Evaluation of Bioenergetics Models for Predicting Great Blue Heron Consumption of Rainbow Trout at Hatcheries

Abstract

We predicted consumption of rainbow trout Oncorhynchus mykiss by great blue herons Ardea herodias by combining a population estimate of predators with bioenergetics models of (1) existence metabolism (ME), (2) existence metabolism plus reproductive costs (ME_Breeding), and (3) field metabolic rate (FM) for individual great blue herons. We then compared predicted consumption to directly estimated consumption by a free-ranging population of great blue herons foraging at a fish hatchery. Although the FM model predicted annual population consumption rates similar to observed annual consumption, model predictions differed significantly from directly estimated consumption for 6 months. Monthly consumption predicted by the existence metabolism models agreed with directly estimated consumption from October to June. During the breeding season (May–July), directly estimated consumption was higher than predicted by the ME model for nonbreeders but was lower than predicted by the ME_Breeding model. This result would be expected if 20–30% of the great blue herons were reproductive. Peak population consumption was observed in August but was not predicted by the models. This discrepancy probably resulted from an influx of migrant great blue herons that were not detected during observations of population-level estimates. Model predictions would be improved most by using more reliable estimates of great blue heron abundance and by explicitly measuring the proportion of breeding birds in the population. Given accurate predator population estimates, bioenergetics models provide reliable, cost-effective estimates of fish losses to piscivorous birds.