Deep Learning Kp Computational Prediction Combined Multimodal Data of Solar Wind and EUV Images
doi: 10.11728/cjss2026.03.2025-0141 cstr: 32142.14.cjss.2025-0141
-
Abstract: This study presents a deep learning framework for multi-scale prediction of the geomagnetic Kp index by integrating multi-modal solar data. Unlike models that rely solely on upstream solar wind measurements from the Advanced Composition Explorer (ACE), which offer a short-term forecast horizon of 0.5~2 h, our approach synergistically incorporates Extreme Ultraviolet (EUV) images of coronal holes from the Solar Dynamics Observatory (SDO). These images provide precursor signals of solar activity, enabling medium-term forecasts. This study constructed a multi-year dataset (2011－2019) with carefully designed features to ensure physical meaning and sample balance. And several models were developed and evaluated: a direct solar wind model, an indirect multi-scale model using EUV images, and a model incorporating coronal hole location information. While all models showed similar and superior performance in 3-hour-ahead predictions, the model incorporating coronal hole location information demonstrated a significant advantage at extended forecast horizons, maintaining a Correlation Coefficient (CC) of approximately 0.23 even at a 72 h lead time. This work underscores the critical value of fusing solar source images with in-situ measurements to achieve reliable, long-lead space weather forecasts.
-
Figure 4. Quantification process of the position information of the solar extreme ultraviolet image
Figure 5. Different forecasting procedures for Kp prediction in terms of ACE parameters and SDO images
Figure 8. Scatter plot of correlation between measurement and prediction value of solar wind speed
Table 1. 23 Features of solar wind parameter with DM construction
Feature construction Parameters of Solar wind max n, v, B, Bz min n, v, B, Bz avg n, v, B, Bz delta n, v, B, Bx, By, Bz avg_abs Bx, By, Bz max (difference of absolute value) max (Bx,delta, By,delta, Bz,delta) true Kp 
下载: 导出CSV
Table 2. Comparison of prediction between random and yearly dataset of DKPW
Dataset form RMSE MAE CC Random 0.675 0.528 0.852 Yearly 0.656 0.505 0.866
下载: 导出CSV
Table 3. Advance time and performance of Kp prediction with MKPI
Advance time/h RMSE SD MAE CC 3 0.63 1.07 0.39 0.86 6 0.90 0.79 0.81 0.70 9 1.00 0.67 1.00 0.63 12 1.05 0.58 1.1 0.57 15 1.05 0.48 1.11 0.52 18 1.07 0.45 1.15 0.47 21 1.10 0.37 1.22 0.42 24 1.13 0.31 1.27 0.38 27 1.15 0.25 1.32 0.33 30 1.16 0.20 1.35 0.30 33 1.18 0.15 1.39 0.24 36 1.19 0.11 1.41 0.18 39 1.19 0.09 1.43 0.15 42 1.20 0.07 1.43 0.16 45 1.21 0.00 1.45 0.00
下载: 导出CSV
Table 4. Earlier advance time and better performance of Kp prediction of KPCL
Advance time/h RMSE SD MAE CC 3 0.63 1.06 0.4 0.85 6 0.93 0.76 0.87 0.71 9 1.05 0.60 1.10 0.62 12 1.08 0.52 1.17 0.56 15 1.10 0.46 1.21 0.50 18 1.13 0.43 1.28 0.45 21 1.14 0.38 1.31 0.39 24 1.16 0.34 1.34 0.35 27 1.18 0.30 1.39 0.30 30 1.18 0.32 1.40 0.25 33 1.18 0.21 1.38 0.24 36 1.18 0.20 1.38 0.23 39 1.18 0.19 1.39 0.23 42 1.19 0.21 1.39 0.22 45 1.18 0.23 1.38 0.23 48 1.18 0.21 1.39 0.23 51 1.18 0.20 1.39 0.25 54 1.17 0.21 1.38 0.25 57 1.18 0.22 1.39 0.22 60 1.18 0.21 1.39 0.22 63 1.18 0.20 1.38 0.24 66 1.18 0.21 1.40 0.23 69 1.18 0.20 1.39 0.23 72 1.18 0.19 1.38 0.24
下载: 导出CSV
-
[1] LUHMANN J G, LI Y, LEE C O, et al. Solar cycle variability in coronal holes and their effects on solar wind sources[J]. Space Weather, 2022, 20(10): e2022SW003110 doi: 10.1029/2022SW003110 [2] WINTOFT P, WIK M, MATZKA J, et al. Forecasting Kp from solar wind data: input parameter study using 3-hour averages and 3-hour range values[J]. Journal of Space Weather and Space Climate, 2017, 7: A29 doi: 10.1051/swsc/2017027 [3] ENGQUIST B, LÖTSTEDT P, RUNBORG O. Multiscale Modeling and Simulation in Science[M]. Berlin, Heidelberg: Springer, 2009 [4] BROWN E J E, SVOBODA F, MEREDITH N P, et al. Attention-based machine vision models and techniques for solar wind speed forecasting using solar EUV images[J]. Space Weather, 2022, 20(3): e2021SW002976 doi: 10.1029/2021SW002976 [5] GONZALEZ R C, WOODS R E. Digital Image Processing[M]. 4th ed. New York: Pearson, 2017 [6] HE K M, GIRSHICK R, DOLLÁR P. Rethinking ImageNet pre-training[C]//Proceedings of 2019 IEEE/CVF International Conference on Computer Vision. Seoul: IEEE, 2019: 4917-4926 [7] BERNOUX G, BRUNET A, BUCHLIN É, et al. Forecasting the geomagnetic activity several days in advance using neural networks driven by solar EUV imaging[J]. Journal of Geophysical Research: Space Physics, 2022, 127(10): e2022JA030868 doi: 10.1029/2022JA030868 [8] CONDE D, CASTILLO F L, ESCOBAR C, et al. Forecasting geomagnetic storm disturbances and their uncertainties using deep learning[J]. Space Weather, 2023, 21(11): e2023SW003474 doi: 10.1029/2023SW003474 [9] E W, HAN J Q, JENTZEN A. Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations[J]. Communications in Mathematics and Statistics, 2017, 5(4): 349-380 doi: 10.1007/s40304-017-0117-6 [10] GALVEZ R, FOUHEY D F, JIN M, et al. A machine-learning data set prepared from the NASA Solar Dynamics Observatory mission[J]. The Astrophysical Journal Supplement Series, 2019, 242(1): 7 doi: 10.3847/1538-4365/ab1005 [11] SHIH C Y, LIN C Y T, LIN S Y, et al. Forecasting of global ionosphere maps with multi-day lead time using Transformer-based neural networks[J]. Space Weather, 2024, 22(2): e2023SW003579 doi: 10.1029/2023SW003579 [12] PARK E, MOON Y J, SHIN S, et al. Application of the deep convolutional neural network to the forecast of solar flare occurrence using full-disk solar magnetograms[J]. The Astrophysical Journal, 2018, 869(2): 91 doi: 10.3847/1538-4357/aaed40 [13] 李晓坤. 多模态多尺度日面信息的深度学习地磁指数预报[D]. 北京: 中国科学院大学, 2022LI Xiaokun. Geomagnetic Index Prediction with Deep Learning Based on Multi-Modal and Multi-Scale Solar Surface Information[D]. Beijing: University of Chinese Academy of Sciences, 2022 [14] HUANG X, WANG H N, XU L, et al. Deep learning based solar flare forecasting model. I. results for line-of-sight magnetograms[J]. The Astrophysical Journal, 2018, 856(1): 7 doi: 10.3847/1538-4357/aaae00 [15] HEGDE1 D V, KIM T K, POGORELOV N V, et al. Ensemble modeling of the solar wind flow with boundary conditions governed by synchronic photospheric magnetograms. I. multipoint validation in the inner heliosphere[J]. The Astrophysical Journal, 2025, 988(2): 154 doi: 10.3847/1538-4357/addf33 [16] GUO Dalei, ZHANG Zhen, ZHU Lingfeng, et al. Generative model-based of flare hierarchic recognition and forecast of extreme ultraviolet images in solar active region[J]. Chinese Journal of Space Science, 2023, 43(1): 60-67 doi: 10.11728/cjss2023.01.220214015 [17] OWOLABI C, RUAN H B, YAMAZAKI Y, et al. Empirical modeling of ionospheric current using empirical orthogonal function analysis and artificial neural network[J]. Space Weather, 2021, 19(11): e2021SW002831 doi: 10.1029/2021SW002831 [18] WINTER P M, BURGER C, LEHNER S, et al. Residual neural networks for the prediction of planetary collision outcomes[J]. Monthly Notices of the Royal Astronomical Society, 2023, 520(1): 1224-1242 doi: 10.1093/mnras/stac2933 [19] ZEWDIE G K, VALLADARES C, COHEN M B, et al. Data-driven forecasting of low-latitude ionospheric total electron content using the random forest and LSTM machine learning methods[J]. Space Weather, 2021, 19(6): e2020SW002639 doi: 10.1029/2020SW002639 [20] HAO Y H, LU J Y, PENG G S, et al. F10.7 daily forecast using LSTM combined with VMD method[J]. Space Weather, 2024, 22(1): e2023SW003552 [21] The SunPy Community, MUMFORD S J, CHRISTE S, et al. SunPY-python for solar physics[J]. Computational Science & Discovery, 2015, 8(1): 014009 [22] GUO D L, ZHANG T Y, XUE B S. Enhanced and quantified features for magnetic unipolarity in HMI magnetogram of coronal hole[C]//Proceedings of 2025 IEEE 11th International Conference on Control Science and Systems Engineering. Beijing, China: IEEE, 2025 -
-
下载:
图(10) / 表(4)
计量
- 文章访问数: 392
- HTML全文浏览量: 69
- PDF下载量: 84
-
被引次数:
0(来源:Crossref)
0(来源:其他)
