Spatio-temporal and interspecific variation in otolith trace-elemental fingerprints in a temperate estuarine fish assemblage

We tested whether estuarine fishes have site-specific differences in the concentrations of trace elements in their otoliths that can be used as \u27fingerprints\u27 to identify them to their estuary of origin. To evaluate the robustness of this approach, we tested whether elemental fingerprints were consistent among individuals of five species that were collected in 1996 from three temperate estuaries in southern California. We also tested whether elemental fingerprints were consistent between spring and autumn 1996 for three species in one of the sites, Carpinteria Marsh. The species evaluated comprised a mid-water-dwelling smelt (Atherinops affinis), two benthic gobies (Clevelandia ios and Ilypnus gilberti), and two flatfish (Paralichthys californicus and Hypsopsetta guttulata). The concentrations of six elements (Mn, Cu, Zn, Sr, Ba, and Pb) were determined in the otoliths using inductively coupled plasma-mass spectrometry (ICP-MS). Within estuaries, the five species exhibited strong variation in elemental concentration, indicating substantial interspecific differences in otolith environmental history. When the five fish species were considered separately, multivariate (MANOVA) and univariate (ANOVA) analyses of variance indicated that the elemental composition of otoliths differed significantly among the estuaries in four of the five species. Based on linear discriminant function analyses (DFA), differences were strong enough that trace element composition could be used to accurately assign fish to their site of origin [mean (range): 93.5% (74-100%)]. However, elemental signatures within Carpinteria Marsh were not consistent between spring and autumn 1996, and this was reflected in a substantial reduction in the accuracy of assigning fish to their true site of origin. When we compared site differences between fish species (site x species interactions), the elemental fingerprints were most similar between closely related species (e.g. the two gobies and the two flatfish) and most dissimilar between distantly related species, both phylogenetically and ecologically. Among the six elements analyzed, Sr and Ba exhibited the most inconsistent pattern among species, with significant differences in 80 and 70% of the pairwise species comparisons, respectively. The remaining four elements showed ≥70% consistency in the pattern of variation among sites for the different species. Thus, while otolith elemental fingerprinting can be a useful tool for inferring estuarine residency, such fingerprints may be temporally variable and species specific. © 2003 Elsevier Science B.V. All rights reserved