Natal signatures of juvenile Coris julis in the Azores: investigating connectivity scenarios in an oceanic archipelago

Abstract

To estimate connectivity between populations, we used trace-element composition in otoliths of the temperate wrasse Coris julis as a proxy for the environmental conditions experienced from hatch to settlement. Recruits collected at different sites in the Azores archipelago (northeastern Atlantic) differed significantly in their natal chemical signatures, and were sufficiently diverse to be separated into in 4 distinct natal types. Types 1 and 4 were both low in Mg and Ba, although Mg was dominant in Type 1, and Type 4 was more Ba enriched. Type 2 had high Ba and intermediate Mg concentrations, while Type 3 showed the highest Mg and intermediate Ba concentrations. We identified at least 2 natal types present in recruits at each site, which suggests multiple larval sources per site. Recruits from 3 sites did not differ greatly in their natal type composition, suggesting a common set of sources, but also limiting our ability to describe meaningful connectivity and dispersal scenarios. In contrast, we found that the prevalent Natal Type 4 was absent at 1 site, suggesting that the source characterized by Natal Type 4 does not contribute to the replenishment of C. julis populations at the site. The most striking difference was observed at the smallest scale (10s of kilometers), where we found distinct natal type compositions between north and south shores on the same island, suggesting that these 2 sites have consistent differences in the contribution of specific natal sources. Further, the south shore site had recruits with elevated Mg:Ca concentrations from the otolith core to edge, which suggests that these larvae developed in inshore areas and experienced local retention.KEY WORDS: Otolith chemistry · Otolith core · Fish larvae · Trace element · LA-ICP-MS · Dispersal · Population connectivity Resale or republication not permitted without written consent of the publisherMar Ecol Prog Ser 387: [51][52][53][54][55][56][57][58][59] 2009 2003). Fortunately, in some cases, the varying physical and chemical characteristics of ambient seawater are reflected in the larval otolith (structures in the inner ears of fishes composed of a calcium carbonate/protein matrix) and are preserved in its structure, forming a permanent record of past environmental conditions experienced by the individual (Campana & Neilson 1985, Thorrold et al. 1997. Over the past 2 decades several authors have found evidence that larvae growing up in different water masses are likely to produce otoliths of distinct elemental composition (e.g. Swearer et al. 1999, DiBacco & Levin 2000, Thorrold et al. 2001, Forrester & Swearer 2002, Miller & Shanks 2004. Such variations in trace-element compositions of fish otoliths (elemental fingerprints) have been used both to discriminate commercial fish stocks (Campana et al. 1994, Fowler et al. 2005 and to infer migration pathways and environmental conditions experienced by fishes (Thorrold et al. 1997, Jones et al. 1999, Swearer et al. 1999, Secor et al. 2001, Warner et al. 2005, Standish et al. 2008...

Selecting statistical models and variable combinations for optimal classification using otolith microchemistry

“…When using otolith fingerprints to discriminate water masses, data transformations are needed to satisfy the requirements of the method, but are not always successful and may lead to spurious requirements (Wilson 2007), and those requirements may not be fulfilled after such transformations. Indeed, a brief review of randomly chosen papers among the vast literature dealing with otolith fingerprints analyzed with parametric methods (56 studies, from Kalish 1991to Fontes et al 2009) indicated that normality is a major breakpoint: 35% of the papers did not clearly address this issue, only 5% of the papers had normality of their entire data set (e.g., Chittaro et al 2006), and the remaining studies transformed their data. After transformation, 30% of the papers acknowledged non-normality of the data (e.g., Gillanders et al 2001, de Vries et al 2005).…”

Inferring dispersal and migrations from incomplete geochemical baselines: analysis of population structure using Bayesian infinite mixture models

“…; Fontes et al . ; Shima & Swearer ). The DPM model produces a marginal distribution over the number of sources, the direct probabilistic interpretation of which is more natural than that of arbitrarily scaled model selection criteria such as the AIC or DIC, and allows for estimation of marginal quantities such as P ( K + > S | M ), the probability that there are more sources in the mixed sample than in the baseline S , given the specified model M .…”

“…Nevertheless, the DPM modelling approach has several advantages over existing methods for discovering structure in a recruit pool based on geochemistry. Current methods that use geochemistry to uncover the number of sources in a recruit data set or a mixed fishery use model selection or resampling criteria to produce a single best model (White et al 2008;Fontes et al 2009;Shima & Swearer 2009). The DPM model produces a marginal distribution over the number of sources, the direct probabilistic interpretation of which is more natural than that of arbitrarily scaled model selection criteria such as the AIC or DIC, and allows for estimation of marginal quantities such as P (K + > S|M), the probability that there are more sources in the mixed sample than in the baseline S, given the specified model M. Furthermore, no previous approaches for eliciting the number of sources explicitly incorporate the baseline into the analysisfor the Bayesian clustering model of White et al (2008), the geographical origin of clusters needs to be determined by comparison of cluster means of mixed sample and baseline fish.…”

High Interannual Variability in Connectivity and Genetic Pool of a Temperate Clingfish Matches Oceanographic Transport Predictions

“…[22, 23]), microchemistry of otoliths as geochemical markers (e.g. [24, 25]) and biophysical models, which are used to reconstruct dispersal tracks, population connectivity and to identify potential sink and source populations (e.g. [26, 27]).…”