Advancing the application of novel causal inference to stratospheric chemistry and chemical-dynamical interactions of ozone-relevant species

• Applicant: Dr. Evgenia Galytska
• Subject Area: Atmospheric Science
• Term: since 2026
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 573930665

This proposal aims to advance the understanding of complex stratospheric chemical-dynamical interactions under a changing climate, focusing on ozone (O3)-related species, in particular nitrogen dioxide (NO₂), by coupling causal inference methods with traditional statistical approaches. These interactions are critical to the distribution of ozone, which not only protects life on Earth by absorbing harmful ultraviolet radiation but also shapes the thermal structure, and thus the circulation, of the stratosphere. Understanding stratospheric ozone variability requires disentangling the interplay of anthropogenic and natural drivers, including ozone-depleting substances (ODSs), greenhouse gas emissions, extreme natural events, and dynamical variability, particularly where transport and chemical processes are nonlinear and coupled. While traditional statistical methods are well-suited for estimating relationships and trends, they are limited in revealing true cause-and-effect connections, particularly in systems influenced by feedbacks, confounding factors, and regime-dependent behavior. Causal inference methods, an emerging class of techniques, combine statistical theory with supervised machine learning (ML). They are specifically designed to uncover direct, indirect, and spurious associations. This makes them a powerful tool for identifying and evaluating the underlying physical drivers of observed variability and assessing the consistency of chemistry-transport and chemistry-climate models with observations. Building on the applicant’s recent work that demonstrated the potential of causal inference in atmospheric science, particularly the successful application to ozone variability in the tropical middle stratosphere, this project will extend causal analysis across different regions, altitudes, and dynamical regimes of the stratosphere. Particular focus will be placed on identifying the drivers of ozone changes in the tropical lower stratosphere and the polar regions, where processes such as stratospheric transport, heterogenous and catalytic chemistry, and the influence of large-scale variability modes such as Quasi-Biennial Oscillation (QBO), El Niño-Southern Oscillation (ENSO) remain sources of uncertainty. By leveraging ground-based and satellite observations, model simulations, and state-of-the-art causal discovery tools, the proposed research will integrate causal reasoning into data-driven research, provide deeper insight into stratospheric chemistry-climate interactions and support the development of more accurate and physically grounded representations of ozone in climate models.

• DFG Programme: Research Grants
• International Connection: Belgium, Finland, United Kingdom
• Cooperation Partners: Professor Dr. Martyn Chipperfield; Dr. Quentin Errera; Dr. Viktoria Sofieva
• Co-Investigators: Dr. Alexei Rozanov; Dr. Mark Weber