δ15N = 0.4 ± 0.4‰) between soil depths of 0–5 cm and 5–10 cm. Sampled sediments at the edge-of-field flume had higher δ13C values than the 0–5 cm soil layer; i.e., Δ13C values were 0.3 ± 0.4‰ and 0.8 ± 0.4‰ for suspended and deposited (materials deposited in a basket downstream of a flume) sediment, respectively. In contrast, δ15N values increased in suspended (Δ15N = 0. 8 ± 0.6‰) but decreased (Δ15N = − 0.4 ± 0.5‰) in deposited sediment, compared to the 0–5 cm soil layer, suggesting that the N isotopes can respond differently in edge-of-field sediment types. Although current fingerprinting work tends to not take explicit account of organic tracer alteration during transport or after deposition, our results demonstrate that correcting sediment δ13C and δ 15N values for tracer alteration in a Bayesian un-mixing model generated robust and reliable estimates of source contributions to both target sediment types. We therefore recommend taking account of the subtle but consistently altered δ13C and δ 15N values along the sediment cascade in un-mixing modelling to help better discriminate sources and to improve un-mixing model estimates at the catchment scale.