Combining expert knowledge and models in participatory workshops with farmers to design sustainable weed management strategies

International audienceCONTEXT: In order to design cropping systems reconciling crop production, biodiversity and reduced herbicide use, we organised participatory workshops with farmers in the Champagne region (North-Eastern France).OBJECTIVE: The main issues of the study were to develop and implement a methodology for participatory workshop-based cropping-system design, combining expert knowledge and models, in the particular case of agroecological weed management.METHODS: Methods used in this study combined cropping-system prototyping by farmers, expert opinion and models. In a first meeting, farmers determined their objectives and constraints, and chose a reference system from one of their farms (oilseed rape/winter wheat/winter wheat/spring barley heavily infested by autumnal grass weeds). In a second meeting, two sets of prototypes were designed by two separate groups, using the Mission Ecophyt'eau (R) tool as support. The reference and the prototypes were evaluated in terms of (1) technical feasibility from farmers' and scientists' expertise, (2) weed harmfulness for crop production and weed contribution to biodiversity with FLORSYS which simulates crop and weed growth and reproduction in cropping systems at a daily time step over several years, and (3) economic, social and environmental sustainability with the DEXiPM model. Steps 2 and 3 were carried out after the meeting. At a third meeting, these results were presented to the farmers who continued improving the prototypes, using the DECIFLORSYS model which includes decision trees to guide farmers during cropping-system design and a fast calculator estimating weed harmfulness and biodiversity of cropping-system prototypes.RESULTS AND CONCLUSIONS: Both prototypes presented increased crop diversification, introducing legumes and other broad-leaved crops, crop mixtures and cover crops. Both prototypes reduced weed harmfulness for crop production (yield loss, harvest contamination, field infestation) and herbicide use intensity (approx. -60%) compared to the reference system. The best solution was a suboption of prototype B replacing most of the herbicides by mechanical weeding, dividing yield loss by nearly two compared to the reference and improving biodiversity by 5-10%. The workshop participants appreciated the knowledge on agroecosystem functioning and the complementarity of models. DECIFLORSYS allowed a direct evaluation during workshops. FLORSYS produced a detailed diagnosis of the technical and meteorological causes of the cropping systems' performance. DEXiPM assessed working times and economic viability of the prototypes.SIGNIFICANCE: Following the workshops, some participants invested in new tools for mechanical weeding and introduced more spring crops into their rotations