Solar Wind Speed Prediction via Graph Attention Network. Issue 7 (11th July 2022)

• Record Type: Journal Article
• Authors: Sun, Yanru; Xie, Zongxia; Wang, Haocheng; Huang, Xin; Hu, Qinghua
• Is Part Of: Space weather. Volume 20:Issue 7(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-07-11
• Subjects: solar wind speed prediction -- deep learning -- graph attention network; Space environment -- Periodicals; 551.509992
• DOI: 10.1029/2022SW003128
• Languages: English
• ISSNs: 1542-7390

Abstract

Abstract: The solar wind is a plasma flow formed by the expansion of the high‐temperature corona and propagates in the interplanetary space with speeds between 200 km/s and 900 km/s. Accurate solar wind speed prediction and longer lead time will help mitigate the impact of solar storms on aerospace equipment and the Earth's magnetic field. Recently, most approaches do not explicitly capture the relationships between different solar wind features, and the prediction accuracy of 96‐hr is still not good enough. This paper elaborately designs an end‐to‐end model: Graph‐Temporal‐AR model (GTA) for solar wind speed prediction. First, our framework considers each feature as the node to construct the graph structure and adopts the graph attention module to learn the complex dependencies among features. Second, our approach employs the dilated causal convolution to extend the receptive field and prolong the prediction time. Furthermore, we leverage the autoregressive model to solve the scale insensitive problem of the neural network, making our model more robust. Specifically, we combine the OMNI data measured at Lagrangian Point 1 (L1) with the extreme ultraviolet images observed by the Solar Dynamics Observatory satellite to predict the solar wind speed at L1. Compared with the baseline models, GTA obtains significant performance improvements. Through visualization, we find GTA excavates the relationships between multiply variables without domain prior knowledge, which may help us find other unknown associations in heliophysics data sets. The data and code are available from https://github.com/syrGitHub/GTA . Plain Language Summary: The solar wind is charged particles stream ejected from the Sun, and high‐speed solar wind may affect Earth's magnetic field. This study proposes a deep learning‐based model to predict the solar wind speed at Lagrangian Point 1 reliably. We elaborately design an end‐to‐end model: Graph‐Temporal‐AR model (GTA) which combines the Graph Attention Network, the dilated causal convolution, with the autoregressive model. We adopt the coronal hole information extracted from the extreme ultraviolet image and the OMNI data in 5 years to train the parameters of the GTA model and obtain good performance compared with the other benchmark models. Key Points: We propose a method combining the linear module and the nonlinear component of the neural network for prediction We propose a method to excavate the complex relationships among solar wind features and learn temporal dependencies We propose a Graph‐Temporal‐AR model model that achieves robust performance in different situations and has good interpretability