The Sooty Ant-Tanager (Habia gutturalis) and White-mantled Barbet (Capito hypoleucus) are endangered and endemic birds of Colombia. Both species have small geographic ranges and presumably low population sizes possibly due to habitat destruction and fragmentation. In order to estimate the effects of landscape features on the occupancy of both species, we sampled a variety of landscape configurations within the buffer zones of two hydroelectric impoundments in the Central Andes of Colombia and applied occupancy models to estimate the proportion of area occupied as a function of these covariates. We surveyed 35 point-counts in each hydroelectric impoundment, between June and July of 2014 and 2015. We used single-season models to estimate occupancy while recognizing imperfect detection. Mean occupancy estimates in the study area were similar for both species (0.61 SD=0.33 for the Sooty Ant-Tanager and 0.63 SD=0.25 for the White-mantled). Nonetheless, occupancy probability within the study area was very different between them. The best model for the Sooty Ant-Tanager indicated a decrease in occupancy with elevation, whereas the top model for the White-mantled Barbet indicated an increase in occupancy with distance from streams. Detection probabilities were similar for both species (>0.4) and declined significantly during the second year. Our results provide quantitative guidelines that can be used to evaluate and monitor the state of these populations on the short and long term.

Study Area

The study was conducted in fragmented secondary forests within the buffer zones of two hydroelectric power plants located in the department of Antioquia in the Central Andes of Colombia: one in the municipality of San Carlos (Punchiná), and the other in the municipalities of San Rafael, Alejandría, Santo Domingo and San Roque (San Lorenzo). These protected areas encompass ~80 km2 (ISAGEN 2008) of transitional forest from premontane to tropical wet forests (Holdridge 1947), a key elevation band (580-1420 m) characterized by high levels of fragmentation (Cardona et al. 2011).

Sampling

Our sampling units were point counts defined as circles of 125 m radius, within which birds were registered during 30 minutes. We defined the maximum radius, as the longest distance at which there was a record of any of the two species. Point counts were distributed along transects of ~1 km each. Transects were located near trails and separated by more than 2 km from each other. Each transect consisted of five point-counts spaced at 250 m intervals along a route. A total of 70-point counts (14 transects) were located within the study area, subject to security and transportation restrictions, with the intention to include a broad range of landscape configurations. The landscape features were based on a remotely-sensed Landsat 2013 image. We surveyed 35 sites in the buffer zone of each hydroelectric impoundment between June and July of 2014 and 2015. One of the occupancy modeling assumptions is that detection histories are independent across sites (Mackenzie et al. 2002). In order to quantitatively assess this assumption, we performed spatial correlograms on the residuals of a logistic regression between detections and occupancy covariates.

We conducted all surveys from sunrise (06:00) until 10:00, waiting silently for three minutes before the beginning of observations. For each target species, we registered all individuals seen or heard during a 30-min period, and we recorded each survey with an omnidirectional microphone Sennheiser ME62. Two observers (AMR in 2014 and JSB in 2015) conducted the surveys for two consecutive days. During a survey, we scored the distance to each detection with prior training using a range finder (Nikon 8397 ACULON). We also quantified the number of individuals of each species and if the detection was visual or auditory. The two species are medium sized birds (~20 cm) that are at least seasonally vocal with distinct and easily recognizable songs and calls (Willis 1972, Hilty & Brown 2001, Laverde et al. 2005). In total we conducted 70-point counts with two repetitions during the course of one year.

References

Cardona, N., H. S. David, H. Gómez & P. Roldán. 2011. Flora de embalses, centrales hidroeléctricas de ISAGEN en el oriente antioqueño San Carlos, Jaguas y Calderas. Guía ilustrada. ISAGEN - Universidad de Antioquia, Herbario Universidad de Antioquia, Medellín, Colombia.

Hilty, S.L., & W.L. Brown. 2001. Guía de las aves de Colombia. American Bird Conservancy, Imprelibros S. A., Cali, Colombia.

Isagen S.A. 2008. Centrales de generación Isagen. Https://www.isagen.com.co/comunicados/centrales_generacion.pdf. 

Laverde, O., C. Munera, & L. M. Renjifo. 2005. Preferencia de hábitat por Capito hypoleucus, ave colombiana endémica y amenazada. Ornitología Colombiana 3:62–73.

Willis, E. O. 1972. Taxonomy, ecology, and behavior of the Sooty Ant-Tanager (Habia gutturalis) and other anttanagers (aves). American Museum Novitates no. 2480.

Dataset contains the following files (detailed information about each in the README section:

Detection_histories.csv: Detection histories for Capito hypoleucus and Habia gutturalis. Numbers refer to number of individuals detected. Notice that we use only presence/absence in our manuscript.

Detection_covariates.csv: Detection covariates for both species. 

Occupancy_covariates.csv: Site covariates for both species. 

Point_counts.csv: Geographic coordinates (WGS84) for each site. 

CH_single_season.R: R script to run occupancy analyses for Capito hypoleucus.

HG_single_season.R: R script to run occupancy analyses for Habia gutturalis.