Beyond classical invasions hypotheses: Investigating the impact of agricultural landscape simplification on wild pig populations

Worldwide intensive agriculture has caused simplification of landscapes, reducing not only native vegetation cover but also landscape heterogeneity. Yet, how landscape simplification may affect biological invasion is uncertain. This is because classical hypotheses of biological invasion focus on single predictors, making contradictory predictions on its effects. Here, we explore combinations of predictors from classical hypotheses to investigate how landscape simplification affects one of the most harmful invasive species worldwide – the wild pigs (Sus scrofa). We rely on a hierarchical sampling design across 55 landscapes maximizing uncorrelated variation between landscape heterogeneity and native vegetation cover. By comparing Royle and Nichols models of wild pig abundance, we consider linear and non-linear, and simple, additive and interactive effects of three groups of predictors: landscape disturbance (changes in landcovers), landscape heterogeneity (changes in landcover diversity/configuration) and biotic interactions (predator presence/absence and changes in native species richness). All best-ranked models included an interactive effect between landscape heterogeneity and open vegetation cover. Our findings suggest no support to classical hypotheses of biological invasion. We find no support for either biotic control, native vegetation loss or a linear impact of landscape heterogeneity on wild pig abundance. Instead, results suggest a strong context dependence, with landscape heterogeneity effects depending on the quantity of open vegetation cover. These results challenge a general nature-based guideline to invasion control in agroecosystems and suggest the relevance that investigations embrace a mosaic view of landscapes and the complex relationship of predictors, giving support to context-specific guidelines to invasion control.