Shifts from complementarity to selection effects maintain high productivity in maize/legume intercropping systems.
Complementarity (CE) and selection effects (SE) have been either invoked to explain the positive diversity-productivity relationship in natural and semi-natural ecosystems. Few studies have, however, separated selection and complementarity effects in economically valuable intercropping systems, which receive significant nutrient inputs throughout the growing season. We performed a 2-year field experiment with five cropping systems (i.e. maize/peanut intercropping, maize/soybean intercropping, maize, peanut and soybean monocultures) under different combinations of nitrogen (N) and phosphorus (P) fertilization. Sequential harvest of subplots was performed five times during the growing season and the additive partitioning method was applied to determine complementarity and selection effects in the two intercropping systems. We found that the land equivalent ratio (LER) based on yield was greater than or close to 1 in both maize/soybean and maize/peanut intercropping systems, suggesting yield advantages in these intercropping systems compared to monocultures. The LER of the two intercropping systems without N fertilization was greater in 2018 than in 2017. Nitrogen fertilization reduced the LER of maize/soybean intercropping in 2018, and that of maize/peanut intercropping across 2 years. Nitrogen fertilization had much stronger effects on maize yield than on soybean or peanut yield. Intercropping increased maize yield but decreased soybean and peanut yield. The CE was more important for net biodiversity effects without N fertilization while the SE was more important with N fertilization in the two intercropping systems across 2 years. Phosphorus fertilization impacts on the LER and biodiversity effects were weaker compared to N fertilization. Finally, net biodiversity effects were greater at later growth stages within a growing season. Synthesis and applications. Our study demonstrates how yield advantages of species-diverse intercropping systems can be maximized either through (a) CE in the absence of N fertilization thanks to complementarity in nutrient use between crops or (b) SE under N fertilization thanks to overyielding of highly productive species (i.e. maize). Yield advantages promoted by CE are more sustainable because they are associated with reduced chemical fertilizer use and greater monetary benefits to farmers. These findings have important implications for the design and management of species-diverse intercropping food-production systems.