Relational structure of illegal wildlife hunting in China: a nationwide hunter–prey network analysis
Description
Illegal wildlife hunting continues to pose a major biodiversity threat in China, yet there remains no systemic relational understanding of the way in which perpetrators are linked to key taxa. To address this, here we provide a novel framework for understanding and addressing the systemic roots of wildlife crime. To characterise the structure of hunter–prey interactions in China, we applied bipartite network analysis, a relational framework not previously applied to nationwide wildlife crime prosecution data, to 6,379 poaching case prosecution records (2014–2020). Results revealed that illegal hunting forms significantly nested, non-random networks at national and provincial scales, producing a structured socio-ecological system. Offenders were overwhelmingly males, with those aged 31–50 with primary to junior middle school education predominating and hunting the broadest prey spectrum. Provincial socio-economic context, particularly regional wealth, shaped these hunter-prey networks, increasing nestedness in mammal poaching. Key prey families (e.g., Phasianidae) were associated with a broad range of hunter groups (defined by age and educational attainment) within the network. Our findings show that poaching forms a structured, non-random network, suggesting limitations to uniform enforcement approaches in China. This nested structure is consistent with a clear core–periphery configuration, in which a small number of prey families are targeted by many hunter groups, while other prey families are embedded within narrower, more specialized hunting relations. More broadly, effective conservation must integrate demographically targeted interventions, network-informed protection of core prey, and regionally tailored policies to disrupt the structure of hunter-prey networks.
Files
Steps to reproduce
To reproduce this research, one must first query the China Judgments Online database (https://wenshu.court.gov.cn/) using the Chinese terms for “wild animal,” “illegal hunting,” and “illegal catching” to retrieve all relevant judgment documents from 1 January 2014 to 31 December 2020. From the initial 10,937 records, documents are manually reviewed—extracting defendant demographics (age, education), prey species, and case details—and excluded if they do not involve birds or mammals, lack hunter age or education, lack prey family information, or are retrials, yielding a final analytic sample of 6,379 cases. Hunters are classified into 20 groups (5 age classes × 4 education levels) and prey into 68 bird families and 25 mammal families, with all interactions converted to binary presence (1)/absence (0). Provincial socio‑economic variables (e.g., GDP, income, education) are obtained from the China Knowledge Resource Integrated Database, prey trait data (body mass, IUCN status, protection status) from public repositories (IUCN Red List, Chinese museum database, Etard et al. 2020), and phylogenetic trees from VertLife. Network construction, nestedness calculation (NODF metric, Curveball null models with 1,000 iterations), and visualisation are performed in R using the bipartite package; spatial regression (Moran’s I, Lagrange multiplier tests, spatial lag/error models) uses GeoDa 1.22; non‑parametric tests (Scheirer–Ray–Hare, Kruskal–Wallis, Dunn’s post hoc) use the FSA package; and phylogenetic generalised least squares models use the caper package.