Causal inference for quantifying chemical–dynamical pathways controlling tropical middle stratospheric ozone variability

Galytska, Evgenia; Hassler, Birgit; Arosio, Carlo; Chipperfield, Martyn P.; Dhomse, Sandip S.; Dubé, Kimberlee; Feng, Wuhu; Iglesias-Suarez, Fernando; Runge, Jakob

doi:10.5194/acp-26-8185-2026

Articles | Volume 26, issue 11

https://doi.org/10.5194/acp-26-8185-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/acp-26-8185-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 26, issue 11

Research article

|

12 Jun 2026

Research article |

| 12 Jun 2026

Causal inference for quantifying chemical–dynamical pathways controlling tropical middle stratospheric ozone variability

Evgenia Galytska, Birgit Hassler, Carlo Arosio, Martyn P. Chipperfield, Sandip S. Dhomse, Kimberlee Dubé, Wuhu Feng, Fernando Iglesias-Suarez, and Jakob Runge

Data sets

Transformed Eulerian mean data from the ERA5 reanalysis (monthly means) F. Serva https://doi.org/10.5281/zenodo.7081721

QBO: monthly zonal stratospheric winds from tropical radiosonde data (mainly Singapore) T. Kerzenmacher and P. Braesicke https://doi.org/10.5281/zenodo.18850668

Model code and software

EyringMLClimateGroup/galytska26acp_CausalInferenceForTropicalMiddleStratOzone: Causal inference for quantifying chemical–dynamical pathways controlling tropical middle stratospheric ozone variability (v1.0) Evgenia Galytska https://doi.org/10.5281/zenodo.20509332

jakobrunge/tigramite: Tigramite 5.2 J. Runge https://doi.org/10.5281/zenodo.7747255

Short summary

We explore how chemical and dynamical processes shape ozone in the tropical middle stratosphere using satellite data and a chemistry-transport model. We apply a framework to identify cause–effect relationships and estimate their strength. For 2004-2021 monthly data, we detect an indirect pathway controlling ozone variability, with effects increasing after two–three months. We further assess how different wind regimes influence this pathway.