Habitat Associations of Golden Eagle Prey Inferred from Prey Remains at Nesting Sites in Utah, USA
Golden Eagles (Aquila chrysaetos) can prey on a wide variety of species, but population persistence is often thought to depend on the abundance of a few key prey species. We investigated Golden Eagle prey remains at 254 nesting sites in north-central Utah, USA, from 1970–2014. We hypothesized that variation in observed prey at the nesting site could be predicted by ecoregion or localized (6.4-km radius) environmental factors. We identified 147 prey species representing a minimum of 26,734 individuals, with the majority of species occurring at low frequencies. Golden Eagle prey remains were dominated by black-tailed jackrabbits (Lepus californicus), with cottontails (Sylvilagus spp.), rock squirrels (Otospermophilus variegatus), and yellow-bellied marmots (Marmota flaviventris) also found frequently, and occasionally in large numbers per nesting site. We found natural groupings of prey species by multivariate analyses. Nonmetric multidimensional scaling (NMDS) identified three prey assemblages typical of sagebrush (Artemisia spp.)steppe, wetland, and mountain ecosystems. Canonical correspondence analysis (CCA) and permutational multiple analysis of variance (PERMANOVA) suggested that prey assemblages were associated with environmental variables, including: (1) forest cover and elevation vs. sagebrush and pinyon pine (Pinus spp.) cover; and (2) alfalfa (Medicago sativa), crop, and wetland cover vs. elevation and forest, sagebrush, and pinyon pine cover. Observed prey were better predicted by measured environmental factors than biogeographic boundaries. The abundance of the four most frequently recorded prey species was influenced primarily by habitat, and to a lesser degree by overall diversity of prey remains, precipitation, and time trend variables, as suggested by Poisson regression models. Our analyses indicate that Golden Eagle prey varied within and between ecoregion boundaries, and that prey were more strongly predicted by localized environmental factors than by climate or time. Asociaciones de Hábitat de las Presas de Aquila chrysaetos Inferidas a Partir de Restos de Presas en Lugares de Cría en Utah, Eeuu
Aquila chrysaetos puede obtener presas pertenecientes a una amplia variedad de especies, pero usualmente se piensa que la persistencia poblacional depende de la abundancia de unas pocas especies de presas clave. Investigamos los restos de presas de A. chrysaetos en 254 lugares de cría en el norte-centro de Utah, EEUU, desde 1970 hasta 2014. Hipotetizamos que la variación en las presas observadas en el lugar de cría podría predecirse por ecorregión o por factores ambientales localizados (6.4 km de radio). Identificamos 147 especies de presas representando un mínimo de 26,769 individuos, con la mayoría de las especies presentes en bajas frecuencias. Los restos de presas de A. chrysaetos estuvieron dominados por Lepus californicus, pero también hallamos con frecuencia restos de Sylvilagus spp., Otospermophilus variegatus y Marmota flaviventris, que ocasionalmente se presentaron en grandes cantidades en los lugares de cría. Encontramos agrupaciones naturales de especies de presas mediante análisis multivariados. El escalamiento multidimensional no métrico identificó tres ensambles de presas típicos de los ecosistemas de estepa de Artemisia spp., de humedal y de montaña. Los análisis de correspondencia canónica y los análisis de varianza múltiple con permutaciones sugirieron que los ensambles de presas estuvieron asociados con variables ambientales, incluyendo: (1) cobertura del bosque y altitud vs. cobertura de Artemisia y Pinus spp.; y (2) cobertura de Medicago sativa, cultivo y humedal vs. altura y cobertura de bosque, Artemisia y Pinus spp. Las presas observadas fueron mejor predichas por los factores ambientales medidos que por los límites biogeográficos. La abundancia de las cuatro especies de presas más frecuentes estuvo influenciada principalmente por el hábitat y en menor medida por la diversidad global de restos de presas, la precipitación y las variables de tendencia de tiempo, como sugirieron los modelos de regresión de Poisson. Nuestros análisis indican que las presas de A. chrysaetos variaron dentro y entre los límites de las ecorregiones, y que las presas fueron predichas con mayor robustez por factores ambientales localizados que por el clima o el tiempo. [Traducción del equipo editorial]ABSTRACT
RESUMEN
Golden Eagle nesting sites (n = 254) in the long-term study area in Utah, USA, with CEC Level III Ecoregions (CEC 1997).
The distribution of environmental variables of land cover in a 6.4-km neighborhood surrounding Golden Eagle nesting sites (n = 254) by ecoregion in Utah, USA. Excludes data from three nesting sites in the Wyoming Basin ecoregion.
First two axes of nonmetric multidimensional scaling (NMDS) ordination of prey remains collected from Golden Eagle nesting sites (dissimilarities calculated using Anderson's 2006 modification of the Gower index) in Utah, USA, with gray circles indicating nesting sites in multivariate space and (A) notable prey species labeled, or (B) environmental gradients overlaid, but not informing the ordination. Environmental covariates are proportion of land cover in a 6.4-km radius around each eagle nesting site (centered to 0 and standardized to unit variance).
Relationships between ecoregion (A) or environmental gradients (B) and prey remains assemblages at Golden Eagle nesting sites in Utah, USA, as illustrated by canonical correspondence analysis (CCA). Eagle nesting sites in multivariate space shown as gray open circles. (A) Centroids of ecoregions (CBR = Central Basin and Range, WUM = Wasatch and Uinta Mountains, CP = Colorado Plateaus) in black. (B) Environmental gradient vectors illustrated as proportionally scaled black arrows. Environmental covariates are proportion of land cover in a 6.4-km radius around each eagle nesting site (centered to 0 and standardized to unit variance).
Comparison of standardized beta parameter estimates (means ± 95 Bayesian credible intervals) for Poisson models explaining the abundance of black-tailed jackrabbits (Jackrabbits), cottontail spp. (Cottontails), rock squirrels (Rock squirrels), and yellow-bellied marmots (Marmots) in prey remains recovered from Golden Eagle nesting sites in Utah, USA, from 1970–2014. Parameter estimates are shown only if that variable is included in the top model for that species. “Crop,” “grass,” “sage,” and “forest” refer to the proportion of land cover dominated by that vegetation type within 6.4 km of the nesting site, “diversity” is the Hill number estimated to describe the overall diversity of species found within the nesting site's prey remains (larger numbers indicate higher diversity), “jackrabbits” is the minimum number of black-tailed jackrabbit remains found at each nest, “year” is the overall time trend, and “dry year,” “wet year,” “previous year dry,” and “previous year wet” all indicate whether the annual precipitation averaged across weather stations within the study area was outside the 0.985 CI of the mean for that year or the previous one.
Annual frequencies of black-tailed jackrabbit (circles) and cottontails (crosses) found in the prey remains (mean of minimum number of individuals) in Golden Eagle nesting sites in Utah, USA.
Contributor Notes
1 Email address: jessilbrown@gmail.com