Séminaire de Jonathan LENOIR (Université de Picardie - Jules Verne)


Le mercredi 22 mars 2017 à 12h45, salle de conférences de l'OSUR, bâtiment 14b, Campus de Beaulieu, UR1

Le mercredi 22 mars 2017 à 12h45, salle de conférences de l'OSUR, bâtiment 14b, Campus de Beaulieu, UR1

Integrating advanced remote sensing data into an invasive species distribution model (iSDM) coupled with a population model to predict the invasion dynamic of a non-native plant at fine spatial resolution

Species distribution models (SDMs) are important tools to investigate changing distributions and to propose efficient management strategies. However, to ensure robust predictions and efficient management decisions it is important to model both the potential and realized distributions, especially so for non-native invasive species which distribution are not in equilibrium with the environmental conditions within the invaded range. To address this issue, I will present iSDM: a novel framework for invasive SDMs comprising an environmental systematic sampling design to optimally collect presence-absence data from the field and a probability index to sort and subsequently separate environmental absences (EAs: reflecting environmentally unsuitable sites) from dispersal-limited absences (DLAs: reflecting sites out of dispersal reach). This framework helps overcoming the conceptual and methodological limitations of the disequilibrium in SDMs inherent to non-native invasive species and enables managers to robustly estimate both the realized and potential distributions within the invaded range. Once the realized and potential distributions are known, the challenge is to incorporate local population dynamics at very fine spatial resolution in order to mechanistically forecast or hindcast the invasion process. Here, we propose to use advanced remote sensing data at fine spatial resolution and to couple iSDM with mechanistic simulations of dispersal, local establishment and population growth to address this challenge. Within the FP7, ERA-NET, BiodivERsA funded project DIARS (Detection of Invasive plant species and Assessment of their impact on ecosystem properties through Remote Sensing), I will present an application based on airborne LiDAR data and which aims at predicting the invasion dynamic of a non-native tree species (Prunus serotina Ehrh.) in the forest of Compiègne. By reconstructing the invasion dynamic of this long-lived tree species with a complex life cycle, I will present the various components of our hybrid model, input data and first results.

Contact : Jonathan LENOIR