Developing Robust Estimates of Population Response to Environmental Change: Moving Towards Integrating Across Scales and Data Types

Natural resources managers generally need information across broad spatial scales to make effective decisions, but generating robust estimates of animal population response to environmental change usually suffers from the all-too-common tradeoff between high quality local and poorer quality, but widespread data. We describe our path towards integrated models by introducing three models for a common stream fish (brook trout) based on data of increasing local complexity; presence/absence, abundance, and individual tags. We use these data types that vary in spatial coverage from 1000s km to 100s km to 1 km to estimate the effects of variation in stream flow and stream temperature on dynamics. We show that it is possible to estimate how environmental variation affects dynamics and that the magnitude and direction of effects are generally similar across models. We then explore the possibility of integrating the models to take full advantage of all data types. The wealth of available data for brook trout combined with independent or integrated models provides the framework for robust predictions and forecasts of the effects of current and future environmental variation on brook trout populations, a key metric for effective management at broad spatial scales.