Human disturbance has the potential to alter competitive interactions, favoring species better able to adapt to areas used by humans. One such species is the spotted hyena (Crocutu crocuta), which has been successful in human dominated areas throughout Africa, competing through kleptoparasitism with other carnivore species (e.g., leopards [Panthera pardus]). In the Udzungwa Mountains, Tanzania, leopard density declines sharply close to human settlements and hyenas are their only competitors. Using camera trap data and a spatio-temporal occupancy model, we assessed the relative dominance of each species through spatial co-occurrence, altered activity patterns and temporary site avoidance. We tested the hypothesis that hyenas gain a competitive advantage over leopards in human-dominated areas due to their relatively higher tolerance for anthropogenic activities. We found that while hyena occupancy was best predicted by prey occupancy and not strongly affected by landscape factors associated with humans, leopards, especially male leopards, were less likely to be detected close to human settlements. Female leopards, which are smaller than males, exhibited activity shifts and temporary site avoidance in response to hyenas, whereas hyenas shifted their activity patterns in response to male leopards. These results suggest that while hyenas may be behaviorally dominant over female leopards, they are subordinate to male leopards. We found that male leopards and hyenas were less likely to co-occur closer to people, especially where prey was scarce, suggesting subordinance of hyenas to male leopards may be mitigated by human disturbance. Furthermore, young male leopards shifted their activity patterns to be more diurnal in response to hyena presence, suggesting that dominance relationship between hyenas and leopards develops with age and is probably related to body size. These results indicate that human disturbance has the potential to affect the competitive relationship between leopards and hyenas in the Udzungwa mountains, but that relationships will vary with sex and body size.

Ethics Statement

Data collection consisted of non-invasive, remotely set camera traps and did not involve direct contact or interaction with the animals. This research was conducted under research permit numbers 2013-274-NA-2013-111 and 2014-137-ER-2013-111, issued to RWH by the Tanzania Commission for Science and Technology (COSTECH).

Study area

The Udzungwa Mountains are in south-central Tanzania (centred on 7^◦46’ S, 36^◦43’ E) and are part of the renowned biodiversity hotspot of the Eastern Arc Mountains (Rovero, Menegon, et al. 2014; Burgess et al. 2007). The Udzungwa Mountains are completely encompassed by agriculture, with no connectivity to nearby protected areas (Rovero, Martin, et al. 2014). Northern Udzungwa is protected as the Udzungwa Mountains National Park (UMNP; 1990 km²) which merges to the south and west with the central part of the Kilombero Nature Reserve (1345 km² (Figure 1).

Camera trapping

We deployed six camera trap arrays covering approximately 2500 km² across five habitat types. Each array consisted of 25-34 paired camera traps and was deployed for 12-49 days (mean=31) for a total of 164 sites monitored for 5038 camera nights of total survey effort (Figure 1; Table S1). We performed the camera trapping during the dry season, August to December 2013 and from June to December 2014. We followed the protocol by Henschel and Ray (2003) for the camera trap placement with average trap spacing of 1.6 km, placed in an evenly spaced random array. Mean distance between the edge of each camera trapping array was 22.8 km (min. 5 km, max. 49 km). At each site a pair of camera traps were placed 30-40 cm above ground level and 3-4 m from the centre of an animal trail facing each other. A minimum of one camera trap per station was a white Xenon flash Cuddeback Ambush (Cuddeback Non Typical Inc., Istanti, WI, USA). In 87 of 164 stations, the second camera was an infrared camera UOVision 565HD IR+ (UOVision Technology, Shenzhen, China) set on 15-second video recording mode. Sexing of leopards was determined by the presence of external male genitalia and confirmed by two independent observers (RWH and Francesco Rovero). Male leopard age and presumed dominance status was determined by presence/absence of dewlap (Hunter, Henschel, and Ray 2013).