Incorporating location uncertainty improves inference with stop-level North American Breeding Bird Survey data

Abstract/Summary

Covariate uncertainty is a challenge for monitoring ecological data. Here, we test a Bayesian method for accounting for covariate uncertainty in multi-species models using informative priors. We conduct a series of simulations to evaluate the effects of incorporating covariate uncertainty into models, using the North American Breeding Bird Survey (BBS) dataset as a case study. This extensive database has annual bird point count data for over 100,000 locations annually, but the precise spatial coordinates of these locations are unknown. We find that an informative prior model produces substantially better inferences than does a simpler model that makes assumptions about locations and covariates.

Publication details