2021 ESA Annual Meeting (August 2 - 6)

A century of avian diversity change in North America

On Demand
Stephen J. Murphy, Evolution, Ecology, and Organismal Biology, Ohio State University;
Background/Question/Methods

Recent evidence suggests that multiple avian populations have experienced precipitous declines in abundance and spatial extent across North America and other parts of the globe. These findings have led researchers to seek answers regarding the underlying drivers of these changes. However, to date, few studies have examined the community-level consequences of such population changes over a similarly long timeframe. Using data from the Audubon Christmas Bird Count Survey, a century-old citizen science data effort that monitors avian population abundance in North America, we examine how community-level metrics (i.e., species richness, Shannon-diversity, evenness, and functional richness) have changed over the past ~100 years. Individual species were modelled using three approaches: 1) a naïve approach based on raw effort-corrected counts; 2) N-mixture model with a 2-year closure assumption; and 3) Dail-Madsen model with a non-robust sampling design. Each of these approaches has its own strengths and weaknesses, and we sought to determine how dependent our overall inferences regarding diversity change were on the modelling approach that was used. In each case, we incorporated relevant covariates relating to population abundance and species detection probability, including sampling effort, minimum temperature, and availability of suitable habitat.

Results/Conclusions

We observed notable trends in all facets of biodiversity examined for each of the three analytical methods used. As expected, changes in taxonomic diversity, especially species richness, showed the least amount of change, and changes that did exist were driven primarily by trends in sampling effort observed across years. In contrast, abundance-weighted and functional diversity metrics showed more significant changes. Interestingly, these trends were highly variable across space, with many circles experiencing declines in functional diversity over time, while others showed notable increases. We did not observe any clear link in these trends to changes in minimum temperature or suitable habitat availability. Given these findings, future efforts will be needed to better understand the patterns and underlying drivers of variation in diversity change in avian communities across North America. Admittedly, this will remain challenging over the long timescales analyzed in this study given important limitations that are present in most historical datasets. Specifically, the data used here make it challenging to accurately estimate detection probabilities, and thus true population abundance. Still, the concordance we observed across analytical methods suggest that these trends are real, and thus important for understanding avian communities over the past 100 years.