Habitat loss and degradation can undermine wildlife communities and ecosystem functioning. However, certain generalist wildlife species like mesopredators and omnivores can exploit these disturbed habitats, sometimes leading to population increases (e.g. ‘mesopredator release’ in degraded areas). Although mesopredator release may cause negative effects on food webs and zoonotic disease management, some disturbance‐tolerant species may help perpetuate important ecological interactions, such as seed dispersal. We evaluated the habitat associations of common palm civets Paradoxurus hermaphroditus, which are widespread generalist mesopredators in Southeast Asia. Common palm civets are also high‐quality seed dispersers, and potential zoonotic disease hosts. We used published and new camera trapping data to map their probability of presence across Southeast Asia and evaluate regional‐scale associations between capture rates and habitat variables such as elevation, ecoregion intactness and Human Footprint Index, among others. We also assessed the influence of habitat variables on their relative abundance at the local scale. At the regional scale, we found that common palm civets showed significant positive associations with landscapes characterized by lower ecoregion intactness, higher Human Footprint Index and lower elevations. At the local scale, their relative abundance showed a significant positive association with higher Human Footprint Index, but only to a certain point, after which it started decreasing. They also favoured lower elevations at the local scale. These multi‐scale results indicate that common palm civets' abundance can increase under certain levels of human disturbances, consistent with the ‘mesopredator release’ hypothesis. This suggests they may be crucial seed dispersers in degraded forest landscapes, especially where more sensitive seed dispersers have disappeared. Our results are also consistent with previous studies reporting that habitat degradation increases populations of potential zoonotic disease hosts, and thus risks of transmission to humans.

The “trophic downgrading of planet Earth” refers to the systematic decline of the world’s largest vertebrates. However, our understanding of why megafauna extinction risk varies through time and the importance of site- or species-specific factors remain unclear. Here, we unravel the unexpected variability in remaining terrestrial megafauna assemblages across 10 Southeast Asian tropical forests. Consistent with global trends, every landscape experienced Holocene and/or Anthropocene megafauna extirpations, and the four most disturbed landscapes experienced 2.5 times more extirpations than the six least disturbed landscapes. However, there were no consistent size- or guild-related trends, no two tropical forests had identical assemblages, and the abundance of four species showed positive relationships with forest degradation and humans. Our results suggest that the region’s megafauna assemblages are the product of a convoluted geoclimatic legacy interacting with modern disturbances and that some megafauna may persist in degraded tropical forests near settlements with sufficient poaching controls.

Re‐establishing extirpated wildlife—or “rewilding”—is touted as a way to restore biodiversity and ecosystem processes, but we lack real‐world examples of this process, particularly in Southeast Asia. Here, we use a decade of aggregated camera trap data, N‐mixture occupancy models, and input from local wildlife experts to describe the unassisted recolonization of two native large herbivores in Singapore. Sambar deer (Rusa unicolor) escaped from captivity (in private or public zoos) in the 1970s and contemporary camera trap data show they have only colonized nearby forest fragments and their abundance remains low. Wild pigs (Sus scrofa), in contrast, naturally recolonized by swimming from Malaysia in the 1990s and have rapidly expanded their range and abundance across Singapore. While wild pigs have not recolonized all viable green spaces yet, their trajectory indicates they soon will. We also note that a third ungulate, the muntjac deer (Muntiacus muntjak), was captured in camera trapping in 2014 and 2015 but was never recorded afterward despite increased sampling effort, and thus we do not focus on their presumably unsuccessful recolonization. The divergent rewilding trajectories between sambar deer and wild pigs suggest different conservation outcomes and management requirements. Sambar deer may restore lost plant–animal interactions such as herbivory and seed dispersal without requiring significant management. Wild pigs, in contrast, have reached high numbers rapidly and may require active management to avoid hyperabundance and negative ecological impacts in regions, such as Singapore that lack both hunting and large predators.

Habitat loss and degradation threaten forest specialist wildlife species, but some generalist mesopredators exploit disturbed areas and human‐derived food, which brings them into closer contact with humans. Mesopredator release is also important for human health for known zoonotic disease reservoirs, such as Asian civets (Viverridae family), since this group includes the intermediator species for the SARS‐CoV‐1 outbreak. Here we use camera trapping to evaluate the habitat associations of the widespread banded civet ( Hemigalus derbyanus ) across its range in Southeast Asia. At the regional scale, banded civet detections among published studies were positively associated with forest cover and negatively associated with human population. At the local scale (within a landscape), hierarchical modeling of new camera trapping showed that abundance was negatively associated with forest loss and positively associated with distance to rivers. These results do not support mesopredator release and suggest a low likelihood overlap with humans in degraded habitats and, therefore, a low risk of zoonotic disease transmission from this species in the wild. We also estimate that banded civet distribution has contracted to under 21% of its currently recognized IUCN Red List range, only 12% of which falls within protected areas, and a precipitous recent decline in population size. Accordingly, we suggest the banded civet's Red List status should be re‐evaluated in light of our findings.

Predator–prey dynamics are a fundamental part of ecology, but directly studying interactions has proven difficult. The proliferation of camera trapping has enabled the collection of large datasets on wildlife, but researchers face hurdles inferring interactions from observational data. Recent advances in hierarchical co‐abundance models infer species interactions while accounting for two species' detection probabilities, shared responses to environmental covariates, and propagate uncertainty throughout the entire modeling process. However, current approaches remain unsuitable for interacting species whose natural densities differ by an order of magnitude and have contrasting detection probabilities, such as predator–prey interactions, which introduce zero inflation and overdispersion in count histories. Here, we developed a Bayesian hierarchical N‐mixture co‐abundance model that is suitable for inferring predator–prey interactions. We accounted for excessive zeros in count histories using an informed zero‐inflated Poisson distribution in the abundance formula and accounted for overdispersion in count histories by including a random effect per sampling unit and sampling occasion in the detection probability formula. We demonstrate that models with these modifications outperform alternative approaches, improve model goodness‐of‐fit, and overcome parameter convergence failures. We highlight its utility using 20 camera trapping datasets from 10 tropical forest landscapes in Southeast Asia and estimate four predator–prey relationships between tigers, clouded leopards, and muntjac and sambar deer. Tigers had a negative effect on muntjac abundance, providing support for top‐down regulation, while clouded leopards had a positive effect on muntjac and sambar deer, likely driven by shared responses to unmodelled covariates like hunting. This Bayesian co‐abundance modeling approach to quantify predator–prey relationships is widely applicable across species, ecosystems, and sampling approaches and may be useful in forecasting cascading impacts following widespread predator declines. Taken together, this approach facilitates a nuanced and mechanistic understanding of food‐web ecology.