Mixed-stock and discriminant models use for assessing recruitment sources of estuarine fish populations in La Plata Basin (South America)

The objective of this study was to identify potential recruitment sources of Prochilodus lineatus from freshwater areas (Parana and Uruguay rivers) to estuarine population of the Rio de la Plata Estuary (La Plata Basin, South America), considering young (age-1) and adult (age-7) fish. LA-ICP-MS chemical analysis of the otolith core (nine element:Ca ratios) of an unknown mixed sample from Rio de la Plata Estuary (2011 and 2017) was compared with a young-of-year baseline data set (same cohort) and classified into freshwater nurseries (Parana or Uruguay river) by using maximum classification-likelihood models (MLE and MCL) and quadratic discriminant analysis (QDA). Considering the three models used, the Uruguay River was the most important contributor for both young and adult populations. The young population (2011) was highly mixed with contributions between 31.7 to 68.3%, while the degree of mixing was found to decrease in 2017 (adult fish) from 97.1 to 100% contributions. The three employed methods showed comparable estimates, however, the QDA showed a high similarity with the MCL model, suggesting sensitivity to evaluate small contributions, unlike the MLE method. Our results show the potential application of maximum likelihood mixture models and QDA for determining the relative importance of recruitment sources of fish in estuarine waters of the La Plata Basin.