Energetic criteria to predict biodiversity loss and keystone species in empirical agricultural food webs

Understanding how biodiversity loss propagates through food web ecological networks is essential for managing ecosystem services in agricultural landscapes. Traditional binary-topological approaches often oversimplify food web extinction dynamics by assuming species go extinct only after the complete loss of resources. In this study, we applied energetic thresholds to model secondary extinctions in 504 empirical food webs from the Farm Scale Evaluations (FSE) of conventional and genetically modified herbicide-tolerant (GMHT) crop systems. By simulating species removals under varying levels of consumer susceptibility to resource loss, we quantified food web robustness across crop types and management regimes. Our results show that agricultural food webs become increasingly fragile with higher energetic thresholds, particularly in GMHT-managed beet fields. Keystone species shift from autotrophs to gastropods as energy sensitivity increases, revealing the limitations of binary-topological food web analyses in capturing real-world extinction dynamics. Moreover, omnivory, especially carabid-gastropod trophic links, enhances food web stability by reducing secondary extinctions. These findings underscore the importance of incorporating energetic considerations to more accurately predict biodiversity loss and identify critical species in agroecosystems. Our approach offers valuable insights for designing resilient agricultural systems in the face of environmental change.