PhD position: PhD Air Quality Modeling - Leveraging Emerging Technologies for Data-scarce environments

Background: ALÁFÍÀ addresses critical air quality challenges in Sub-Saharan West Africa (SSWA), where rapid urbanization, biomass burning, and Saharan dust transport are driving pollution levels that threaten public health and quality of life. Despite these pressures, the region lacks comprehensive air quality monitoring infrastructure, hindering effective policy-making. This project propose to leverage satellite observations and deep learning to develop a regional air quality forecasting model. Using a physically informed Graph Neural Network (GNN), the model will integrate atmospheric processes like advection and diffusion, perform source apportionment to distinguish pollution sources (urbanization, biomass burning, dust), and downscale satellite observations to urban scales. Case studies in Lagos, Abidjan, and Accra will demonstrate the tool's applicability across diverse urban environments, ultimately delivering a scalable, cost-effective solution that empowers SSWA governments to promote healthier, more sustainable communities.

Job description: The PhD candidate will develop and implement a physically informed GNN at the heart of the ALÁFÍÀ project. Key tasks include processing Sentinel-5P and Sentinel-3 satellite data using Python workflows, designing hybrid ML-physics models with atmospheric constraints (boundary layer processes, emission inventories), and implementing physics-constrained loss functions and neural operators. The candidate will develop source apportionment algorithms to distinguish pollution sources and create interpretability frameworks for policymakers. Additional work includes downscaling satellite data to urban resolution and validating model outputs against observations in case study cities. This interdisciplinary role involves collaboration with atmospheric scientists, remote sensing experts, and West African stakeholders, offering the opportunity to contribute to research that directly supports environmental policy and public health in the region.

Methodology: The project stems from a hybrid ML-physics approach that enables accurate pollutant dispersion modeling and urban-scale downscaling, even in data-scarce environments. At its core is the development of a physically informed Neural Network that integrates Earth Observation data with atmospheric physics constraints.

Responsibilities and Tasks:

1. Engage in supervised scientific research that will ultimately result in a doctoral thesis,
2. Participate in consortium meetings, workshops, conferences, and contribute to scientific discourse on air quality and machine learning,
3. Publish research findings in scientific journals and present at international conferences,
4. Follow a PhD training programme at the Graduate School and relevant research schools,
5. Develop and implement the physically informed Graph Neural Network for air quality forecasting,
6. Process and integrate satellite data from Sentinel missions using Python-based workflows,
7. Design hybrid ML-physics models incorporating atmospheric physics constraints and source apportionment capabilities,
8. Validate model performance against ground observations in West African case study cities,
9. Collaborate with atmospheric scientists, remote sensing experts, and stakeholders in Sub-Saharan West Africa.

We are seeking a highly motivated PhD candidate with a Master's degree in Atmospheric Science, Meteorology, Mathematics, Physics, or Environmental Engineering. The ideal candidate will have experience in regional air quality modeling using chemistry-transport models and a demonstrated interest in AI tools, specifically deep learning and graph neural networks for spatiotemporal applications. Essential qualifications include a strong foundation in atmospheric physics—particularly boundary layer processes and emission inventories relevant to biomass burning and Saharan dust in West Africa—as well as proficiency in Python, along with skills in linear programming optimization. We particularly welcome candidates enthusiastic about hands-on work with satellite data, and those interested in developing hybrid ML-physics models employing physics-constrained losses. Familiarity with HPC environments (SLURM/GPU workflows) is advantageous. The successful candidate should demonstrate a strong commitment to transdisciplinary research and possess the collaborative skills necessary for working effectively in a multidisciplinary team. Good verbal and written communication skills in English are required (TOEFL iBT 90+ or equivalent). Proficiency in Dutch is a plus.

Station

Heerlen. You are present in Heerlen (at least) two days a week

Salary

The salary is determined in accordance with salary scale P of Appendix A of the Collective Labour Agreement of Dutch Universities and ranges from € 3.059,-- gross per month upon commencement to € 3.881,-- gross per month in the fourth and final year, in case of full employment.

The PhD candidate will be appointed for a period of 15 months. The appointment will be extended to 4 years when progress and performance are good. A PhD training program is part of the agreement.

The Open Universiteit provides good secondary benefits such as training, mobility, part-time employment and paid parental leave.

Working at the Open University means contributing to academic education for the most motivated students in the Netherlands and Flanders, as well as to scientific research with visible impact for people and society. With a broad range of bachelor’s and master’s programmes and short courses, the Open University makes lifelong learning accessible to everyone. There are no prior education requirements for bachelor’s programmes: anyone aged eighteen or older can enrol. Education is hybrid and flexible, available anytime and anywhere, at a pace that suits the student.

The majority of the nearly 16,000 students combine their studies with work. This creates a valuable interaction: experiences and questions from professional practice enrich education and research and increase their societal relevance. It is therefore no coincidence that the Open University has ranked in the top three of the Dutch Keuzegids for universities for many years, and this year holds the number one position.

In research, the university focuses on three central themes: healthcare, digitalisation and AI, and society and citizenship. Solutions to today’s challenges are sought across disciplinary boundaries and in collaboration with partners.

The Open University employs approximately 800 staff members, has its main campus in Heerlen, and operates study centres in the Netherlands and Flanders. Its culture is people-centred, collaborative, and inclusive: an environment in which students, staff, and partners feel connected, with ample opportunities for personal development and a healthy work–life balance.

This position sits at the nexus of environmental science  and computer science, combining cutting-edge AI techniques with atmospheric modeling to address critical air quality challenges. 

The Department of Environmental Sciences is part of the Faculty of Science. The Department embodies the commitment of the Open Universiteit to excellence in the environmental sciences, science for impact, and lifelong learning in the sustainability domain. The Department brings together an ambitious and inspiring group of people working on integrated environmental modelling, sustainability learning, and environmental governance and has about 20 team members. The research of the Department aims to contribute to the understanding of social-ecological systems, the development of solutions for environmental issues, and to the wider body of knowledge that helps societies reach their sustainability goals.

Department of Computer Sciences  

The Department of Computer Science is an ambitious and enthusiastic group of approximately 40 people (33 FTE), broadly focused on improving the impact of computers and computer science on society. The department’s research program “Towards high-quality and intelligent software” (2020–2025) consists of four research lines, focusing on: Techniques for quality assurance of software systems, Software and computer system security, and privacy-by-design, Responsible artificial intelligence, including methods and applications of AI and Educational tools and computing education.

Information

For more information about this vacancy you can contact: lyana.curier@ou.nl.