Using spatially-explicit plant competition models to optimise crop productivity in intercropped systems

Abstract

Intercropping, by capitalizing on positive biodiversity–productivity relationships, represents a promising option to increase agricultural sustainability. However, the complexity and context-dependency of plant–plant interactions can make it challenging for farmers to find suitable crop combinations. Furthermore, intercropping is usually implemented with standard inter-row spacing and plant densities based on monoculture practices, which might not be the ideal configuration to maximize yield.Here we present a spatially-explicit yield analysis method based on plant ecological interaction models that allowed to optimize crop species combinations and spatial configurations for maximal yield in intercropped systems. We tested this method with three crop species, namely oat, lupine, and camelina. In a first step, field experiments in which crop density and adjacent crop type were varied provided us with indications on which species would compete more with each other. The results showed us that oat and camelina strongly competed with each other. In addition, the distance experiments allowed us to understand how the changes in yield associated with the presence of neighbors vary with distance. This allowed us to find the sets of parameters (identity of neighbors, sowing density, distances between individuals) that optimise intercrop yield (measured as Land Equivalent Ratio [LER]) for the three considered species. Specifically, we show that alternating rows of species led to higher LERs than a homogeneous species mixing, and that 3-species combinations are not necessarily more performant than the best 2-species combinations. In addition, we show that increasing the density of oat is generally beneficial for LER, while increasing the density of lupine is not.By modelling crop yield from simple and reproducible density and distance experiments, our results allow to optimize crop mixtures in terms of species combinations and spatial configurations, for maximal crop yield.