The identification of disturbance thresholds is important for many aspects of aquatic resource management, including the establishment of regulatory criteria and the identification of stream reference conditions. A number of quantitative or model-based approaches can be used to identify disturbance thresholds, including nonparametric deviance reduction (NDR), piecewise regression (PR), Bayesian changepoint (BCP), quantile piecewise constant (QPC), and quantile piecewise linear (QPL) approaches. These methods differ in their assumptions regarding the nature of the disturbance-response variable relationship, which can make selecting among the approaches difficult for those unfamiliar with the methods. We first provide an overview of each of the aforementioned approaches for identifying disturbance thresholds, including the types of data for which the approaches are intended. We then compare threshold estimates from each of these approaches to evaluate their robustness using both simulated and empirical datasets. We found that most of the approaches were accurate in estimating thresholds for datasets with drastic changes in responses variable at the disturbance threshold. Conversely, only the PR and QPL approaches performed well for datasets with conditional mean or upper boundary changes in response variables at the disturbance threshold. The most robust threshold identification approach appeared to be the QPL approach; this method provided relatively accurate threshold estimates for most of the evaluated datasets. Because accuracy of disturbance threshold estimates can be affected by a number of factors, we recommend that several steps be followed when attempting to identify disturbance thresholds. These steps include plotting and visually inspecting the disturbance-response data, hypothesizing what mechanisms likely generate the observed pattern in the disturbance-response data, and plotting the estimated threshold in relation to the disturbance-response data to ensure the appropriateness of the threshold estimate.

Models for making preseason forecasts of adult abundance are an important component of the management of many stocks of Pacific salmon Oncorhynchus spp. Reliable forecasts could increase both the profits from fisheries and the probability of achieving conservation and other management targets. However, the predictive performance of salmon forecasting models is generally poor, in part because of the high variability in salmon survival rates. To improve the accuracy of forecasts, we retrospectively evaluated the performance of eight preseason forecasting models for 43 stocks of pink salmon O. gorbuscha over a total of 783 stock-years. The results indicate that no single forecasting model was consistently the most accurate. Nevertheless, across the 43 stocks we found that two naïve time series models (i.e., those without explicitly modeled mechanisms) most frequently performed best based on mean raw error, mean absolute error, mean percent error, and root mean square error for forecasts of total adult recruits. In many cases, though, the best-performing model depended on the stock and performance measure used for ranking. In 21% of the stocks, a new multistock, mixed-effects stock-recruitment model that included earlysummer sea surface temperature as an independent variable along with spawner abundance demonstrated the best performance based on root mean square error. The best-performing model for each pink salmon stock explained on average only 20% of the observed variation in recruitment. Owing to the uncertainty in forecasts, a strong precautionary approach should be taken to achieve conservation and management targets for pink salmon on the West Coast of North America.

To improve the understanding of effects of environmental factors on spawner-to-recruit survival rates of pink salmon (Oncorhynchus gorbuscha), we developed several spatial hierarchical Bayesian models (HBMs). We applied these models to 43 pink salmon stocks in the Northeast Pacific. By using a distance-based, spatially correlated prior distribution for stock-specific parameters, these multistock models explicitly allowed for positive correlation among nearby salmon stocks in their productivities and coefficients of early summer coastal sea surface temperature (SST). To our knowledge, this is the first time that such distance-based, spatial prior probability distributions for parameters have been applied to fisheries problems. We found that the spatial HBMs produce more consistent and precise estimates of effects of SST on productivity than a single-stock approach that estimated parameters for each stock separately. Similar to earlier results using mixed-effects models for the same stocks, we found significant positive effects of SST on survival rates of northern pink salmon stocks, but weaker negative effects of SST on survival rates of southern pink salmon stocks. However, we show a smoother transition in magnitude of effects between these regions.

In this paper, we present an improved methodology for estimating salmon escapements from stream count data. The new method uses a hierarchical Bayesian model that improves estimates in years when data are sparse by "borrowing strength" from counts in other years. We present a model of escapement and of count data, a hierarchical Bayesian statistical framework, a Gibbs sampling approach for evaluation of the posterior distributions of the quantities of interest, and criteria for determining when the model and inference are adequate. We then apply the hierarchical Bayesian model to estimating historical escapement and escapement timing for pink salmon (Oncorhynchus gorbuscha) returns to Kadashan Creek in Southeast Alaska.Résumé : On trouvera ici une méthodologie améliorée pour estimer les échappées de saumon à partir des données de dénombrement dans un cours d'eau. La nouvelle méthode utilise un modèle hiérarchique bayésien (HBM) qui améliore les estimations pour les années où les données sont rares en " empruntant de la puissance " aux dénombrements des autres années. Nous présentons un modèle pour les données d'échappées et de dénombrement, un cadre statistique hiérarchique bayésien, une stratégie d'échantillonnage de type Gibbs pour l'évaluation des distributions a posteriori des valeurs recherchées et des critères pour déterminer quand le modèle et les inférences sont adéquats. Nous appliquons le modèle HBM à l'estimation des échappées des années antérieures et du moment des retours des Saumons roses à Kadashan Creek dans le sud-est de l'Alaska.[Traduit par la Rédaction] Su et al. 1662