PhD Project Tile: Bridging Causality and Forecasting in Space Weather: A Transparent Complex Systems Modelling and Interpretable Machine Learning Framework”

Project Descriptions

Accurate space weather forecasting is essential for safeguarding satellite infrastructure and terrestrial power grids against geomagnetic disturbances and radiation hazards. While data-driven models have gained prominence, balancing high-fidelity predictive accuracy with physical interpretability remains a formidable challenge. This project introduces a novel hybrid framework designed to bridge the gap between causality and prediction in space weather. By integrating information-theoretic causality analysis, Sparse Nonlinear System Identification (S-NSID) and Interpretable Machine Learning (IML), the framework offers a transparent alternative to traditional ‘black-box’ approaches. The process begins with rigorous nonlinear causality tests to quantify information flow between solar wind drivers, such as velocity, density, pressure, and interplanetary magnetic field components, and their corresponding magnetospheric responses. These causal insights act as a principled feature-selection mechanism, ensuring the resulting models are both parsimonious and mathematically explicit. By identifying the sparse functional structures governing Sun-Earth coupling, this research provides a robust, physically grounded tool for real-time risk mitigation and the advancement of predictive space science.

The PhD studentship, funded by the STFC, is intended for home students graduating in a relevant subject (Mathematics, Computing, Computer Science, Engineering, Physics, Astronomy/Astrophysics, etc.). It includes full UK tuition fees, a living allowance (stipend), and additional funding to cover research costs, collaborative visits, and national and international travel to conferences and workshops.

Application Deadline: March 16th, 2026
Expected Start Date: October 2026

For further information, details on the project and application procedure, please contact Dr. Hua-Liang (Leon) Wei: w.hualiang@sheffield.ac.uk.