post-doc position on spatial modelling of ecological networks

When:
25/12/2019 – 26/12/2019 all-day
2019-12-25T01:00:00+01:00
2019-12-26T01:00:00+01:00

Annonce en lien avec l’Action/le Réseau : aucun

Laboratoire/Entreprise : LBBE Lyon / LECA Grenoble
Durée : 2 years
Contact : vincent.miele@univ-lyon1.fr
Date limite de publication : 2019-12-25

Contexte :
The study of species interactions, modelled by ‘networks’, is crucial to understand the functioning of ecological communities and their resilience to global changes. Nowadays, the ever-increasing availability of multiple ecological networks sampled at different spatial locations (see Kortsch et al) allows for studying species interactions at large spatial scales. New questions arise on the variability and plasticity of species interactions in space. This variability is necessarily driven by species turnover that induce changes in network composition (i.e. species/nodes identity can change between different locations) but not always in network structure. In other words, if we are able to characterize the internal structure (or the “shape”) of the ecological networks, it could be possible to compare this structure along spatial gradients putting aside species identity. For instance, two networks with different species can have a similar shape because they share common ecological compartments (e.g. in food webs, see Ohlmann et al).

Kortsch et al, Food‐web structure varies along environmental gradients in a high‐latitude marine ecosystem, Ecography 42 (2), 295-308 (2018)
Ohlmann et al, Diversity indices for ecological networks: a unifying framework using Hill numbers, Ecology Letters 22, 737-747 (2019)

Sujet :
The postdoc project aims at developing a new mathematical framework to study the spatial process driving the variations of network structure. It will rely on two intertwined objectives. The first one consists in proposing the most appropriate way to characterize/measure and compare network structure (that is, converting the network data into quantitative therefore comparable information). Different approaches could be considered, including (but not only):
• machine learning-based techniques, such as nodes/network embedding techniques (see Hamilton et al)
• statistics-based frameworks, such as network models (see Kéfi et al) or network statistics (e.g. beta-diversity, see Ohlmann et al).
The second objective consists in modelling the spatial process that drives changes observed in the light of the aformentionned structure measure. This process will ultimately integrate spatial information (coordinates, ecological barriers,…) and environmental variables (e.g. climate or landscape configuration). In the end, the implemented framework will allow for mapping the biogeography of the internal structure of networks.

Hamilton et al, Representation learning on graphs: Methods and applications, arXiv preprint arXiv:1709.05584 (2017)
Kéfi et al, How structured is the entangled bank? The surprisingly simple organization of multiplex ecological networks leads to increased persistence and resilience, PLoS biology 14-8 (2016)

Profil du candidat :
Motivation to work in an interdisciplinary team.

Formation et compétences requises :
PhD in ecology or machine learning or statistics or network science, with a strong interest for at list a second element in this list.

Proficiency in at least one programming language (Python, R and/or C++).

Adresse d’emploi :
Lyon (LBBE, campus de la Doua, Lyon) or Grenoble (LECA, Univ. Grenoble Alpes) depending on the candidate’s preference

Document attaché : postdoc-econet2019.pdf