ARCANE -- Early Detection of Interplanetary Coronal Mass Ejections

Interplanetary coronal mass ejections (ICMEs) are major drivers of space weather disturbances, posing risks to both technological infrastructure and human activities. Automatic detection of ICMEs in solar wind in situ data is essential for early warning systems. While several methods have been proposed to identify these structures in time series data, robust real-time detection remains a significant challenge. In this work, we present ARCANE - the first framework explicitly designed for early ICME detection in streaming solar wind data under realistic operational constraints, enabling event identification without requiring observation of the full structure. Our approach evaluates the strengths and limitations of detection models by comparing a machine learning-based method to a threshold-based baseline. The ResUNet++ model, previously validated on science data, significantly outperforms the baseline, particularly in detecting high-impact events, while retaining solid performance on lower-impact cases. Notably, we find that using real-time solar wind (RTSW) data instead of high-resolution science data leads to only minimal performance degradation. Despite the challenges of operational settings, our detection pipeline achieves an F1-Score of 0.37, with an average detection delay of 24.5% of the event’s duration while processing only a minimal portion of the event data. As more data becomes available, the performance increases significantly. These results mark a substantial step forward in automated space weather monitoring and lay the groundwork for enhanced real-time forecasting capabilities.

💡 Research Summary

Interplanetary coronal mass ejections (ICMEs) are among the most hazardous space‑weather phenomena, capable of disrupting satellites, power grids, and communication systems. While many automatic detection methods have been proposed, they typically rely on high‑quality, post‑processed science data and require observation of the full event before a detection can be made, limiting their usefulness for real‑time warning. In this context the authors introduce ARCANE (Automatic Real‑time detection And forE‑cast), a modular framework explicitly designed to detect ICME onsets using only streaming real‑time solar‑wind (RTSW) measurements.

The data source is the NOAA Space Weather Prediction Center’s real‑time solar‑wind feed, which aggregates measurements from ACE, DSCOVR, and Wind. To ensure cross‑spacecraft compatibility the authors restrict inputs to six core parameters: the three magnetic‑field components (Bx, By, Bz), the total field magnitude |B|, proton density (Np), proton temperature (Tp), bulk speed (V), and plasma beta (β). The data are resampled to a 10‑minute cadence, and missing values are handled with linear interpolation and masking.

For the detection model the study adopts the ResUNet++ architecture, previously validated on high‑resolution science data, and treats ICME identification as a time‑series segmentation problem. Training uses a standard split (70 % train, 15 % validation, 15 % test) with class‑weighting to mitigate the strong imbalance between ICME and quiet‑solar intervals. As a baseline, a conventional threshold‑based detector that flags events based on magnetic‑field strength, β, and velocity drops is implemented.

Evaluation focuses on early‑detection capability: the model must issue an alert before the full ICME structure is observed. Results show that ResUNet++ achieves an overall F1‑score of 0.37, substantially higher than the baseline, and excels at high‑impact events where it reaches a recall of 0.55—over 30 % improvement. The average detection delay corresponds to 24 % of the event’s total duration, meaning alerts are raised well before the event’s end. Importantly, performance on real‑time data is only marginally lower (≤0.02 F1 loss) than on science‑quality data, demonstrating robustness to increased noise and occasional gaps. Performance scales positively with the amount of training data, with the F1‑score improving by roughly 0.05 for each doubling of the dataset.

The paper also discusses operational challenges: catalog inconsistencies, data gaps, and the trade‑off between early alerts and false‑alarm rates. ARCANE mitigates these issues through a modular design, post‑processing steps such as confidence scoring and minimum‑duration filtering, and by providing a flexible evaluation framework tailored to real‑time constraints. The authors suggest future extensions including multi‑sensor data fusion, Bayesian uncertainty quantification, and integration with downstream forecasting modules. In summary, ARCANE represents a significant advance toward operational, real‑time ICME early warning, offering a practical solution that bridges the gap between research‑grade detection and space‑weather forecasting needs.