Relationships between salmon abundance and tree-ring δ15N: three objective tests

Abstract

Quantification of a relationship between salmon escapement in rivers and riparian tree-ring δ15N could allow reconstruction of prehistorical salmon abundance. Unfortunately, attempts to quantify this link have met with little success. We examined the feasibility of the approach using natural abundance of δ15N in riparian tree rings formed before and after extirpation of salmon and 15N tracer studies in a river and riparian soils. We concluded that (i) extractable (sap) N must be removed for interpretation of tree-ring δ15N because it contains up to 78% of the N in wood, is mobile, and differs from structural N in isotopic composition, (ii) no significant change in structural tree-ring δ15N was associated with salmon extirpation in a natural system, (iii) 500‰ 15NH4+ added to a stream was detected in riparian tree rings spanning at least 8 years, demonstrating interring movement of N that confounds detection of an annual signal, and (iv) addition of 28 000‰ 15NH4+ to riparian soils at a rate equaling 7.25 kg salmon·50 m–2 resulted in maximum tree-ring δ15N of ∼100‰–600‰. Thus, the calculated maximum signal possible from salmon was 0.08‰–0.43‰, which is within the range of natural variation. Evidence suggested that neither total nor structural tree-ring δ15N was useful for reconstructing salmon abundance.