Prey availability and temporal partitioning modulate felid coexistence in Neotropical forests

Abstract

Carnivores have long been used as model organisms to examine mechanisms that allow coexistence among ecologically similar species. Interactions between carnivores, including competition and predation, comprise important processes regulating local community structure and diversity. We use data from an intensive camera-trapping monitoring program across eight Neotropical forest sites to describe the patterns of spatiotemporal organization of a guild of five sympatric cat species: jaguar (Panthera onca), puma (Puma concolor), ocelot (Leopardus pardalis), jaguarundi (Herpailurus yagouaroundi) and margay (Leopardus wiedii). For the three largest cat species, we developed multi-stage occupancy models accounting for habitat characteristics (landscape complexity and prey availability) and models accounting for species interactions (occupancy estimates of potential competitor cat species). Patterns of habitat-use were best explained by prey availability, rather than habitat structure or species interactions, with no evidence of negative associations of jaguar on puma and ocelot occupancy or puma on ocelot occupancy. We further explore temporal activity patterns and overlap of all five felid species. We observed a moderate temporal overlap between jaguar, puma and ocelot, with differences in their activity peaks, whereas higher temporal partitioning was observed between jaguarundi and both ocelot and margay. Lastly, we conducted temporal overlap analysis and calculated species activity levels across study sites to explore if shifts in daily activity within species can be explained by varying levels of local competition pressure. Activity patterns of ocelots, jaguarundis and margays were similarly bimodal across sites, but pumas exhibited irregular activity patterns, most likely as a response to jaguar activity. Activity levels were similar among sites and observed differences were unrelated to competition or intraguild killing risk. Our study reveals apparent spatial and temporal partitioning for most of the species pairs analyzed, with prey abundance being more important than species interactions in governing the local occurrence and spatial distribution of Neotropical forest felids.

Fig 1. Target Neotropical cat species and summary hypotheses.

From large to smaller species: A–Jaguar, B–Puma; C–Ocelot, D–Jaguarundi, and E–Margay. Spatial partitioning hypothesis (including jaguar, puma and ocelot): 1) prey availability would be more important in determining felid habitat use than landscape covariates; 2) based on body weight ratios, jaguar exert negative effects on puma and ocelot, and puma exerts negative effects on ocelot. Temporal partitioning hypothesis (including all five species): higher temporal segregation between species pairs experiencing higher chances of competition. Black arrows indicate strong relationship and grey arrows indicate weaker relationship. Photos by: CAX (A, C and E), COU (B) and YAN (D).

Fig 2. Location of the eight Neotropical study sites and a map of a typical camera trap array at Caxiuanã National Forest (CAX), Brazil.

Each point represents a camera trap location. Camera traps are distributed in two sampling arrays of 30 camera traps each (North and South of Caxiuanã River) See site codes on Table 1.

Fig 3. Relative importance of environmental and interaction covariates on the habitat use of three Neotropical forest cats.

Row A–Sum of models weights (AIC_wt/QAIC_wt) of occupancy models to assess habitat factors; row B–Sum of models weights (AIC_wt/QAIC_wt) of occupancy models to assess both habitat factors and species interactions.

Fig 4. Covariates effect on habitat use of jaguar, puma and ocelot.

Beta estimates with 95% of confidence interval estimated from single-season species models: row A—Beta estimates from occupancy models to assess habitat factors; row B–Beta estimates from occupancy models to assess both habitat factors and species interactions (The beta estimates has an effect on the dependent variable when confidence interval do not include 0).

Fig 5. Coefficient of overlap in daily activity patterns between jaguar, puma and ocelot in Neotropical forest sites.

X and Y axis represent time of the day and activity density, respectively. Overlap is represented by blue shaded areas and Δ is the coefficient of overlap (varying from 0 –no overlap to 1 –total overlap). (*) indicates significant differences. Study site is indicated in the top left corner.

Fig 6. Coefficient of overlap in daily activity patterns between ocelot, jaguarundi and margay in Neotropical forest sites.

X and Y axis represent time of the day and activity density, respectively. Overlap is represented by blue shaded areas and Δ is the coefficient of overlap (varying from 0 –no overlap to 1 –total overlap). (*) indicates significant differences. Study site is indicated in the top left corner.

Fig 7. Intraspecific variation in daily activity patterns in felid species across eight Neotropical forest sites.

X and Y axis represent time of the day and activity density, respectively.

Fig 8. Daily activity level of felid species across the eight Neotropical forest sites.

Proportion of active hours per day. Error bars represent the standard error.