The impact of West Nile virus on the abundance of selected North American birds

Ivo M Foppa,Raphaelle H Beard,Ian H Mendenhall

doi:10.1186/1746-6148-7-43

Abstract

BackgroundThe emergence of West Nile virus (WNV) in North America has been associated with high mortality in the native avifauna and has raised concerns about the long-term impact of WNV on bird populations. Here, we present results from a longitudinal analysis of annual counts of six bird species, using North American Breeding Bird Survey data from ten states (1994 to 2010). We fit overdispersed Poisson models to annual counts. Counts from successive years were linked by an autoregressive process that depended on WNV transmission intensity (annual West Nile neuroinvasive disease reports) and was adjusted by El Niño Southern Oscillation events. These models were fit using a Markov chain Monte Carlo algorithm.ResultsModel fit was mostly excellent, especially for American Crows, for which our models explained between 26% and 81% of the observed variance. The impact of WNV on bird populations was quantitatively evaluated by contrasting hypothetical count trajectories (omission of WNV) with observed counts. Populations of American crows were most consistently affected with a substantial cumulative impact in six of ten states. The largest negative impact, almost 60%, was found in Illinois. A regionally substantial decline was also seen for American Robins and House Sparrows, while the other species appeared unaffected.ConclusionsOur results confirm findings from previous studies that single out American Crows as the species most vulnerable to WNV infection. We discuss strengths and limitations of this and other methods for quantifying the impact of WNV on bird populations.

Highlights

The emergence of West Nile virus (WNV) in North America has been associated with high mortality in the native avifauna and has raised concerns about the long-term impact of WNV on bird populations
High proportions of the variance in the abundance of American Robins was explained by WNV in California (CA), MA, MN and Tennessee (TN)
The most deleterious effect was found in IL (-58%; 95% credible intervals (CIs): -76%, -31%), where in 2010 less than half of the birds that would have been expected in absence of WNV were counted

Summary

Introduction

The emergence of West Nile virus (WNV) in North America has been associated with high mortality in the native avifauna and has raised concerns about the long-term impact of WNV on bird populations. Counts from successive years were linked by an autoregressive process that depended on WNV transmission intensity (annual West Nile neuroinvasive disease reports) and was adjusted by El Niño Southern Oscillation events. These models were fit using a Markov chain Monte Carlo algorithm. Monitored populations of these birds suffered catastrophic mortality due to WNV infection [21,22] These observations spawned concerns about the threat WNV might pose to the North American avifauna. Statistical analysis of count data from well-established national or international bird surveillance programs, such as the North American Breeding Bird Survey (BBS), may offer clues on the impact of WNV on seasonal bird populations. These routes are primarily surveyed in June [23], coinciding with early seasonal WNV transmission

Objectives

Methods

Results

Discussion

Conclusion