Elizabeth A. Barnes Research Group

Boston University

How can interpretable AI improve forecasts of Earth system variability across days-to-decades timescales? The group develops and implements artificial intelligence tools that mimic scientific human reasoning to improve intrinsic interpretability for climate applications. Researchers apply data-driven forecasting methods across temporal and spatial scales to quantify predictability and change in Earth systems using observational and modeled datasets. The team investigates mechanisms of Earth system variability and predictability by combining statistical analysis with interpretable machine learning to produce actionable scientific understanding. The group emphasizes responsible AI practices to anticipate human–Earth system futures and guide decision-relevant forecasting.