Introduction

Mammalian carnivores typically occupy higher trophic levels and broadly affect ecosystem dynamics through trophic cascades (Ripple et al. 2014), understanding their community structure is fundamental to the development of conservation planning (Ritchie & Johnson 2009). In terrestrial carnivore guilds, both direct interference (e.g. killing and harassment) and resource competition among sympatric carnivores constrain the survival, fecundity, and eco-behavioural patterns of each species, affecting population dynamics as well as shaping community structure (Linnell & Strand 2000, Ritchie & Johnson 2009). According to the theory of competitive exclusion (Hardin 1960) and the niche variation hypothesis (Bolnick et al. 2007), competitive species need to partition their ecological niches to achieve sympatry in an environment. Partitioning of niches for food, use of space, and activity time in predator guilds is fundamental for the coexistence of multiple species (Schoener 1974, Amarasekare 2003). However, sympatric mammalian carnivores show substantial overlap in their trophic niches and share common resources (Manlick & Pauli 2020, Lanszki et al. 1999). Thus, spatial and temporal partitions play major roles in reducing competitive confrontations and facilitating successful sympatry (Monterroso et al. 2014, Karanth et al. 2017, Tsunoda et al. 2020).

Niche partitioning within a carnivore guild is affected by the availability of regional resources, landscape structure, and anthropogenic disturbances (Seveque et al. 2020, Manlick & Pauli 2020). For example, seasonal scarcity of food resources increases trophic overlap among competitive mesocarnivores, facilitating their spatiotemporal partitioning (Tsunoda et al. 2020). Further, anthropogenic landscape modifications decrease large canid and felid predators and allow dominance of superior competitors or introduced species, altering the spatiotemporal partitioning among the subordinate mesocarnivores (Wang et al. 2015, Tsunoda et al. 2018, Carricondo-Sanchez et al. 2019). Anthropogenic landscape alterations are major drivers of the depletion of available resources and eco-behavioural modification in wild animals (Bateman & Fleming 2012). Therefore, regional scale assessment of spatiotemporal patterns and intraguild interactions in carnivore guilds associated with environmental and landscape gradients is a central interest for conserving regional biodiversity (Ritchie & Johnson 2009).

Fig. 1.

Location of the study area and camera-trapping stations in central Bulgaria.

The present study aims to elucidate patterns in spatial distribution and diel activities among medium- and large-sized species in carnivore guilds across central Bulgaria. Although there have been some local-scale observations of spatiotemporal partitioning among mesocarnivores in this study region (Georgiev et al. 2015, Tsunoda et al. 2018, 2020), no previous assessment has been made at a regional scale. We hypothesized that spatial and temporal patterns in carnivore guilds might be associated with geo-environmental conditions, anthropogenic landscape modifications (e.g. less forested cover) as well as intraguild competition. To test our hypothesis, we investigated distributional patterns and diel activities among carnivores by analysing camera-trapping data from various altitudinal and landscape areas.

Study area

The study area is in central Bulgaria, including the Upper Thracian Plain and the three surrounding mountains, the Central Stara Planina, the Sredna Gora, and the Eastern Rhodopes (Fig. 1). The climate across the region is transitional between moderate-continental and continental-Mediterranean, but is cooler in high-altitude areas of the Central Stara Planina Mountains (> 1,000 m a.s.l) (Velikov & Stoyanova 2007). Vegetation and land use vary according to altitude. Lowland areas up to 500 m a.s.l. are a mosaic of agricultural fields, orchards, pastures, buildings, oriental hornbeam shrublands (Carpinus orientalis Mill.), and patches of oak (Quercus dalechampii Ten.) forest. Mountainous areas from 500 to 800 m a.s.l. are predominantly covered by secondary oak forests (Quercus spp.) and black pine (Pinus nigra J.F. Arnold) plantations, with natural beech forests (Fagus sylvestris L.) at higher elevations. On the basis of a faunal analysis using mammalian inventory records (Popov 2007), two zoogeographic zones in the study region are suggested: the western-central mountainous subdistrict and the southern lowland subdistrict.

Material and Methods

Statistical analysis

We analysed spatial distribution patterns of the carnivore species using a non-metric multidimensional scaling (NMDS) method with Bray-Curtis matrix. For standardization of sampling efforts, we calculated the number of observations per camera-day for each camera-trapping station for the nine focal carnivores as variables for the NMDS analysis. To assess factors associated with carnivore distribution patterns, we then fitted eight covariates for altitude, ruggedness, land cover (the relative proportion of forest, shrubland-grassland, cropland-pasture, or artificial construction coverage), camera locations inside/outside protected areas, and direct human activities (observations per camera-day of humans and domestic animals such as cows, horses and goats) (Mori et al. 2020) as vectors onto the NMDS ordination plot. Prior to the analyses, we checked multicollinearity among numerical parameters of the vectors (i.e. other than “camera locations inside/outside of protected areas” as nominal scale) using Spearman's correlation and variance inflation factor (VIF). These analyses were performed using R 4.1.3 (R Core Team 2022) and the “vegan” package (Oksanen et al. 2020) for R.

Table 1.

Overview of camera-trapping campaigns in the present study.

Table 2.

Sum of images captured (per 100 camera-trapping days) and total number of observed stations (proportions per total 42 stations) by camera-trapping of the focal carnivore species studied in central Bulgaria.

We estimated diel activities of focal species using Kernel density interference (Ridout & Linkie 2009) for the time of day (0:00-23:59) that the animal images were captured, and transformed into circular data (0-2π). In this analysis, we pooled the data from all camera-trapping stations. For this analysis, we used the “overlap” package (Meredith & Ridout 2021) for R.

Fig. 2.

Ordination plots of non-metric multidimensional scaling (NMDS) on carnivore species composition observed by camera-trapping in central Bulgaria (stress value = 0.153), fitted with environmental and anthropogenic vectors (indicated by arrows and ellipses) of camera-trapping stations. Red arrows represent vectors with statistical significance (P < 0.001), and grey arrows represent nonsignificant vectors (P > 0.05). Blue-dotted and black-dashed ellipses represent inside and outside protected areas, respectively.

Results

We obtained a total of 3,364 images of the nine focal carnivores from 13,988 camera-trapping days across 42 stations between 2015 and 2020. The species most frequently detected was the European badger, with 1,108 independent events, and the least frequently detected was the wolf, with eight independent events (Table 2).

Fig. 3.

Diel activity patterns (estimated using Kernel's density) of nine carnivore species observed by camera-trapping in central Bulgaria. The centre of the x-axis represents midnight. Vertical bars represent time (minutes) when the animals were observed.

Although there were relatively high correlations (r > |0.7|) among some parameters (e.g. altitude-ruggedness and forests-croplands/pastures), the low VIF values (< 2.5) indicated little multicollinearity (Table S1). With regard to spatial distribution patterns, the NMDS ordination plot indicated that the compositions of the focal carnivore guilds were different along the NMDS axis 1 (with stress value = 0.153), which was significantly associated with altitudinal gradients, ruggedness, forest cover and cropland-pasture surrounding the camera-trapping stations (P < 0.001; Fig. 2 and Table 3). Three of the focal species; grey wolf, brown bear, and European wildcat, were mainly distributed in forested mountains, while the golden jackal, European badger, and domestic cat were distributed in agricultural lowland habitats (Fig. 2 and Fig. S1). The remaining three carnivores (domestic dog, red fox, and martens) were mainly distributed in sites with intermediate levels of altitude, ruggedness and forest/agricultural coverage (Fig. 2). Moreover, camera locations inside/outside of protected areas showed marginal significance (P = 0.075; Table 3), indicating both wildcat and domestic dog were more frequently observed at camera stations inside protected areas (Fig. 2).

With regard to diel activity, six of the focal carnivores (grey wolf, red fox, European badger, martens, European wildcat, and domestic cat) had mostly nocturnal patterns (Fig. 3). The remaining three species had different patterns of diel activity. The brown bear was cathemeral, the golden jackal was crepuscular, and the domestic dog was diurnal (Fig. 3).

Discussion

The present study showed low detection rates for the two largest species, the grey wolf and brown bear, by camera-trapping survey. This outcome was similar to the result from a previous study conducted in the Osogovo Mountains on the Bulgarian-North Macedonian border (Zlatanova & Popova 2018). Low detectability of these large carnivores was typically owing to their large home-range sizes and low occurrence rates at a given trapping station (Chutipong et al. 2014). Baits or lures at camera-trapping stations potentially improve the detectability of large carnivores (Satterfield et al. 2017) with potential to alter their behavioural responses (Meek et al. 2014). For this reason, we decided not to use baits or lures in the present study. Although our sample sizes were small for these two species, we reported them in this study because there was little information on the spatial distributions and daily activity in central Bulgaria.

Table 3.

Results of non-metric multidimensional scaling (NMDS) scores (axes 1 and 2) on carnivore guilds fitting with geological, landscape and anthropogenic variables.

Our findings indicated that the spatial distribution of the focal carnivore guilds varied with changes in altitudinal gradient, ruggedness and landscape features. The results of NMDS indicated that the two largest species, the grey wolf and brown bear, might be associated with forest, high altitudes and ruggedness, consistent with previous reports from other study sites in Bulgaria (Zlatanova & Popova 2013, Gavrilov et al. 2015). These species are typically intolerant of anthropogenic landscape modifications, such as urbanization, intensive farming, and habitat fragmentation (Cimatti et al. 2021). In contrast to the large carnivores, the golden jackal and European badger were mainly distributed in the less-wooded agricultural-lowlands. Jackals prefer human-modified landscapes with fewer forested areas and avoid mountainous habitats (Šálek et al. 2014b, Spassov & Acosta-Pankov 2019). The presence of wolves may also be a major driver constraining the distribution of the jackal in high-altitude areas (Krofel et al. 2017, Spassov & Acosta-Pankov 2019), as was demonstrated in our NMDS ordination plot (Fig. 2). Badgers were also found mainly in agricultural lowlands, but in more-wooded sites than those in jackals (Fig. 2). Badgers often use pasture to forage for their staple food, earthworms (Kruuk et al. 1979), and select open-wood pastures for their sett (Virgos & Casanovas 1999). Smaller mesocarnivores (foxes, wildcats, and martens) were mainly found at higher altitudes and in more forested habitats than jackals and badgers, although these species can also adapt to human-altered landscapes (Šálek et al. 2014a). Local-scale competitive avoidance of large predators is a key driver of the spatial structures of mesocarnivore guilds (Monterroso et al. 2020). Our previous observations demonstrated spatiotemporal separation between jackals and the smaller carnivores due to avoidance of direct confrontation at small scales (Tsunoda et al. 2018, 2020). Although based on a marginally not-significant result, we showed that wildcats tended to occur inside protected areas (Fig. 2 and Table 3). This species has been listed as protected under Bulgarian national law since 2007 and conservation actions to maintain their habitats have been implemented in nature parks and reserves, including the Central Balkans National Park and the Eastern Rhodopes (Spassov et al. 2011).

The two domestic species, the dog and the cat, had different spatial distributions. Dogs were found at higher altitudes and in more forested habitats than cats in the present study, although free-roaming animals typically were found closer to settlement cores than wild species (Vanak & Gompper 2010, Krauze-Gryz et al. 2012). There were two possible reasons for the distribution patterns of the dog. First, feral and stray dogs are culled by hunters in Bulgaria (Raichev et al. 2013), limiting their densities in rural areas. Second, because of methodological limitations on identification of the animal images, the dogs observed by our camera-trapping may have included hunting dogs, which occur in forested areas. Dogs were also found inside protected areas (with a marginal statistic-significance; Table 3). Our camera stations were located in buffer zones of protected areas (i.e. non-strict reserves) and their fringes, including pasture, agricultural fields, small hamlets as well as hunting areas.

With regard to diel activities, six of our focal carnivores were mainly nocturnal. Generally, diel activity patterns in carnivores are synchronized with those of their sympatric prey (Monterroso et al. 2014). For example, wolves predominantly prey on red deer and wild boar (Zlatanova et al. 2014), and the smaller mesocarnivores (foxes, wildcats, domestic cats, and martens) prey on rodents (Hisano et al. 2013, Tsunoda et al. 2017, 2019). All of these prey animals (i.e. ungulates and rodents) are typically nocturnal (e.g. Monterroso et al. 2014, Mori et al. 2020). Badgers are also nocturnal (Barrull et al. 2014, Tsunoda et al. 2020), seeking and foraging on earthworms during the night (Kowalczyk et al. 2003). The remaining three species are typically omnivorous, predominantly consuming vegetable matter such as fruits and acorns (bear) (Bojarska & Selva 2012) or anthropogenic food resources, such as human food waste (dogs and jackals) (Vanak & Gompper 2009, Raichev et al. 2013).

Moreover, nocturnal activity in mesocarnivores (fox, wildcat, martens, and domestic cat) may be facilitated by temporal niche separation from sympatric competitive dogs and jackals, thereby reducing direct confrontations (Scheinin et al. 2006, Vanak et al. 2009). Temporal partitioning between competitive mesocarnivores plays a key role in successful sympatry in many carnivore guilds (Monterroso et al. 2013, Barrull et al. 2014, Carricondo-Sanchez et al. 2019, Tsunoda et al. 2020).

Conclusions

In agreement with our hypothesis, the present study demonstrated distributional patterns and diel activities in carnivore guilds that might be associated with landscape features and intraguild interactions. We found that modifications of the landscape by humans (e.g. forest fragmentation and agricultural land use), rather than direct human presence, may affect spatial distributions in carnivore guilds. Our findings also indicated that potential interspecific competition (interference or confrontation) might result in partition of their spatial distribution and temporal activities. Our results showed that intense human activity in lowland areas might constrain the distribution of wolves and bears, resulting in a broader distribution of the golden jackal in the local mesocarnivore guild. We conclude that spatial and temporal partitioning may be key to successful sympatry among trophic competitive mesocarnivores (Monterroso et al. 2020, Tsunoda et al. 2020).