Forecasting precipitation in the Arctic using probabilistic machine learning informed by causal climate drivers.
Chaos., 2025
Forecasting precipitation in Arctic maritime regions such as Bear Island and Ny-Ålesund is critical for climate risk assessment and early warning systems. This work proposes a probabilistic machine learning framework that integrates wavelet coherence and Synergistic-Unique-Redundant Decomposition to uncover scale-dependent and causal relationships between precipitation and atmospheric drivers. These insights guide the development of data-driven forecasting models incorporating both historical data and exogenous variables.
