The response of weed and crop species to shading: Which parameters explain weed impacts on crop production?

Crops compete with weeds for light, and choosing competitive crop species contributes to managing weeds. The objective was to identify which crop and weed parameters related to competition for light drive weed harmfulness for crop production. In a previous experiment, we measured parameters to characterize species potential plant morphology in unshaded conditions and species response to shading for a range of 60 crop and annual weed species. Here, we integrated the measured parameter values into an existing simulation model that uses an individual-based 3D representation of crop-weed canopies to predict weed dynamics and crop production from pedoclimate and cropping system information. The model, i.e. FlorSys, was used to run virtual experiments in seven French and Spanish regions, with 272 cropping systems varying in terms of crop rotations, herbicide use and tillage intensity etc. A series of statistical methods (RLQ, fourth corner analysis, Principal Component Analysis, Pearson correlation coefficients, analysis of variance) were used to identify the key weed and crop parameters that drive crop yield loss and other weed harmfulness indicators. The weed species that caused the highest yield loss had a large leaf area at emergence. When young, they presented a large specific leaf area and a uniform leaf area distribution along plant height. They were also taller per unit plant biomass and their populations were more homogeneous in terms of plant width. At later stages, harmful weed species presented a smaller interception area to herbicides, with thicker leaves located lower on the plant. When shaded, harmful weed species shifted their leaves upwards and decreased their plant width per unit biomass. Weed-suppressive crop species had a large specific leaf area, wider plants per unit biomass, and a uniform leaf area distribution along plant height. When shaded, they increased their plant height and width per unit biomass. There was a trade-off between parameters driving potential crop production and those minimizing weed-inflicted yield losses.