Camera trapping and occupancy analyses to study tropical and temperate mammals from local to global scale

Abstract

The contemporary biodiversity crisis caused by anthropogenic disturbances calls for urgent actions to reduce the pervasive loss or decline of species that are jeopardizing the functioning of the ecosystems. Halting species decline is only achievable by understanding the causes of biodiversity changes, the responses of species to disturbances, and consequently by promoting conservation actions that can mitigate impacts. In this context, reliable population and community assessments based on sound data and robust analytical frameworks can help to guide conservation strategies. Particularly, information for monitoring changes should come from data collected systematically with tools that can also detect cryptic and rare species and acquire data on the whole community, and from analytical frameworks that can account for imperfect detection. Camera traps as a detection tool, and occupancy models as an analytical framework perfectly suit these requirements. This thesis uses tools to address various aspects of anthropogenic impact on wild mammals, by assessing the potential vulnerability of populations and communities in temperate and tropical protected areas, at both local and global scale, and with a focus on tropical forests. I targeted mammals both for the key ecological roles they play and because camera traps are particularly suitable for targeting this taxon. In particular, Chapters 1 - 3 focus on single populations of overlooked and elusive mammal species at a local scale (in Tanzania and Mongolia), while Chapters 4 - 5 focus on mammalian communities at a larger scale (across protected areas in Italy and in the tropics). In Chapter 1, I analysed data from a systematic camera trapping protocol deployed in a remote reserve in Tanzania to determine the species checklist and assess vulnerability of most threatened species. Among these, the African golden cat (Caracal aurata) was detected for the first time in this country, confirming a range expansion for this species and the establishment of the easternmost African population. Results from the occupancy analyses of this species indicated marked vulnerability to human encroachment in the reserve, as I estimated greater occurrence in areas of primary, closed-canopy and dense forest in the interior of the Nature Reserve and away from settlements. In Chapter 2, I used bycatch data from a systematic camera trapping project in the Mongolian steppe that primarily targeted the snow leopard (Panthera uncia) to study the habitat selection preferences of the neglected Pallas’s cat and its relationship with livestock presence. I estimated that the preferred habitat of the Pallas’s cat (Otocolobus manul) in the study region coincided with areas encroached by livestock, with diel activity segregation that limits direct encounters. Its dependence on areas that are used for grazing may eventually threaten the cat with habitat degradation, prey depletion, predation by dogs, and poisoning from pest control, ultimately undermining the long-term viability of the population. In Chapter 3, I used density data from acoustic distance sampling with camera trapping data to carry out the first comprehensive assessment and spatially-explicit density estimation of the endemic Sanje mangabey (Cercocebus sanjei), across its entire distribution range in Tanzania. I found that the higher density found in the forest of lower protection compared to previous studies instil optimism for the conservation of the species, although occupancy analysis highlighted the vulnerability of this primate to human disturbance. In Chapter 4, we investigated the spatial and temporal patterns of mammalian communities in relation to human frequentation in four protected areas in central and northern Italy, monitored with a standardized camera trapping protocol. By using a Bayesian multi-area occupancy model for communities that also accounts for the time interval of the detection events (i.e., diurnal, crepuscular and nocturnal events), we found that wildlife consistently increased their nocturnal activity at locations with higher rates of human passage, therefore diminishing the chances to encounter humans. We also found differences in behavioural responses according to species’ body size and trophic guild: larger species and omnivores tended to avoid humans in space as well as in time, while larger species also became more diurnal further away from urban areas. Human presence may indeed promote risk perception and the creation of a landscape of fear that can result in cascading effects on the whole community. However, the adoption of a primarily nocturnal activity can act as coping mechanism to allow coexistence between humans and wildlife. Finally, in Chapter 5, I investigated pan-tropical patterns of changes in mammal communities in relation to proximity to settlements, increasing human density, habitat loss and forest fragmentation in the broader landscape surrounding protected forests. Although protected areas represent indeed a critical tool for conservation, evidence warns that human impacts around reserve threatened biodiversity within. By using camera trapping data systematically collected in 37 tropical forests, I simultaneously estimated patterns of community richness and occupancy by accounting for imperfect detection. I found that the escalating human densities around reserves impact richness within, while occupancy was not affected. Moreover, the extent of forest cover in the landscape is associated with higher richness and occupancy within reserves, with occupancy increasing with less fragmentation. These results suggest an extinction filtering of sensitive species, and a potential increase in the abundance of adaptative species. While the anthropogenic impact on the landscape highly affects wildlife within reserves, I found that the viability of mammals depends on environmental changes happening beyond protected areas’ boundaries. Overall, the results of this thesis demonstrated how the use of systematic camera trapping data coupled with robust analytical framework that accounts for imperfect detection could represent a valid tool for population and community assessments in view of promoting wildlife conservation and mitigation strategies. Hence, mammal assessments are fundamental in view of the new challenge to doble the protected area estate within the upcoming years to ensure that current and future reserves can sustain viable wildlife populations.