PhD position on Causal Machine Learning & Earth Observation for Urban Flood Resilience (AUROrA Project)

01.10.2025, Wissenschaftliches Personal

This position brings together four research groups at TUM: the Chair of Data Science in Earth Observation (Prof. Xiao Xiang Zhu), the Chair of Hydrology and River Basin Management (Prof. Markus Disse), the Chair of Geoinformatics (Prof. Thomas H. Kolbe), and the Chair of Algorithmic Machine Learning & Explainable AI (Prof. Stefan Bauer). The project aims to develop an integrated urban flood analysis framework that supports flood-resilient urban design by combining multi-sensor Earth observation (EO) data, hydrological–hydrodynamic models, urban digital twins, and advanced AI methodologies.

Within this broader framework, the advertised PhD project focuses on leveraging EO data and causal machine learning to systematically uncover the drivers of urban flooding and quantify their impacts on flood resilience. The research will integrate multi-modal EO data (e.g., Sentinel-1/2, PlanetScope), urban morphology, drainage networks, and socio-hydrological information to (i) enhance EO-based urban flood mapping, (ii) identify causal links between rainfall, drainage, and built-up morphology, and (iii) deliver actionable insights for resilient urban planning. Results will be validated against hydrodynamic models and 3D urban semantic models in collaboration with project partners, with applications to recent major flood events in Germany, Brazil, and Spain.

Supervision & collaborations:
The PhD will be supervised by Prof. Xiao Xiang Zhu (SiPEO, TUM), and Dr. Jie Zhao acts as the day-to-day supervisor, with close collaborations inside TUM (The chair of Hydrology and River Basin Management, The chair of Geoinformatics, and The chair of Algorithmic Machine Learning & Explainable AI) and access to external partners and datasets.

Your tasks will include:
• Build a comprehensive multi-modal urban flood dataset integrating Sentinel-1/2, PlanetScope and other data sources
• Develop robust AI models for urban flood mapping using sparse multi-temporal, multi-source EO data (SAR intensity, InSAR coherence, optical), including cross-modal fusion and modality distillation
• Design a causation analysis framework combining deep learning with causal discovery & inference to quantify the influence of rainfall, drainage capacity, and 3D urban form on flood severity
• Validate results with hydrodynamic simulations and 3D urban semantic models, benchmark against state-of-the-art methods, and publish in leading international journals and conferences
• Literature research
• Scientific publishing

Your qualifications:
• Completed academic university degree (university diploma / M.Sc.) in Computer Science, Geoscience, Remote Sensing, Hydrology, Data Science, Physics, or related fields
• Experience in machine learning (ML), artificial intelligence (AI) or related fields
• Software skills in ML languages such as Python
• Ability and enthusiasm to learn new technologies quickly
• Ability to work highly motivated both independently and in a team
• Very good written and spoken English skills
• Some knowledge or background in the SAR is an advantage
• Knowledge of causal ML, graphical models, or multimodal data fusion is an advantage

We offer:
• An exciting and challenging job at a university ranked among the best worldwide
• Compatibility of job and family
• Possibility of remote work (home office)
• A friendly and cooperative environment
• A PhD position remunerated according to TV-L E 13 100% (Tarifvertrag für den öffentlichen Dienst der Länder). The successful applicant will have a 48-month contract. As an equal opportunity and affirmative action employer, TUM explicitly encourages applications from women as well as from all others who would bring additional diversity dimensions to the university’s research and teaching strategies. Preference will be given to disabled candidates with essentially the same qualifications.
• Start date: early 2026 (ASAP)

Did we catch your interest? We are looking forward to receiving your comprehensive application, including your letter of motivation, CV, and academic transcripts of records, preferably in English via an email to ai4eo@tum.de until 1. November 2025 at the latest. Please indicate "PhD application for Causal ML & EO for Urban Flood Resilience” in the subject line.

The position is suitable for people with severe disabilities. Applicants with severe disabilities will be given preference if they are otherwise essentially equally qualified, capable, and professionally competent.

Kontakt: ai4eo@tum.de