Quantifying the Influence of Climate on Storm Activity Using Machine Learning

Abstract Midlatitude storms vary due to the slowly evolving climate and the rapidly changing synoptic conditions. While the impact of both factors has been studied extensively, their relative contributions remain poorly quantified. We use 84 years of ERA‐5 reanalysis data and convolutional neural networks to assess the relative importance of seasonal climatology versus synoptic conditions in controlling averaged and individual storm activity. Our models successfully predict over 90% of the variability in mean storm activity, showing climatic conditions dominate it. However, only one‐third of the variability in individual storm properties is attributed to climatic factors, indicating that synoptic conditions dominate individual storm characteristics. Further isolating the impact of long‐term climate trends on individual storms shows that it contributes to storm‐intensity variability only . In contrast, its contribution to storms' associated heat anomalies is over three times greater, demonstrating that variables directly linked to global warming provide a clearer pathway for weather attribution.