Site selection for otoliths (hypotheses 3,4)
To determine short-term population structure, and allow comparisons with genetic results, within-lake and coast patterns were evaluated by examining otolith trace element signatures from juveniles returning to streams sampled from a subset of sites used in the genetics study. Otolith trace element signatures provide a measure of potential mixing or separation of populations from a given tributary (Campana, 1999). As returning juveniles are all from the same cohort, and otolith trace element signatures are influenced by a large number of factors (Elsdon & Gillanders, 2005; Grammer et al., 2017; Jaecks, Bond, & Quinn, 2016; Thomas, Ganio, Roberts, & Swearer, 2017), the observation of a similar otolith trace element signature in all juveniles from a given location suggests a common environment and therefore implies recruitment from a single larval pool (Campana, 1999), most likely indicating restricted dispersal from the natal stream. Wide variation in juvenile otolith trace element signatures suggests variation in larval environment and therefore larval recruitment from a diverse range of sources (Gillanders, 2005). Comparison of trace element signatures in juveniles collected from different streams can thus indicate whether larvae developed in one common pool or many (Campana, 1999; Carlson, Fincel, & Graeb, 2017; Gillanders, 2005; Grammer et al. 2017). However, as water chemistry is influenced by geology, regions with similar geologies may be difficult to distinguish. To allow comparison with genetic structuring, the five tributaries of Wanaka and Wakatipu used in the genetic portion of this study were again selected as sampling sites for otolith trace element analyses. Due to logistic and financial constraints, only five coastal streams grouped into three broader regions (Region one: Breccia Creek, Grave Creek; Region two: Thirteen Mile Creek, Kararoa Creek; Region three: Fox River) were selected from the 16 sites used in the genetics portion of this study. These coastal streams were selected to have similar catchment size to the tributaries selected within Wanaka and Wakatipu, and to span as much of the north-south range as possible (Fig. 1).