Weed seeds in exogenous organic matter and their contribution to weed dynamics in cropping systems. A simulation approach.

摘要

Nitrogen fertilizers are crucial for ensuring crop production. Mineral nitrogen fertilizer are often complemented or replaced (e.g. in organic agriculture) by exogenous organic matter (EOM) to limit environmental impacts of mineral fertilization (e.g. reduced carbon storage, nitrate leaching to groundwater, biodiversity erosion). EOM can though cause new problems as it comprises pest propagules, particularly weed seeds. The objective of the present work was to evaluate the impact of EOM on grass-weed dynamics, combining EOM analyses and model simulations. The first step analyzed seven EOM types for their chemical properties (e.g. nitrogen, carbon) and weed seed contents (using germination tests). The tested EOMs were of agricultural or urban origin (e.g. co-compost of green waste and sewage sludge or organic fraction of municipal waste). The most infested EOMs were composted farmyard manure and cattle slurry while fresh manure contained few weed seeds. Urban and green waste composts were nearly seed-free. In total, nineteen weed species and one crop species (Triticum aestivum) were identified with cattle slurry comprising the highest species number (thirteen). Weeds were mostly grass species (Lolium sp., Bromus sterilis, Echinochloa crus-galli, Festuca sp., Poa annua), except for Trifolium sp. In the second step, an existing model (AlomySys) which quantifies the effects of cropping systems on the dynamics of an autumnal grass weed similar to Lolium sp. was amended to account for EOM applications, considering both weed seed addition to the native soil seed bank and additional nitrogen. Finally, the amended model was used to simulate the effects on grass-weed dynamics of applying composted farmyard manure, using the observed EOM characteristics. Simulations focused on a small number of factors potentially interacting with EOM, i.e. EOM application frequency and seed content, tillage strategies, mineral nitrogen fertilization, and initial weed infestation. These simulations showed that for autumnal grass-weed species, adding weed seeds via EOM only increased multi-year weed infestation if fields were initially weed-free or if the simulated weed species was highly dormant in summer. Conversely, the additional EOM nitrogen reduced the fitness of the newly produced grass-weed seeds by affecting their dormancy and germination pattern, having more of them emerge in summer and thus unable to flower. Burying manure by mouldboard ploughing decreased or even cancelled the EOM effect.