Different sampling designs to improve jaguar (Panthera onca) population estimates by SECR in one of the smallest viable populations of the species: simulations and first field results

Link To Share This Poster: https://cdmcd.co/zY9Jjz
Live Discussion Link: https://cdmcd.co/PJ9xzb

Cristiano T. Trinca, Trinca – Ciência, Tecnologia e Consultoria Ambiental, Ribeirão Preto, SP, Brazil, Francesca B. L. Palmeira, RLadies Ribeirão Preto, Ribeirão Preto, SP, Brazil and Beatriz M. Beisiegel, NGI Iperó, Instituto Chico Mendes de Conservação da Biodiversidade, Brazil

Background/Question/Methods Spatially explicit capture-recapture (SECR) analysis, using photographic identification of individually distinct animals, is the main method to assess jaguar (Panthera onca) populations. Previous studies have indicated that at the Atlantic forest, the Paranapiacaba Remnant in Southeastern Brazil protects one of the smallest viable population of the species, with a very low population density (0.66 ind/100 km2 , IC95% 0. 29-1.17). Obtaining enough jaguar captures of this population to allow population estimates by SECR is challenging, because at least five animals with enough spatial recaptures are needed. Therefore, we conducted a simulation study to (1) determine the better sampling grid prior to field allocation of the camera stations and (2) obtain a robust estimate of jaguar density. We tested three designs of trap array: (i) a single block, (ii) in clusters and (iii) an elongated design. The designs were applied to populations of 13, 20 and 35 individuals. We also varied the number of stations (60-62 or 100) and the sampling period (60, 90, 180, 360 days). At least 100 results were simulated for each combination. We used the secr and spatstat packages for R, with a habitat mask considering a buffer of 62 km, defined on the basis of the distance traveled by the area's male jaguars, covering the entire habitat mask (7,490 km2) or just the Paranapiacaba Remnant.

Results/Conclusions
Elongated designs and longer sampling periods resulted in more precise and accurate estimates. More stations resulted in more precise estimations. Based in these simulations, we designed a elongated sampling, with four consecutive sampling periods of 90 days and at least 25 stations each, with a first sampling period allocated along the longer axis of the Paranapiacaba Remnant, in preferential jaguar paths (old trails and dirt roads). Incomplete results of the first sampling period (27 stations, 70 days, D=0.246 ind/100 km2 , IC95% 0.09-0.62, g0=0.584+-0.266, sigma=5,333m+-870) indicate that although the sampling design is spatially adequate, because we captured five identified males, it needs refinement, since no identified female was detected. The latter may be due to few females in the population or their avoidance of the trails used by males. Therefore, we need to sample also outside the preferential jaguar paths to try to detect females, which in turn poses a challenge due to the large area to be sampled.