Publications

Weed suppression in annual crop species mixtures

Gu, Chunfeng

Summary

Weeds are a major threat to crop production in agriculture. If left uncontrolled, weeds have a potentially large adverse effect on crop production through competition with crops for limiting light, water, and nutrients. Intercropping, as a spatial diversification measure, has shown potential for ecological weed management in many cases. This thesis aimed to investigate the extent to which weed suppression can be achieved by annual intercrops consisting of two main crop species and to explore the possible underlying mechanisms for weed suppression.

A global meta-analysis showed that intercropping is generally a useful approach for species with poor weed suppressiveness to improve weed management. On average, weed biomass in intercrops was substantially (58%) lower than that in weaker weed-suppressive sole crops, and it was similar to that in stronger weed-suppressive sole crops. This pattern emerged across species combinations such as maize/legume and small-grain cereal/legume intercrops. An additive design gave stronger weed suppression than a replacement design, showing the important role of plant density on weed suppressiveness. Spatial arrangement mattered in replacement systems: a mixed pattern gave stronger weed suppression than a row pattern. Weed suppressiveness of intercrops was not affected by being either simultaneous or relay systems, and it was not affected by the level of nitrogen fertilizer input. Three models (i.e., arithmetic mean, weighted arithmetic mean, and weighted harmonic mean) were developed to predict weed biomass in intercrops. The model valuation was used to explore the relevance of possible mechanisms (i.e., complementarity, selection, and plant density) underlying weed suppression in intercrops. The arithmetic mean and weighted arithmetic mean resulted in a systematic overestimation of the weed biomass in intercrops. The weighted harmonic mean was able to provide a good prediction of the observed weed biomass in intercrops. This model reflects how much weed biomass to expect based on the relative densities of the species in a mixture and species-specific weed suppression in pure stands, therefore the model does not account for the contribution of complementarity between intercropped species to weed suppression. The good prediction suggests that except plant density effect, selection of the stronger weed-suppressive species is an important mechanism while complementarity is of less importance for weed suppression. For intercropping with a less intimate intermingling of the two component species, either due to temporal (relay intercrops) or spatial (strip intercrops) separation, the harmonic mean tended to underestimate observed weed biomass. In the subsequent meta-analysis, findings based on crop yield or biomass data supported the important role of the selection effect on weed suppression in intercrops. The net effect ratio (NER) of the strongly weed suppressive species was 1.33 times as large as the NER of the weaker weed suppressive species, showing the occurrence of species competitive dominance. A better weed-suppressive ability of a species in the pure stand was associated with a better competitive position in intercrop. Species dominance was stronger in intercrops with a mixed configuration than in intercrops with a row or strip arrangement. Species dominance occurred in small-grain cereal/legume intercrops but not in maize/legume systems. Differences in species’ weed-suppressive abilities resulted in stronger species dominance in additive designs than in replacement designs. In addition to the analysis of literature data, I conducted a preliminary analysis using functional structural plant (FSP) modelling. Simulations from FSP modelling revealed that a selection effect of species with stronger weed suppressiveness occurs particularly during early growth, resulting in weed suppression in intercrops that is better than the average of sole crops

Overall, this thesis showed that intercropping is a promising strategy for crop species with poor weed suppressiveness to increase “within-season” weed suppression. Findings confirm the important role of plant density in weed suppression. In intercrops with close interaction between species, a selection effect steered by species with stronger weed suppressiveness is an important mechanism underlying weed suppression. In such cases, the weighted harmonic mean model provides a good prediction of the weed biomass. Future research should focus on the competitive balance between intercropped species and between crop and weed species in order to obtain a satisfying yield and weed suppression in intercrops.