1. There is a long-standing relationship between humans and vertebrate scavengers, as scavengers’ contributions take on regulating (e.g. nutrient recycling, disease control), material (e.g. competition, livestock depredation) and non-material (e.g. sky burials, ecotourism) roles in society. A social-ecological approach to studying biodiversity is increasingly needed, since the inclusion of local perceptions and knowledge has proven critical for effective conservation programs and ecosystem management.

2. We examine livestock farmers’ perceptions and knowledge related to vertebrate scavengers in the highly diverse Chitwan-Annapurna Landscape (Nepal), and assess the sociodemographic traits that influence their perceived value of scavengers’ ecosystem services provisioning (ESP), and function via scavenging services (SS). 

3. Farmers’ perceptions of functional importance (SS) showed species-specific gradation, unlike ESP, where only avian scavengers were perceived as beneficial. Our results show that the perception of scavenging as a beneficial ecosystem service and its importance as a biological function are decoupled for facultative scavengers, and coupled for obligate scavengers. Relatedly, we identify that affluence-related traits drove positive perceptions of ESP, and local ecological knowledge-based traits were linked to increased knowledge of function via SS. 

4. Thus, this increased awareness of functional importance based on close contact with nature does not guarantee positive valuations of scavengers’ contributions, whereas formal education did influence positive perceptions despite reduced awareness of function. Additionally, our findings suggest that existing environmental education measures are targeting the right groups, as these respondents coincide with lower favorability of scavengers’ ecosystem services, but may be unable to overcome existing human-wildlife conflict. 

5. For the first time in South Asia, we survey relevant community stakeholder’s attitudes towards an entire scavenging guild and their associated benefits, detriments, and functional importance. Our study illustrates the varied perceptions that exist for different scavenger species, and closely examines a wide-ranging set of sociodemographic traits that show disparate influences on farmers’ knowledge of ecological function and perceived ecosystem service benefits. Crucially, these findings can guide conservation and management priorities by considering the differences in public perception and awareness of scavenging, as well as the interpretation of nature’s contribution to people. 

Between 2018 and 2019, we conducted 141 interviews with livestock farmers across the Chitwan-Annapurna Landscaoe of Central Nepal. In each of the three study areas, we selected 16-24 villages according to their accessibility, and based on communications with local governmental officials and community leaders. At each village, we approached 1-6 individuals that identified as keeping livestock by a combination method of random and snowball sampling (Cortés-Avizanda et al. 2018, García-Alfonso et al. 2019). All data was collected by hand in the field, and then the corresponding author manually digitized all responses into corresponding entries within Microsoft Excel. 

Our fieldwork, including survey design and methodology, was conducted with the approval of The City University of New York’s Human Research Protection Program (HRRP) under the category of Human Subject Research (IRB File #2019-0413). In addition, we also received approval for our survey methodology and fieldwork from Nepal’s Department of National Parks and Wildlife Conservation. We sought verbal informed consent before proceeding with the survey, rather than written consent, due to variability in literacy rates, and farmers’ comfort with reading written documents and ability to sign. In accordance with the guidelines of our institution’s HRRP and Institution Review Board, we first read a pre-approved oral consent script aloud to participants that explained the purpose of our study, our local collaborators, and the nature of questions. Participants were assured that their identities would remain anonymous, and no personal identifiers would be recorded from the information collected. 

Analyses were conducted using R software 3.3.1 (R Core Team 2016) with ‘glm’ from the stats package for “univariate” (with fixed factor) models, and ‘glmulti’ from the glmulti package (version 1.0.7.1) for multivariate model selection (Calcagno and de Mazancourt 2010).

Missing values (e.g., farmer was not asked about species as it was not included in the survey for a given survey area, farmer did not respond to the specific question) are designated as "NA". All analyses omitted NAs, unless otherwise specified in the manuscript.