Skip to Main content Skip to Navigation

Modélisation de la dynamique des adventices dans un agroécosystème

Abstract : Many species have a dormant stage in their life cycle, such as seeds for plants. These species have different types of survival strategies. In particular, plants are known have survival strategies dependent on dormancy and dispersal of seeds. The metapopulation model, which does not consider a dormancy stage and is often used to analyse a species' dynamic, applied to a species which undergoes dormancy can lead to wrongly declare extinction in a patch where dormant individuals can still be present. In order to include dormancy in a model it is preferable to use hidden variables to model dormant individuals as they are often unobservable. Several Markovian models with hidden variables have already been proposed to study species with hidden stages. However, they all have different limitations : only presence/absence observations are modelled ; the dormancy stage is limited to one year or colonisation from neighbour patches is not taken into account. We propose a hidden Markov model with data feedback which describes the local and regional dynamics of a species with hidden stages where only observables stages may influence other patchs. The model allows species to undergo potentially time infinite dormancy using abundance classes. One would expect estimation, restoration and prediction of the next non-dormant populations to have an exponential computational time in terms of patches, however we have demonstrated that estimation, restoration and prediction are all achievable in a linear in terms of patches. The regional dynamic is modeled using the indistinguishable influence of neighbour non-dormant populations states on a dormant or non-dormant population. Numerical experiments on simulated data show that the state dormant populations can easily be retrieved as well as the future non-dormant populations' state. Results on weed species highlight that the state of the seed bank is mostly influenced by seed survival. Our framework provides a simple and efficient tool that could be further exploited to analyse and compare annual plants' dynamics, like weeds survival strategies in crop fields and even for species with hidden stages.
Document type :
Complete list of metadata

Cited literature [97 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Wednesday, September 2, 2020 - 2:35:09 PM
Last modification on : Wednesday, May 12, 2021 - 8:09:21 AM
Long-term archiving on: : Wednesday, December 2, 2020 - 4:35:36 PM


Files produced by the author(s)


  • HAL Id : tel-02786981, version 1
  • PRODINRA : 470661



Sebastian Le Coz. Modélisation de la dynamique des adventices dans un agroécosystème. Modélisation et simulation. Université Paul Sabatier - Toulouse III, 2019. Français. ⟨NNT : 2019TOU30034⟩. ⟨tel-02786981⟩



Record views


Files downloads