Editorial Type:
Article Category: Research Article
 | 
Online Publication Date: 04 Feb 2021

Environmental Factors Shape the Nonbreeding Distribution of the Harlan's Red-Tailed Hawk: A Maximum Entropy Approach

,
,
, and
Page Range: 79 – 92
DOI: 10.3356/0892-1016-55.1.79
Save
Download PDF

ABSTRACT

Several lines of evidence support the hypothesis that abiotic and biotic factors directly or indirectly control species' distributions. Despite the importance of assessing the environmental factors governing species' distributions, it is not clear how such factors influence migratory species. We evaluated environmental factors related to the nonbreeding distribution (September through March) for Harlan's Red-tailed Hawks (Buteo jamaicensis harlani), using a maximum entropy (Maxent) approach. We evaluated records spanning the nonbreeding season from 1 September to 31 March, gathered from four different sources including online data and published literature; we used maximum and minimum temperatures, precipitation, solar radiation, wind speed, and cloud cover as abiotic factors. Based on the most influential environmental factors (determined using percent contribution, a measure representing the increase in likelihood associated with each environmental factor), we delineated the current winter distribution (December through February) and evaluated niche overlap. From December through February, we found that minimum temperature showed the highest percent contribution (62.2%, 55.9%, and 66.9% in each month, respectively). Response curves for minimum temperature were quadratic in shape in September, non-monotonic in shape in October, and sigmoid in shape in November, December, January, and February, with high suitability values for temperature ranging from 0–20°C and low suitability for temperature,0°C. However, habitat suitability decreased at higher levels of solar radiation (.6000 kJ m–2 d–1) during fall migration (September, October) and spring migration (March). Our revised winter distribution map, based on ecological niche modeling, includes areas of ecoregions not previously included in Harlan's Hawk range maps in the western USA, and northeastern Mexico.

RESUMEN

Los Factores Ambientales Moldean la Distribución no Reproductiva de Buteo jamaicensis harlani: Un Enfoque de Máxima Entropía

Varias líneas de evidencia apoyan la hipótesis de que factores bióticos y abióticos controlan directa o indirectamente la distribución de las especies. A pesar de la importancia de evaluar los factores ambientales que gobiernan la distribución de las especies, no es claro cómo estos factores influyen sobre las especies migratorias. Evaluamos los factores ambientales relacionados con la distribución no reproductiva (septiembre a marzo) de Buteo jamaicensis harlani usando un enfoque de máxima entropía (Maxent). Evaluamos registros que abarcan la estación no reproductiva desde el 1 de septiembre al 31 de marzo, reunidos a partir de cuatro fuentes diferentes, incluyendo datos en línea y literatura publicada; usamos temperaturas máximas y mínimas, precipitación, radiación solar, velocidad del viento y cobertura de nubes como factores abióticos. Basados en los factores ambientales más influyentes (determinados usando el porcentaje de contribución, una medida que representa el aumento en probabilidad asociado con cada factor ambiental), delineamos la distribución invernal actual (diciembre a febrero) y evaluamos la superposición de nicho. De diciembre a febrero, encontramos que la temperatura mínima mostró el mayor porcentaje de contribución (62.2%, 55.9% y 66.9% en cada mes, respectivamente). Las curvas de respuesta para la temperatura mínima tuvieron forma cuadrática en septiembre, forma no monotónica en octubre y forma sigmoidea en noviembre, diciembre, enero y febrero, con altos valores de idoneidad para temperaturas que fluctuaron entre 0–20°C y baja idoneidad para temperaturas,0°C. Sin embargo, el hábitat idóneo disminuyó a altos niveles de radiación solar (.6000 kJ m–2 d–1) durante la migración de otoño (septiembre, octubre) y la migración de primavera (marzo). Nuestro mapa de distribución invernal revisado, basado en modelos de nicho ecológico, incluye áreas de ecorregiones previamente no incluidas en los mapas de distribución de B. j. harlani en el oeste de EEUU y noreste de México.

[Traducción del equipo editorial]

Copyright: © 2021 The Raptor Research Foundation, Inc. 2021
pdf
Figure 1.
Figure 1.

Recorded winter distribution (polygon in diagonal hatch; generally depicted from December through February) in the South Central Plains ecoregion (Wood 1932, Mindell 1983, 1985), and the hypothetical vagrant nonbreeding distribution (polygon in cross hatch) of Harlan's Hawks as suggested by Howell and Webb (1995). We created the vagrant nonbreeding distribution with the alpha-hull method (Burgman and Fox 2003). The dotted lined shows the calibration area (“M”) for ecological niche models. White points indicate records used for calibration (“final data set”, n ¼ 1379 records) during overall nonbreeding season.


Figure 2.
Figure 2.

Percent contribution (based on the Maxent output) of environmental variables for each best-supported monthly model (including the Average model) of Harlan's Hawk nonbreeding season distribution. We selected best-supported models based on DAICc.


Figure 3.
Figure 3.

Predicted suitability and potential geographic distribution of Harlan's Hawks derived from Maxent ecological niche models from the nonbreeding season. Higher values and warmer (red) colors indicate areas with higher predicted suitability. Black points indicate the total of occurrences used for developing models (i.e., training), whereas gray points show the total of occurrences used to evaluate distribution models. Please refer to the online version of this article for interpretation of the references to color.


Figure 4.
Figure 4.

Binary maps showing the predicted potential geographic distribution for Harlan's Hawks during winter. Values of 1 (low probability of occurrence) and 2 (moderate probability of occurrence) indicate areas with Harlan's Hawks present, and the value of 3 represents the spatial coincidence among all winter months, which was considered as the winter distribution derived from our ecological niche models.


Figure 5.
Figure 5.

Niche similarities between models of Harlan's Hawk nonbreeding season distribution based on the Schoener's D (lower triangular). Higher indexes (intensity of gray) indicate more similarity.


Contributor Notes

Associate Editor: David E. Andersen

Received: 13 Jul 2019
Accepted: 08 Sept 2020
  • Download PDF