Ecological Archives E091-239-A1

Beth Gardner, Juan Reppucci, Mauro Lucherini, and J. Andrew Royle. 2010. Spatially explicit inference for open populations: estimating demographic parameters from camera-trap studies. Ecology 91:3376–3383.

Appendix A. Details about Pampas cats and the Pampas cat study.

The Pampas cat is currently listed as Near Threatened by the IUCN (Cat Specialist Group 2002) and as Vulnerable in Argentina by the Sociedad Argentina para el Estudio de los Mamíferos (Díaz and Ojeda 2000). The Pampas cat is a small felid, weighing 3–7 kg, and measuring 30–35 cm length (García-Perea 1994). Pampas cats have distinct spot patterns, allowing individuals to be uniquely identified using camera trapping techniques (Fig. A1). The Pampas cat has a wide distribution in South America, extending to all Argentinean continental regions, Uruguay, southern and central Brazil, Bolivia, and Chile (Nowell and Jackson 1996, García-Perea 1994). While the Pampas cat can exist in a diversity of habitats, knowledge of this species' biology and habitat use are very limited (Nowell and Jackson 1996).

photoA1
 

   FIG. A1. A Pampas cat captured by a camera trap during the study in northwestern Argentina. Note the spot and stripe patterns that allow cats to be individually identified.


2.  Study Design

The study area is located in the province of Jujuy in the Northwest of Argentina (22° 30''S, 66° 30''W) near the borders with Chile and Bolivia, within the altitude range from 3900 to 4600 m. The topography of the area is very fragmented, with a large number of canyons and steep cliffs, and with very sparse vegetation. Due to the elevation, the temperature can range from 35°C during the day to -10°C at night, and the annual rainfall is less than 400 mm and concentrated in summer.

Sampling was conducted from 6 October to 14 December in 2006 and from 22 April to 16 June in 2007 using an array of camera traps. We treat these two sampling sessions as primary periods with closure over the secondary sampling intervals. Some of the traps were moved within each year, and the trap locations were not consistent between years (Fig. A2). There were 22 camera trap stations (2 cameras per trap, to allow individuals to be identified by the coat spot patterns, Fig. A1) deployed in each year, of which 7 traps were moved once in 2006 and 4 traps were moved once in 2007. As is typical in camera trapping studies (e.g., Karanth and Nichols 2002), the trap locations are strategically placed at ecologically optimal sites selected by biologists and based on signs of usage including animal tracks and scat. The mean distance between sampling stations was 1-2 km, to ensure that no individual's home range would exist between traps and therefore have no probability of capture (Karanth and Nichols 2002). There is no published information about the home range size of the Pampas cat in this region, thus the best guess of home range size was inferred from similar sized cats and based on knowledge of the area. This strategic placement of the trap locations results in an irregularly spaced array (Fig. A2). The coordinates for the trap locations were recorded in UTM, WGS 84 Zone 19. For the purposes of the analysis, we scaled the UTM coordinates from meters into kilometers.

There were 6 sampling intervals in 2006 and 2007 (approximately 9 trap days per sample occasion for each trap location). For individual identification, all photos were examined separately by authors and collaborators, resulting in encounter histories for 22 individuals. Of those individuals, 7 were captured only in 2006, 9 were captured only in 2007, and 6 were captured in both years. In 2006, 8 individuals were captured once, 4 individuals were captured twice, and 1 was individual captured three times. In 2007, 13 individuals were captured once and 2 individuals were captured twice. In the combined data set for both years, 15 individuals were captured once, 2 captured twice, 3 captured 3 times, and 2 captured 4 times. The encounter information for each individual is presented in Table A1.

For this particular data set, time was discretized into approximately 1 week sampling occasions; multiple captures within each interval at the same camera station were allowed. Because the cameras are operating almost continuously (i.e., there are times when malfunctions occur and the camera has some time delay in resetting), an individual could conceivably be caught an arbitrary number of times in a particular trap during a sample occasion. Thus, we used a Poisson encounter process to analyze the data set. However, the actual data set did not result in multiple recaptures of any individual in a particular trap during any one sample occasion. As such, results using the Poisson encounter process should be very similar to results using a Binomial encounter process.

FigA2
 
   FIG. A2. The locations of traps, separated by year. The top figure is 2006, the bottom figure is 2007. The traps are identified by a trap number which corresponds to the trap number listed in Table A1. The trap numbers shown in blue indicates traps that were repositioned slightly during the study.

3.  Analysis

Bayesian analysis of the model was conducting using data augmentation and was implemented in the software package WinBUGS (Gilks et al. 1994). The model specification for WinBUGS is provided in the Supplement (WINBUGS.TXT). The MCMC algorithm was run for 20,000 iterations, with 3 chains, the first 10,000 were discarded, and posterior summaries were computed from the remaining iterations after thinning by 5. We used noninformative priors for the parameters in the model, specifically, ψ, φ, and γ were all specified as Uniform(0,1). For λ0 we used a Gamma(0.1, 0.1) distribution for the prior and for σ2 we assigned a Uniform(0, 10) prior. We set M = 300 for the analysis. Here, we define λ0 as constant across the two primary sampling occasions; however, in a study with more data, this assumption could be relaxed and the encounter rate could vary by year.

TABLE A1. Each individual encounter and the trap identification. See Fig. A2 for the trap locations.

   Individual

 Year

 Occasion

 Trap Number

1

 2006

 1

 2

1

 2006

 3

 2

1

 2006

 6

 10

1

 2007

 4

 53

2

 2006

 1

 8

2

 2006

 5

 13

2

 2007

 6

 43

3

 2006

 1

 11

3

 2007

 4

 40

4

 2006

 1

 4

5

 2006

 2

 21

5

 2006

 3

 12

5

 2007

 5

 34

6

 2006

 2

 12

7

 2006

 3

 3

7

 2006

 6

 18

7

 2007

 1

 32

8

 2006

 3

 21

8

 2006

 4

 21

8

 2007

 1

 37

8

 2007

 2

 37

9

 2006

 4

 21

10

 2006

 5

 4

11

 2006

 5

 1

12

 2006

 6

 14

13

 2006

 6

 22

14

 2007

 2

 53

14

 2007

 4

 35

15

 2007

 2

 53

16

 2007

 2

 37

17

 2007

 3

 44

18

 2007

 3

 48

19

 2007

 3

 54

20

 2007

 5

 54

21

 2007

 5

 33

22

 2007

 6

 37

Density estimates that we provide for the Pampas cat appear to be high relative to other felid species (Trolle and Kéry 2005, Cuéllar et al. 2006). However, this was a pilot study meant to develop techniques and to establish a baseline for density estimates in the upper end of the habitat quality spectrum. As such, the study area is relatively small and comprised of a large amount of very good habitat for Pampas cats. As a result, our density estimates cannot be extrapolated to any meaningfully large portion of the range for this species. To do so would require multiple study sites chosen in a statistically valid manner across the range of the species.

In addition to density, we also estimated the mean per capita recruitment (ρ = 0.23, ρ is defined as R/N1) and mean apparent survival from the first survey to the second (φ = 0.79), which is consistent with other felid species (Karanth et al. 2006, Haines et al. 2005). However, it should be noted that the Pampas cat is a mid-sized felid and the other published studies are on larger felids. One reason for this lack of information on smaller felids is because estimating vital rates for populations of endangered species and those hard to detect is challenging (Thompson 2004). While difficult to obtain, estimates of vital rates, such as survival and recruitment, are useful in understanding what drives changes in population abundance. This information can be important in developing sound conservation strategies and management plans, and assessing population viability.

LITERATURE CITED

Cat Specialist Group, A., 2002. Oncifelis colocolo. In IUCN 2007. IUCN Red List of Threatened Species. Downloaded on 05 October 2008.

Cuéllar, E., l. Maffei, R. Arispe, and A. Noss. 2006. Geoffroy's cats at the northern limit of their range: Activity patterns and density estimates from camera trapping in Bolivian dry forests. Studies on Neotropical Fauna and Environment 41:169–177.

Díaz, G., and R. A. Ojeda, editors. 2000. Libro rojo de mamíferos amenazados de la Argentina. Sociedad Argentina para el Estudio de los Mamíferos, Mendoza, Argentina.

García-Perea, R. 1994. The pampas cat group (genus lynchailurus severtzov, 1858)  (carnivora: Felidae), a systematic and biogeographic review. American Museum Novitates 3096:1–36.

Gilks, W., A. Thomas, and D. Spiegelhalter, 1994. A language and program for complex  bayesian modelling. The Statistician 43:169–177.

Haines, A., M. Tewes, and L. Laack. 2005. Survival and sources of mortality in ocelots. Journal of Wildlife Management 69:255–263.

Karanth, K., J. Nichols, N. Kumar, and J. Hines. 2006. Assessing tiger population dynamics using photographic capture-recapture sampling. Ecology 87:2925–2937.

Karanth, K. U., and J. D. Nichols. 2002. Monitoring tigers and their prey: a manual for researchers, managers, and conservationists in Tropical Asia. Centre for Wildlife Studies.

Nowell, K., and P. Jackson. 1996. Wild Cats. Status Survey and Conservation Action Plan. IUCN - SSC Cat Specialist Group. IUCN, Gland, Switzerland. (online version).

Thompson, W. L. 2004. Sampling rare or elusive species: concepts, designs, and techniques for estimating population parameters. Island Press, Washington, DC, USA.

Trolle, M., and M. Kéry, 2005. Camera-trap study of ocelot and other secretive mammals in the northern Pantanal. Mammalia 69:405–412.


[Back to E091-239]