About the job
Research Project
General objectives
The central objective of the PhD is to develop an integrated model coupling an Agent-Based Model (ABM) with a parsimonious hydrological-hydro-sedimentary model, embedded in a participatory process grounded in Mayotte’s catchments.
This framework aims to:
– Represent bidirectional feedbacks between social decisions and the physical state of the hydrological system;
– Reconstruct past vulnerability trajectories and identify future adaptation levers;
– Detect early-warning configurations associated with crisis regimes and tipping dynamics;
– Deliver a simulation-based deliberation and decision-support tool co-constructed with local stakeholders.
Study sites
The PhD focuses on the Ourovéni and Hajangua-Salim Bé catchments, selected for their representativeness of the dynamics at play in Mayotte: combined urban and rural pressures, agricultural water uses, ecological stakes, and governance complexity. Together, these two catchments offer sufficient diversity to test and generalise the approaches developed throughout the project.
Methodological architecture
a) Participatory Agent-Based Modelling (ABM)
The first component of the PhD involves co-constructing an agent-based model with local stakeholders — public institutions, water operators, municipalities, NGOs, farmers, and inhabitants — through companion modelling workshops. Agents represent heterogeneous decision-makers whose choices are shaped by water access, perceived risk, institutional rules, economic constraints, social norms, and crisis memory. Governance is represented through both formal arrangements and de facto practices, enabling the simulation of mismatches between policy intent and on-the-ground implementation. The Companion Modeling approach will be used to engage stakeholders to co-design the ABM [Barreteau et al., 2003; Etienne, 2014]. The agent-based model will be implemented using the GAMA platform [Taillandier et al., 2019], already used in many SES modeling (e.g. [Gaudou et al., 2013]).
b) Hydrological-hydro-sedimentary modelling
The second component relies on coupling the ABM with a parsimonious hydrological model, calibrated for the target catchments that represent rainfall-runoff dynamics, hydro-sedimentary pressures, and water availability and quality indicators. The coupling architecture is staged: a loose coupling is first implemented to stabilise variable exchange and calibration, after which dynamic coupling supported by a surrogate ML proxy model is progressively introduced to restore two-way feedback between society and the hydrosystem.
c) Simulation experiments
The coupled framework is used for two types of experiments. Retrospective simulations replay past crisis episodes and policy interventions to analyse underlying mechanisms (climatic variability, infrastructure limits, governance fragmentation, illegal extraction, demand growth). Prospective simulations explore adaptation trajectories under alternative climate and societal scenarios to identify robust, equitable, and effective combinations of adaptation measures.
Application
Applications should be submitted before the 14th of July 2026 to Kevin Chapuis and Benoit
Gaudou at kevin.chapuis@ird.fr and benoit.gaudou@ut-capitole.fr.
The application package must include:
– A detailed curriculum vitae;
– A cover letter (maximum 2 pages) outlining the applicant’s motivation and vision for
the project;
– Master’s degree transcripts (or equivalent).
For selected candidates, an interview will be organised quickly (in the week of the 13th of
July, if possible).
Requirements added by the job poster
• Master
