| I examined the effectiveness of lake order, a metric of landscape position, relative to landscape-level measures of surface connectivity and local factors at predicting the presence and types of fish assemblages and explaining variation in fish species richness in 102 small lakes in three regions of boreal Alberta. Fish presence in these lakes was best predicted in logistic regression models by isolation, a simple metric of surface connectivity. Once fish presence was established, maximum depth was most successful in predicting whether piscivore-dominated assemblages were present. Lake order was the best individual metric at explaining species richness among lakes in Poisson regression models. Thus, lake order, while not the best landscape-level metric for predicting fish assemblages, was best at explaining variation in species richness. These models can be used to predict how anthropogenic disturbance to hydrologic systems will affect fishes in boreal Alberta. |
