Three key directions toward the development of a holistic crop modelling framework

Diversified farming practices—like intercropping, riparian buffers, organic amendments, and integrated pest management—are considered reliable strategies to enhance global food production while reducing negative environmental impacts of fertilizers, pesticides, and other practises associated with modern agriculture. In the 1960s, monocultures, agrochemicals, and high-yielding crop varieties revolutionized global food production. Since then, crop and farm simulation models have guided field, farm, and watershed management. Because most of these models were developed within a management paradigm centered on monocropping, it remains unclear whether they adequately represent fundamental processes and mechanisms underlying diversified farms. Our exploratory review of 12 models commonly used to predict crop yield shows that most models lack mechanisms indispensable for that purpose in diversified farms, like positive interactions among plants or between plants and other organisms (e.g., pollinators, fungi). Most models oversimplify weed invasions and focus on isolated fields without landscape contexts. We propose a strategy to improve modelling diversified farms, based on three complementary questions that such models should be able to address: 1) Under which conditions intercropping increases long-term yield and reduces weed pressure? 2) How can the spatial and temporal configuration of diversified agricultural practices be optimized to enhance pollination, weed and pest suppression, water retention, and nutrient uptake? 3) how can ecosystem-based solutions be redesigned and optimized to reduce phosphorus legacies? We discuss how integrating principles from agronomy, ecology, and evolutionary biology into crop models can address these questions, offering insights for soil, crop, and farm modelling and management in diversified farming systems. • Diversified agronomic management practices could help address key environmental issues. • Most crop models were developed under paradigms inconsistent with these management practices. • Models focusing on crop, ecology, evolution, and others can complement each other. • Three research questions explore ways to integrate them to better assess diversified practices.