Toward systems agroecology: Risk-reward balance, emergent plant communities, and temporal weather map in multiplant farming.

Although monoculture is a common practice, its fragility, high input requirements, and significant environmental impact underscore the need to seek alternatives. Polyculture, inspired by naturally occurring plant communities, could be one such alternative. Unfortunately, our understanding of how plant communities respond to human intervention while being subject to weather variations remains limited. Here, using data from long-term experiments, we demonstrate that the effects of soil treatment and weather conditions on biomass production can be dissociated. The choice of soil treatment modifies not only the average yields of multiplant systems (i.e., rewards) but also their sensitivity to weather variations (and thus the risks). Analysis of botanical survey data also reveals that observed plant communities can be grouped into distinct types, differing in their species composition. It is noteworthy that communities of different types occupy distinct regions in the risk-reward space. We also show that certain soil treatments can lead to higher yields without increasing weather sensitivity, while preserving plant biodiversity. Using daily meteorological data, we created an agricultural temporal "weather map" identifying distinct zones favorable or unfavorable to biomass production. Our analysis demonstrates the importance of comprehensive data for establishing the principles of systems agro-ecological design. Faced with the increasing environmental risks associated with intensive monoculture, such principles could prove invaluable for developing less risky polyculture alternatives.