Research
Current Research Projects
Hybrid modelling frameworks that combine process-based hydrology with machine learning to improve streamflow and river temperature prediction across UK catchments.
Project 1 — Signature-Enhanced PIML for Streamflow Prediction
Developing physics-informed machine learning models that combine HBV-derived
hydrological states with LSTM and catchment signatures to improve streamflow
prediction across diverse UK catchments, with a focus on interpretability,
robustness, and performance in heterogeneous and human-influenced catchments.
→ View EGU 2025 Abstract
Project 2 — Reservoir-Impacted Catchments & Hybrid Modelling
Investigating how reservoir regulation affects streamflow behaviour and developing hybrid modelling approaches to improve prediction in human-influenced catchments. Research focuses on identifying where process-based models fail and how machine learning can fill those gaps while preserving physical consistency.
Project 3 — River Temperature Hybrid Modelling
Developing hybrid models that combine process-based understanding and machine learning to improve river temperature prediction and assess climate-related impacts on freshwater systems. Produced practical recommendations for the Environment Agency on developing a pilot forecasting system to support ecological protection and river management.
Research Interests
- Hybrid & Physics-Informed Machine Learning
- Flood Forecasting & Streamflow Prediction
- River Temperature Modelling
- Environmental Data Science & Climate Adaptation