Factors Associated with American Kestrel (Falco sparverius) Nest Box Occupancy and Reproductive Success in an Agricultural Landscape
The American Kestrel (Falco sparverius), a small, cavity-nesting falcon that breeds across most of North America, is experiencing an apparent population decline across much of the United States that has yet to be sufficiently explained. With landscape change as a potential factor, the association of landscape composition with kestrel reproduction may assist not only in managing declines but also in improving the effectiveness and efficiency of nest box programs, which can bolster local populations and support monitoring and research efforts. We monitored 163 unique American Kestrel nest boxes over a 12-yr study period (2008–2019) in western Virginia to investigate the factors influencing kestrel reproductive success and we documented relatively high rates of both box occupancy (80%) and nest success (78.9%). In a study area dominated by both high- and low-intensity agriculture, we focused on identifying patterns in land cover composition that may limit reproductive success and associated each box with land cover composition metrics at four spatial scales. Generalized linear mixed models indicated nest boxes with a lower proportion of developed land within 500 m, and a higher proportion of crops and pasture/hay within 3 km were more likely to be occupied by kestrels. Although probability of kestrel nest success increased with reduced proportions of developed land within 100 m, the primary factors influencing success were clutch initiation date and the presence of European starlings (Sturnus vulgaris), with earlier nesting and the absence of starling activity resulting in higher nest success. Overall, cover-type associations with nest success and box occupancy were not strong and establishment of nest boxes in targeted kestrel habitat, while likely maximizing reproductive output, limits research opportunities to identify critical land cover thresholds for the species. Factores Asociados Con La Ocupación de Cajas Nido y el Éxito Reproductivo de Falco sparverius en un Paisaje Agrícola
Falco sparverius, un pequeño halcón que anida en cavidades y que se reproduce en la mayor parte de América del Norte, está experimentando una aparente disminución de la población en gran parte de los Estados Unidos, una situación que aún no ha sido suficientemente explicada. Con el cambio del paisaje como un factor potencial, la asociación entre la composición del paisaje y la reproducción de F. sparverius puede ayudar no sólo a gestionar las disminuciones, sino también a mejorar la eficacia y la eficiencia de los programas de cajas nido, lo que puede reforzar a las poblaciones locales y apoyar los esfuerzos de seguimiento e investigación. Seguimos 163 cajas nido de F. sparverius durante un período de estudio de 12 años (2008–2019) en el oeste de Virginia para investigar los factores que influyen en el éxito reproductivo de la especie, y documentamos tasas relativamente altas tanto de ocupación de cajas (80%) como de éxito de nidos (78.9%). En un área de estudio dominada por agricultura de alta y baja intensidad, nos enfocamos en identificar patrones en la composición de la cobertura del suelo que puedan limitar el éxito reproductivo, y asociamos cada caja con métricas de composición de la cobertura del suelo a cuatro escalas espaciales. Los modelos mixtos lineales generalizados indicaron que las cajas nido con una menor proporción de suelo desarrollado en un radio de 500 m, y con una mayor proporción de cultivos y pastos/heno en un radio de 3 km, tuvieron más probabilidades de estar ocupadas por F. sparverius. Aunque la probabilidad de éxito del nido de F. sparverius aumentó con la presencia de pequeñas proporciones de suelo urbanizado dentro de los 100 m, los factores principales que influyeron en el éxito fueron la fecha de inicio de la nidada y la presencia de Sturnus vulgaris, con una anidación más temprana y la ausencia de actividad de S. vulgaris resultando en un mayor éxito de anidación. En general, las asociaciones de tipo de cobertura con el éxito del nido y la ocupación de las cajas no fueron fuertes, y el establecimiento de cajas nido en el hábitat objetivo de F. sparverius, aunque probablemente maximiza el rendimiento reproductivo, limita las oportunidades de investigación para identificar umbrales críticos de cobertura del suelo para la especie. [Traducción del equipo editorial]
Abstract
Resumen

The Shenandoah Valley Raptor Study Area (dotted line) showing American Kestrel nest box locations (gray circles), major towns, and roadways (gray lines). Forested areas are shown in gray (or green in the online version). Inset shows boundaries of Virginia and surrounding states, and indicates the area represented by main figure.

Coefficients of standardized variables from a generalized linear mixed model (logit link), predicting probability of nest box occupancy by American Kestrels in the Shenandoah Valley Raptor Study Area, Virginia, USA. See Table 1 for variable definitions. Bootstrap 95% confidence intervals are shown as whisker lines from point estimates. Coefficient values provided in Table S2.

Modeled relationship between probability of nest box occupancy by American Kestrels (A) and the proportion of developed land within a 500-m radius, (B) and the proportion of pasture/hay within a 3-km radius, (C) and the proportion of crops within a 3-km radius. Curves are shown both for boxes in their first year of availability, and all other years. Raw data (n = 570) are shown as points, which are artificially dispersed along the y axis to improve visibility. Note: X-axes are at different scales.

Coefficients of standardized variables from a generalized linear mixed model (logit link), predicting probability of nest success by American Kestrels in the Shenandoah Valley Raptor Study Area, Virginia, USA. See Table 1 for variable definitions. Bootstrap 95% confidence intervals are shown as whisker lines from point estimates. Coefficient values provided in Table S2.

Predicted probability of nest success in response to (A) changes in the proportion of developed land within 100 m and (B) clutch initiation date for American Kestrels in nest boxes in the Shenandoah Valley Raptor Study Area. Raw data (n = 361) are shown as points, which are artificially dispersed along the y axis to improve visibility. Depicted curves are for nesting attempts in boxes without starling activity that were not in their first year of installation, and only these raw data points are shown.
Contributor Notes
1 Email address: kolowskij@si.edu
Associate Editor: Christopher J. W. McClure