Drivers of human-wildlife impact events involving mammals in Southeastern Brazil

Highlights

• Reducing accidents with fauna benefit conservation, economy, and human safety.

• We evaluated natural and anthropogenic factors that affect accident risk for mammals.

• Requested removals, traffic collisions, and electrocution were the main accident types.

• Deforestation, fire, and size of the urban area are the main correlates of accidents.

• Hindering deforestation and fires and adopting other actions should reduce accidents.

Abstract

Annually millions of animals are killed as a result of human-wildlife impacts. Each year the NGO Associação Mata Ciliar (NGOMC), in Southeastern Brazil, receives and rehabilitates thousands of animals. We evaluated how natural and anthropogenic characteristics affect the risk of different types of human-wildlife impacts for mammals that arrive at the NGOMC; and explore the relationship between both the animal's size and the type of human-wildlife impact event, survival rates and the likelihood that these animals can be fully rehabilitated. To test our hypotheses regarding the drivers and consequences of the total number of human-wildlife impact events, traffic collisions, electrocutions, and requested removals, we used records of the mammals that arrived at the NGOMC between 2012 and 2018, and obtained data on environmental attributes and anthropogenic factors at the municipality level, as well as species weights. The total number of human-wildlife impact events and of requested removals were both positively correlated with deforestation rate and urban area. The number of traffic collisions was positively related to the number of fires. Municipalities with larger urban areas were more likely to have at least one electrocuted mammal. Temporally, the number of fires two months before was positively correlated with the number of human-wildlife impact events. Traffic collisions and electrocutions more frequently resulted in the death of the animal, than did other events. Animals that died were heavier on average than those that remained in captivity or were successfully released back into the wild. We conclude that human-wildlife impact event rates should decline with lower rates of deforestation, less anthropogenic fires and the adoption of other specific measures to avoid both traffic collisions with fauna and electrocutions.

Introduction

Anthropogenic activities have contributed to the decline and loss of animal and plant species in almost all the terrestrial ecosystems of the Earth (Barnosky et al., 2011), and consequently contributed to a decrease in ecosystem resilience, functional integrity (Larsen et al., 2005) and the loss of associated ecosystem services (Valiente-Banuet et al., 2015). Habitat loss and degradation, overexploitation, pollution, invasive alien species and disease are the main drivers of defaunation (Dirzo et al., 2014; Pereira et al., 2012). However, other direct factors such as large, uncontrolled fires (Lewis, 2020), traffic collisions (Abra et al., 2019; Abra et al., 2021), and indirect factors, such as the expansion of urban centres without proper planning (McKinney, 2002) can add to these key drivers and amplify their effects, increasing the challenges associated with the mitigation of threats to biodiversity conservation (e.g. Grilo et al., 2010). Beyond this, each year millions of animals are killed globally as a result of accidents caused by human activities, such as road traffic collisions (Abra et al., 2019; Adania et al., 2017; Lester, 2015; Malo et al., 2004), electrocution on power lines (Hernández-Matías et al., 2015; Katsis et al., 2018; Lokschin et al., 2007) and burn injuries sustained in uncontrolled fires (Mendonça et al., 2015; Sgardelis and Margaris, 1992).

There are both conservation, practical, and ethical motivations to prioritize the prevention of such accidents. Firstly, many of these species are threatened with extinction (Hernández-Matías et al., 2015; Medici and Desbiez, 2012; Mumme et al., 2000; Pinto et al., 2020), such that the loss of these individuals represents a serious threat to already endangered populations (Diniz and Brito, 2013; Hernández-Matías et al., 2015; Medici and Desbiez, 2012). Secondly, there are human costs associated with these accidents including risks of injury or death in traffic collisions with wild fauna (Abra et al., 2019; Lester, 2015), potential damage to property and inconvenience as a result of power outages caused by electrocution events (FTI Consulting, 2016), and risk of disease transmission due to contact of wild animals with humans and domestic animals (Chapman et al., 2005; Curi et al., 2012; Thompson, 2013). Given this two-way nature, these types of accidents, together with other events such as the taking of livestock by large predators and retaliatory attacks on predators, are collectively referred to as human-wildlife impacts (Redpath et al., 2013). Finally, from an ethical viewpoint, each of these individual animals is a sentient being, and we should value its life, avoiding suffering when possible (Singer, 1990). Given this range of possible motivations, there is a need for effective strategies to minimize the likelihood of human-wildlife impact events occurring. In the case of road-traffic collisions and electrocutions, strategies commonly used include the suppression of vegetation along roadsides, use of speed bumps and speed management (Collinson et al., 2019; Lester, 2015), implementation of fences and wildlife crossing structures as underpasses, overpasses, and adaptation of drainage culverts and bridges (Malo et al., 2004; Polak et al., 2019; Abra et al., 2020), increased spacing and better insulation of power lines (Tintó et al., 2010), or alternatively installation of subterranean power lines (Richardson et al., 2017).

To ensure that mitigation measures are effective, their design requires a better understanding of the ultimate causes of the human-wildlife impact events (hereafter referred to as events). However, most studies have focussed on the proximate causes, such as travel velocity and visibility on roads (Collinson et al., 2019; Hobday and Minstrell, 2008; Lester, 2015), or the structure of power transmission lines (Dwyer et al., 2013; Tintó et al., 2010). Relatively few studies have attempted to evaluate the ultimate causes, including the ways in which natural and anthropogenic characteristics of the environment lead to increased or decreased probabilities of events occurring. Those that have examined such effects suggest that, for example, the probability of roadkill events is higher where roads pass through areas with more native vegetation (Freitas et al., 2015; Malo et al., 2004), lower in areas with intense road traffic (Grilo et al., 2015; Kreling et al., 2019), and that habitat structure has an effect, with a lower probability of roadkills in more open habitats (Kioko et al., 2015). Perhaps unsurprisingly, the density of both animals and electricity networks have been shown to be important predictors of the number of electrocution events (Katsis et al., 2018). A better understanding of how the probability of human-wildlife impact events caused by human activities varies over space and time is essential in order to design effective mitigation actions and plan for their implementation. Minimising the probability of such events not only has positive conservation impact, but also direct economic benefits, as treatments required to rehabilitate injured animals are expensive, and for those animals that cannot be fully rehabilitated, there will be permanent costs of their upkeep in captivity.

In this study, we use data on human-wildlife impact events held by the NGO Associação Mata Ciliar (hereafter, NGOMC), which receives and, when possible, rehabilitates thousands of animals each year in the state of São Paulo, in the southeast of Brazil (Fig. 1). We aim to: 1) relate spatial patterns of events of different types to natural and anthropogenic landscape characteristics, to evaluate how such characteristics may affect the risk of human-wildlife impact events occurring; and 2) determine survival rates after different types of events, as well as the likelihood that animals can be fully rehabilitated and released, as opposed to being kept in captivity, or not surviving their injuries and relate this to animal's size and to the event type. We also discuss the implications of our results in the context of actions that could help to prevent these types of human-wildlife impact events, as well as the effectiveness of rehabilitation of injured animals.