Ecology and conservation of the jaguar (Panthera onca) in the Cerrado grasslands of central Brazil

The jaguar Panthera onca is the largest felid of the Americas and threatened by habitat loss and direct conflict with humans. Owing to low population densities and cryptic habits it remains little studied, especially in the central Brazilian Cerrado grassland. The present dissertation investigated population status and ecology of the jaguar in Emas National Park (ENP), one of the most important Cerrado reserves. Abundance and density is baseline information for conservation planning. In chapter 2, I estimated jaguar density based on data from a park-wide camera trapping study using regular and spatially explicit capture-recapture models. I observed 10 adult individuals and estimated a density of 0.3-0.6 individuals 100km-2. The spatially explicit approach accounts for animal movement on and off the trapping grid in a formal way and is therefore preferable over the non-spatial approach. Due to the considerable degree of isolation, the small size of the ENP jaguar population exposes it to a risk of extinction by stochastic factors. Since predator ecology is largely governed by their prey, understanding a predator’s foraging ecology can contribute to its conservation. In chapter 3, I analyzed jaguar feeding ecology based on scats collected with the aid of scat detector dogs. Despite a large effort, we only obtained 39 genetically identified jaguar scats. Since the sample was ill-suited for standard analyses, I measured prey availability based on distributional overlap with the predator derived from occupancy models to investigate selection. I further developed a qualitative optimal foraging model assessing expected foraging costs and benefits to predict prey preferences. Giant anteaters had the second lowest distributional overlap with jaguars but accounted for 75 % of its diet, indicating selection. The model predicted giant anteaters to be the most profitable prey, suggesting that selecting for them is the optimal foraging strategy for jaguars in ENP. Resource partitioning is a mechanism fostering co-existence. In chapter 4, I investigated space partitioning between jaguars and pumas in ENP using camera trapping data and occupancy models that account for imperfect species detection and spatial autocorrelation. Jaguar occurrence was positively correlated with the amount of dense habitat in the trap area and negatively correlated with the distance to water. Puma occurrence was negatively correlated with jaguar presence. Both species occurred less often together than expected under independent distributions. Jaguars used more of ENP than pumas. The more generalist puma is frequently found beyond the park boundaries, whereas jaguars are largely restricted to the park. Though outcompeted by the jaguar within the park, the conservation situation of the puma is less worrying than of the isolated small jaguar population. Overall, camera trapping proved more efficient than scat collection to study jaguar ecology in ENP because of the low population density and the ability of camera traps to simultaneously accumulate a larger amount of effort. Hierarchical models are flexible to suit specific sampling and data situations and explicitly account for the observation process, and are thus ideal to investigate carnivore ecology. Jaguar conservation in ENP and the Cerrado should focus on the issue of small population isolation and the potential for dispersal corridors. In the same context, research should focus on the identification of remaining Cerrado jaguar populations and landscape use of jaguars outside of protected areas.