Spatiotemporal variation in mechanisms driving regional‐scale population dynamics of a Threatened grassland bird

Abstract

To achieve national population targets for migratory birds, landscape‐level conservation approaches are increasingly encouraged. However, knowledge of the mechanisms that drive spatiotemporal patterns in population dynamics are needed to inform scale‐variant policy development. Using hierarchical Bayesian models and variable selection, we determined by which mechanism(s), and to what extent, changes in quantity and quality of surrogate grassland habitats contributed to regional variation in population trends of an obligatory grassland bird, Bobolink (Dolichonyx oryzivorous). We used North American Breeding Bird Survey data to develop spatially explicit models of regional population trends over 25 years across 35 agricultural census divisions in Ontario, Canada. We measured the strength of evidence for effects of land‐use change on population trends over the entire study period and over five subperiods. Over the entire study period, one region (Perth) displayed strong evidence of population decline (95% CI is entirely below 0); four regions displayed strong evidence of population increase (Bruce, Simcoe, Peterborough, and Northumberland). Population trends shifted spatially among subperiods, with more extreme declines later in time (1986–1990: 28% of 35 census divisions, 1991–1995: 46%, 1996–2000: 40%, 2001–2005: 66%, 2006–2010: 82%). Important predictors of spatial patterns in Bobolink population trends over the entire study period were human development and fragmentation. However, factors inferred to drive patterns in population trends were not consistent over space and time. This result underscores that effective threat identification (both spatially and temporally) and implementation of flexible, regionally tailored policies will be critical to realize efficient conservation of Bobolink and similar at‐risk species.

Spatial modeling of Audubon Christmas Bird Counts reveals fine‐scale patterns and drivers of relative abundance trends

“…Understanding the fine‐scale drivers of trends creates opportunities to tailor conservation plans to local conditions (Ethier and Nudds , Ethier et al. ).…”

“…), and limits opportunities for tailoring conservation plans to local conditions (Ethier and Nudds , Ethier et al. ). Abundance, distribution, and population trends of birds are affected by land cover, climate, topography, and other environmental conditions and processes at a variety of spatial scales ranging from local (e.g., home range or smaller) to landscape (Thogmartin et al.…”

Varying dataset resolution alters predictive accuracy of spatially explicit ensemble models for avian species distribution

“…Grassland species in areas half the size of Oklahoma in a single ecoregion have shown spatial and temporal differences in variable importance (Ethier, Koper, & Nudds, 2017). Forest species, which likewise occupy a single habitat type, also show spatial and temporal variation in predictor importance (Zuckerberg et al, 2016).…”

“…Grassland bird species are declining faster than other groups of birds (Askins et al, 2007;Hill, Egan, Stauffer, & Diefenbach, 2014;Knopf, 1994) and continue to be imperiled by ongoing and expanding threats to their habitat. Range-wide species distribution predictions have been made for grassland birds but some species with smaller ranges are not accurately modeled (O'Connor, Jones, Boone, & Lauber, 1999), perhaps because some drivers of distribution vary regionally (Askins et al, 2007;Bakker, Naugle, & Higgins, 2002;Ethier et al, 2017), at a scale smaller than the study region. Additionally, spatial and temporal variation in habitat needs and selection pressures (Davis, 2005;Winter, Johnson, & Shaffer, 2005) or interactions with weather events (Pipher, Curry, & Koper, 2016) are known to be important in grassland birds; therefore, they are particularly suitable as a testing ground for a spatially explicit approach to modeling.…”

“…The open habitats of Oklahoma, which contains over a third of its land area as grasslands and an additional 15% as croplands (Diamond & Elliott, ), contain grassland birds characteristic of habitats ranging from southeastern pine savannahs to tallgrass, mixed‐grass, and shortgrass prairies (Askins et al, ; Diamond & Elliott, ). Grassland species in areas half the size of Oklahoma in a single ecoregion have shown spatial and temporal differences in variable importance (Ethier, Koper, & Nudds, ). Forest species, which likewise occupy a single habitat type, also show spatial and temporal variation in predictor importance (Zuckerberg et al, ).…”