Modeling the dynamics and emergence of a multispecies weed seed bank with species traits

Models quantifying the effects of cropping systems on weed life-cycles, in interaction with environmental conditions, are invaluable tools to evaluate and design integrated cropping systems. In the past, we have developed such a model called AlomySys for a single grass weed species. The objective of the present paper was to extend the monospecific emergence model to a multi-specific weed flora, by removing a few non-generic processes (e.g. maternal effects on dormancy and germination) and by adapting the germination sub-model to a larger diversity of species.The model is based on a daily simulation of weed seedbank mortality, seasonal dormancy variations, germination and pre-emergent growth until seedling emergence, in interaction with cropping systems and environmental conditions. In order to parameterize the model for numerous and contrasting species, we developed an innovative approach consisting in estimating 25 parameters from easily accessible species traits or from empirical expert knowledge. Relationships were established between seed traits (e.g. seed coat thickness, mass or shape) and seed mortality rates, dormancy levels, speed of germination as well as pre-emergent mortality. Base temperature for germination and dates of dormancy variations were related to usual dates of the species emergence in fields, obtained from expert knowledge. The generic model structure and the trait–parameter relationships make it possible to simulate a wide diversity of weed species and to predict their response to a large range of cropping systems. Thanks to a daily time-step and the integration of the interactions with soil and climatic conditions, FlorSys permits to optimize the choice of cultural techniques for weed management, as a function of field characteristics and previous technical operations. For instance, the model was used to compare the emergence of different weed species depending on tillage, in interaction with the preceding technical operations.

► Modeling the effects of cropping systems on weed dynamics for integrated weed control.
► Development of a generic weed dynamics model called FlorSys.
► Proposition of a new approach to parameterize this model for a diverse weed flora.
► Parameters can be estimated from species traits or expert knowledge.
► This approach will help to understand the roles of traits in community assembly.