Activity patterns of Tayra Eira barbara populations from Costa Rica and Colombia: evidence of seasonal effects
 
Patrones de actividad de poblaciones de la Tayra Eira barbara en Costa Rica y Colombia: evidencia de efectos estacionales

José F. González-Maya1, Diego Zárrate-Charry1,2, I. Mauricio Vela-Vargas1, Juan S. Jiménez-Alvarado1, Diego Gómez-Hoyos1

1 Proyecto de Conservación de Aguas y Tierras, ProCAT Colombia/Internacional. Bogotá, Colombia.  e-mail: jfgonzalezmaya@gmail.com
2 Oregon State University, Corvallis, Oregón, Estados Unidos.
    Date received: October 16, 2014                Date approval: February 19, 2015               Editor asociado: Jiménez AM.

Abstract

Objective: The objective of this study was to evaluate the activity patterns of Tayra Eira barbara and the effect of seasonality in two different landscapes, one in Colombia (Canal del Dique Watershed) and another in Costa Rica (Talamanca). Methodology: Between 2005 and 2009 camera trapping samplings were conducted to determine activity patterns of the species. We estimated capture frequencies, activity patterns, and made comparisons between sites and seasons, classifying data in 24-hours and in time-periods: Diurnal, Nocturnal and Crepuscular. Comparisons were made using Kuiper tests for homogeneity, Chi-square for differences in patterns and Chao-Jaccard tests for testing overlapping activities. Results: Differences between activity periods (X2=124.26, p<0.0001) and hours (X2=46.38, p=0.0001) were found, with three activity peaks during the day, being this period in which the species is more active in general (X2=124.26, p<0.0001). Seasonality seems to be the most important factor influencing activity patterns (X2=14.15, p=0.0002), especially influencing the hours of higher activity (X2=12.42, p=0.0033). Conclusions: Activity patterns in Tayras seem to slightly vary both in space and time, with seasonality been the most evident driver. Our results coincide with previous studies in similar habitats, but differ with other locations with different ecosystems, highlighting variation within the species and potentially related with different subspecies. The evaluation and identification of basic aspects of natural history of E. barbara are necessary for a better understanding of the species ecology, life history, habitat requirements and their role in ecosystems.

Keywords: Canal del Dique, Capture frequency, Carnivores, Circadian Rythms, Talamanca.

Resumen

Objetivo: El objetivos del presente estudio fue evaluar los patrones de actividad de la Tayra Eira barbara y el efecto de la estacionalidad en dos paisajes diferentes, uno en Colombia (Canal del Dique) y uno en Costa Rica (Talamanca). Metodología: Entre el 2005 y el 2009 se realizaron muestreos con cámaras-trampa para determinar las horas de actividad de la especie. Estimamos frecuencias de captura y patrones de actividad y estos fueron comparados entre sitios y estaciones, clasificando los datos en las 24 horas y según periodos: Diurno, Nocturno y Crepuscular. Las comparaciones se realizaron mediante pruebas de Kuiper de homogeneidad, Chi-cuadrado para las diferencias en los patrones y pruebas Chao-Jaccard para evaluar la sobreposición. Resultados: Se encontraron diferencias entre los períodos de actividad (X2=124.26, p<0.0001) y las horas (X2=46.38, p=0.0001), mostrando tres picos de actividad durante el día, siendo este periodo durante el cual la especie está más activa en general (X2=124.26, p<0.0001). La estacionalidad parece ser el factor más importante afectando los patrones de actividad de la especie (X2=14.15, p=0.0002), especialmente influenciando las horas de mayor actividad (X2=12.42, p=0.0033). Conclusiones: Los patrones de actividad en Tayras parecen variar levemente tanto en el espacio como el tiempo, con la estacionalidad sido el conductor más evidente. Nuestros resultados coinciden con estudios previos en hábitats similares, pero difieren en otras localidades con diferentes ecosistemas, destacando la variación dentro de las especie y  potencialmente relacionados con diferentes  subespecies.  Evaluar e identificar los aspectos básicos de la historia natural de E. barbara son necesarios para entender la ecología, historia de vida, los requerimientos de hábitat y el papel que cumple la especie dentro de los ecosistemas.

Palabras clave: Canal del Dique, Carnívoros, Frecuencias de Captura, Ritmo circadiano, Talamanca.

Introduction

Activity patterns are important characteristics of the natural history of a species, and represent an important input for understanding species ecology and their role in ecosystems (González-Maya et al. 2009, Vela-Vargas and Pérez-Torres 2012). Tayra (Eira barbara) is a small carnivore species distributed from central Mexico to southern Brazil and northern Argentina, and includes seven sub-species (Presley 2000), with E. b. sinuensis distributed from Costa Rica to Colombia (Presley 2000).
The species is considered as Least Concern with decreasing populations by the IUCN Red List of Threatened Species across its distribution (Cuarón et al. 2008).
Tayra has been historically understudied in general, with scarce field studies in few locations within its distribution, and most of the existing information has been obtained as “by-catch” from samplings aimed at more charismatic species (González-Maya et al. 2011). In general, there is a lack of information regarding several aspects of the species ecology and natural history (González-Maya et al. 2009, Presley 2000), being difficult to assess with certainty the conservation status of the species, and understand its biology and ecology (Cuarón et al. 2008, Tewksbury et al. 2014, Valenzuela 1998).

This scarcity of ecological information has difficulted understanding the species role in conserved and disturbed ecosystems or their response to human intervention. Previous studies have suggested Tayras to tolerate certain degrees of human intervention (Cuarón et al. 2008) and thus, it could potentially play a role on meso-carnivore release processes (González-Maya et al. 2011); thereof, Tayras could play a key role in ecosystems where there is no presence of larger predators, controlling populations of small and medium mammals (Camargo and Ferrari 2007, Lindenmayer 1999), however, the trophic dynamics under these scenarios are still mostly unknown.

Here we present the first comparative study on activity patterns for Tayra in Costa Rica and Colombia, specifically addressing geographic variation and the effect of seasonality over these patterns.

Methodology

Study area. The present study was conducted in two areas, one in Colombia and another in Costa Rica (Figure 1), both with different landscapes configuration and characteristics, but sharing the same Tayra sub-species.


Talamanca region is located in southern Costa Rica and is shared with Panama. It represents the most representative forest patch, and includes the largest protected area (La Amistad National Park) in the country. The region is traditionally divided in two slopes of the Cordillera de Talamanca, Caribbean (0-3600m asl) and Pacific (1,000-3600m asl) slopes, and is located from 8°37´- 9°38´N and 82°24´-83°25´W. Precipitation ranges from 2000 to 5000 mm annually with a mean rainfall of 3500 mm; a bi-seasonal weather occurs in the area with a dry season from November/December to March/April and a rainy season from April/May through October/November for the Pacific (González-Maya et al. 2013a), and almost permanent rainy season on the Caribbean with small dry seasons in March-April and September-October (Schipper 2010). The mean temperature is 27ºC and a relative humidity between 70% and 90%, during most of the year (United Nations Environmental Programme/World Conservation Monitoring 2011).

The Canal del Dique basin is located in northern Colombia, on the Caribbean coast near Cartagena city. This region has been historically used for cattle and cropping and the remnant natural vegetation is scarce and fragmented and embedded in a matrix of intensive agricultural use (Conservación Internacional Colombia and Fundacion Herencia Ambiental Caribe 2010); it ranges between 0 and 700 m asl and includes two protected areas and a matrix of pastures, crops and forest remnants (Conservación Internacional Colombia and Fundacion Herencia Ambiental Caribe 2010). The mean annual precipitation is around 1236 mm with a rainy season from May through November and a dry season between December and April. The mean annual temperature ranges from 27.2°C to 28.3°C and the relative humidity from 79% to 84%, with the maximum in 90% and the minimum in 68% (Conservación Internacional Colombia and Fundacion Herencia Ambiental Caribe 2010).

Methods. We evaluated the activity patterns derived from camera-trapping samplings in both areas through 12 sampling events and 324 camera stations. The sampling included 11 samplings for Talamanca from 2005 through 2009 and one for Colombia in 2009. Most of the effort was concentrated in Finca Las Alturas located in Las Tablas Protected Zone on the pacific slopes (González-Maya and Mata-Lorenzen 2008), and in the Corredor Biológico Talamanca-Caribe in the Caribbean slopes of Talamanca (Schipper 2010). The surveys were established with different purposes, in Costa Rica as part of ProCAT International continuous work in the region to evaluate conservation status and assessment of species for conservation planning. In Colombia as part of the projects of the Plan de Conservación de Felinos para el Caribe Colombiano (Colombian Caribbean Felids Conservation Plan) a joint initiative from several national institutions to preserve felid species and prey in the Caribbean region of the country (Castaño-Uribe et al. 2013).

Cameras were settled across the landscapes to be active in the field for a minimum of one month and a maximum of three months, using a standard delay of one minute. Some stations included two cameras to acquire images of both sides of the animal, however, such events were considered as one capture (González-Maya et al. 2009). All the stations were established in order to maximize capture probabilities; each area was surveyed to define the best probable sites for camera trapping, sometimes using human and animal trails, and also with the support from local communities’ knowledge (Chávez et al. 2013). Each camera was set 30 to 60 cm above ground on standing trees looking towards the selected sites, taking special caution to protect cameras from direct sunlight, flooding and even delinquency.

Data analysis. Every Tayra capture was considered an independent event and was summed for the total count; consecutive captures in 1h-periods were considered as one capture. We estimated capture frequencies by dividing the number of photos per sampling efforts and were organized in 1-hour discrete periods, and the capture histories were organized in periods following van Schaik and Grigffiths (1996) and González-Maya et al. (2009) in nocturnal (18h00-05h00), diurnal (06h00-17h00) and crepuscular (17h00-18h00 and 05h00-06h00h) activity periods. Capture frequencies, as proportions, were plotted across a complete day cycle (24h) to explore the activity patterns, and capture frequencies to determine the highest capture rates for each sampling. Kuiper tests (K) were used to assess if distribution of captures was homogenous across the 24-hour day periods for the overall dataset for each country and season (Vela-Vargas and Pérez-Torres 2012). Chi-square tests and contingency tables were used to test for significant differences among countries, periods, seasons and hours and we used Chao-Jaccard index to determine time overlap (Chao et al. 2004), as previously used for these type of analyses (González-Maya et al. 2009). All analyses were performed in language R (R Team Development Core 2008).
Results

A total of 324 camera stations were used in 12 samplings for a total of 8602 trap-days sampling effort (Table1). For Costa Rica, 289 camera stations were used, with a mean (±SD) of 26.2±35.2 cameras per sampling, and 35 stations were used for Colombia (Table1).


A total of 94 Tayra captures were obtained with 77 for Costa Rica and 17 for Colombia. Tayras were captured in 35 locations in Talamanca and in 6 locations in Canal del Dique, being the Talamanca Caribbean sampling the one with the highest number of captures, but Colombia the one with the highest frequency of the entire study (Figure 2). For Talamanca, three samplings showed the highest capture frequencies, two in the Pacific and one in the Caribbean, all of them in the rainy season and in three different years, and the lowest was the dry season sampling of the Pacific during 2007. When comparing the seasonality, the total capture frequencies for the entire study and for the Costa Rica portion showed a higher capture frequency for the rainy season (p=0.0025 and 0.001, respectively).


Activity patterns showed a heterogeneous distribution along the 24-hour day periods for the entire study (K=99.02, p<0.001), for Colombia and Costa Rica (K=70.47, p<0.001 and K=74.55, p<0.001, respectively) and for both rainy and dry seasons (K=69.96, p<0.001 and K=62.30, p<0.001, respectively). Tayras showed a predominant diurnal activity with respect to nocturnal and crepuscular activities for the entire dataset (X2=124.26, p<0.0001; Figure 3). In the two countries the species showed a highly similar pattern with diurnal activities being predominant both in observations (88%), and significantly in frequency (Colombia: X2= 22.27, p<0.0001; Costa Rica: X2= 102.19, p<0.0001).


For the Canal del Dique area, morning hours (8h00-9h00) represented the highest peak of activity of the species, followed by noon (12h00-13h00), while for Talamanca, mid-afternoon hours (15h00-16h00) was the main peak followed by before-noon hours (11h00-12h00). For all sites, three peaks of activity were notorious: 8h00 to 9h00, 11h00 to 12h00 and 15h00 to 16h00, following the same pattern of the independent country sites (Figure 4). Significant differences were found among the activities per hour for both countries (X2=46.38, p=0.0001) and for Colombia (X2=30.04, p=0.0118) but not for Costa Rica (X2=6.38, p=0.3819). The Chao-Jaccard index estimated an overlapping between activities patterns of 0.685 (68.5%) between countries, and a mean (±SD) of 75.6%±11.6% along sites, with overall higher similarity between sites in Costa Rica (87.6%±7.6%) than in Colombia (77.35%±8.6%).

Comparing activity patterns among seasons, we found significant differences for all periods in all sites (X2=14.15, p=0.0002), and for both seasons along day-hours (X2=28.42, p=0.0281) with overall higher significant mid-day and afternoon (17h00) activity during rainy seasons (X2=34.30, p=0.0031) and mid-morning (10h00) during the dry seasons (X2=12.42, p=0.0033; Figure 4).

Discussion

Tayras are among the most widespread and yet least known carnivore species in the tropics (Cuarón et al. 2008, Emmons and Feer 1999), with very scarce information regarding the most basic aspects of its ecology and natural history. Despite Tayras are commonly recorded in camera-traps in most tropical countries, and it is in general considered common (Cuarón et al. 2008), for our study sites is not as frequently recorded, with a relatively low number of detections despite a large trapping-effort. From the few details that can be derived from our study in terms of capture frequency, it seems Tayras are more commonly detected during rainy season but in few stations, which could indicate is not as common or widespread across habitats, or just supports the idea of Tayra as an elusive species or avoiding certain areas. However, even with a large sampling effort, camera specific locations can exert an important influence on the detection of the species.

According to frequency of captures, Tayras in both landscapes (i.e., Costa Rica and Colombia) showed similar patterns across sites with slight differences despite the heterogeneous landscapes analyzed. For Talamanca, Tayra represents a small carnivore under the dominance of big predators such as jaguar and puma, overlapping only on 45% of the time with jaguars (González-Maya et al. 2012), probably for predatory evasion, while in Canal del Dique, Tayra is one of the predominant predator species due to the low abundance and potential absence of top large predators (González-Maya et al. 2011, González-Maya et al. 2013b, González-Maya et al. 2013c); it is possible that for Canal del Dique there is some “plasticity” for the selection of activity-hours derived from the absence of dominant predators and suggested by the relatively higher concentrated detections found in Talamanca. The activity hours found can also be related with the availability of food or by seasonality, as found, and therefore by the consequent climatic constraints and variation.

Previous studies have suggested the species as crepuscular in unrelated ecosystems of Belize (Konecny 1989), in other areas of Colombia (Delgado-V et al. 2011) and nearby human settlements (Calderón-Capote et al. 2015), and mostly diurnal in similar ecosystems in Costa Rica (González-Maya et al. 2009), which in our study seems to coincide, with low activity during crepuscular hours and mostly active during the day. However, the specific hours during the day significantly vary according to season. The intraspecific variation in activity periods for the species across its distribution seem to highlight the effect of ecosystem types, human intervention and presence of top predators, and could potentially reflect variation at subspecific levels, such as E. b. senex in Belize. Diurnal activities are more frequent for Tayra probably because the easiness of foraging during day hours since the species is considered mainly omnivorous (Cuarón et al. 2008). However, Delgado-V et al. (2011) found an important peak of activity of Tayra in crepuscular periods in a fragmented forest in Colombia, suggesting that other factors may be influencing their activity patterns; further research should be done to understand these differences.

The strong effect of seasonality over Tayra activity patterns is an interesting finding, and is probably related with forest phenology and the consequent availability of food resources across habitats. Also, water sources can become scarce during dry seasons, therefore affecting space use, thus, ecosystems with more dramatic seasonality would likely reflect also considerable variation in these patterns. Activity patterns assessments in more localities and with larger seasonality variations would allow to draw better inferences regarding Tayra activities according to ecosystem, season and potentially more direct ecological drivers such as food availability and interspecific competition.

Activity periods for the species has been pointed as one of the main reasons for low inter-specific competition with other carnivores (Sunquist et al. 1989), so the results herein included can be a useful input for carnivore guild analyses in tropical ecosystems, and as a base to understand sympatry and interactions of small carnivore communities in the tropics (González-Maya et al. 2009). Furthermore, our results helps to understand better daily activities patterns of Tayras providing basic data for expanding our knowledge on the natural history of the species (Delgado-V et al. 2011) and potentially as the basis for further research into the role of small carnivores in tropical ecosystems, and their potential effect as a mesocarnivore, among other related questions both in ecology and conservation.

Acknowledgements

The funding for the present study was gently provided by ProCAT International, Finca Las Alturas and Wildlife Conservation Society for Costa Rica and ECOPETROL/REFICAR SA, Conservation International Colombia, Fundación Herencia Ambiental Caribe, CARSUCRE, MAVDT, and ProCAT Colombia for Colombia. The authors would like to thank C. Ange, C. Castaño, M. González, A. Cepeda, SA Balaguera-Reina and V. Gálvis for their support and aid to the entire project. We thank the National Parks Unit (UAESPNN) specially SFF Los Colorados and SFF El Corchal “El Mono Hernández” for their continued support.

Literature cited