31/03/2026 14:00 - Amphi Cauchy, Bâtiment Carnot, ENPC, Champs-sur-Marne

Description

 Titre de la thèse : Exposure to urban air pollution and daily mobility: is model complexity worth it? Insights from an agent-based approach

·  Lien visio :

https://teams.microsoft.com/meet/37420088244407?p=n1oMuX7LRlmupQLgbn

Numéro de réunion : 374 200 882 444 07

Code secret : Nz98ZF66

·  Composition du jury

• Rafael BORGE GARCíA (rapporteur), professeur des universités, Universidad Politécnica de Madrid
• Moez KILANI (rapporteur), professeur des universités, LEM, Université du Littoral Côte d’Opale
• Isabelle COLL (examinatrice), professeure des universités, LISA, Université Paris Est Créteil
• Ouassim MANOUT (examinateur), chargé de recherche, LAET, École nationale des travaux publics de l'État
• Sandrine MATHY (examinatrice), directrice de recherche, GAEL, CNRS
• James WOODCOCK (examinateur), professeure des universités, University of Cambridge
• Yelva ROUSTAN (directeur de thèse), directeur de recherche, CEREA, Ecole Nationale des Ponts et Chaussées
• Nicolas COULOMBEL (co-directeur de thèse), chercheur IPEF, LVMT, Ecole Nationale des Ponts et Chaussées
• Pierre-Olivier VANDANJON (membre invité), directeur de recherche, SPLOTT, Université Gustave Eiffel

·  Résumé

Exposure to air pollution contributes to chronic cardiovascular and respiratory diseases and premature death, especially in urban areas where regulated pollutants, such as nitrogen dioxide and fine particulate matter, have significant spatial variations. Assessments of population exposure and associated health impacts generally rely solely on pollutant concentration estimations at home. However, recent studies have highlighted the potential of integrated mobility – emissions – air quality – exposure modeling chains to represent individual exposure more accurately by accounting for exposure at workplaces or in transportation environments, referred to as dynamic exposure.

This thesis aims to place the modeling of individual exposure to urban air pollution into perspective within a mobility – emissions – air quality – exposure modeling chain, adopting an agent-based approach, and the implications of this framework for the analysis of environmental inequalities in light of the complexity of the modeling chain. In particular, I examine to what extent the results are sensitive to an increasing level of complexity within this modeling chain. To this end, I first propose an analytical framework for exposure as an Integrated Environmental Model, along with the associated uncertainties. These approaches link mobility and air quality modeling in several respects: on the one hand, pollutant emissions from individuals' daily mobility and, on the other hand, the dynamic exposure of individuals in the micro-environments they pass through.

Secondly, I present a modeling chain of individual exposure at a fine spatial scale, with the main contributions being (a) the representation of road traffic emissions from a car fleet that integrates the type of car owned by each household according to their socioeconomic and mobility characteristics; (b) modeling the city as a network of canyon streets based on OpenStreetMap; and (c) developing a model of dynamic exposure to air pollution at the street level. The modeling framework is applied to the Île-de-France region and is based on the eqasim population synthesis model, the MATSim agent-based mobility model, the HBEFA emission model, and a coupled air quality model. The latter consists of the Polair3D regional chemistry-transport model and the MUNICH street network model for simulating street canyons, both of which include the SSH-aerosol chemistry module for simulating secondary pollutants formation. Thus, this work provides tools for assessing individual exposure that can be replicated in other metropolitan areas, based on open-access models.

Third, I offer a cross-perspective on the socio-spatial analysis of individual exposure enabled by agent-based approaches, and on the scope of modeling uncertainties arising from common assumptions in exposure assessment — that is, to what extent the results vary with the level of complexity of the modeling chain. These analyses focus on spatial inequalities and socioeconomic disparities in exposure among different population groups, as well as how activity spaces contribute to individual exposure to air pollution. This work aims to inform researchers, practitioners, and public decision-makers about methods for modeling exposure to urban air pollution, as these assessments are essential both for evaluating the health impacts of ambient air pollution and for evaluating policies aimed at reducing them, and should thus improve the environmental assessment of transportation policies.