Near-real-time 3D Reconstruction of the Solar Coronal Parameters Based on the Magnetohydrodynamic Algorithm outside a Sphere Using Deep Learning Sumiaya Rahman, Hyun-Jin Jeong, Ashraf Siddique, Yong-Jae Moon, Bendict Lawrance Astrophysical Journal Supplement Series, 2024 For the first time, we generate solar coronal parameters (density, magnetic field, radial velocity, and temperature) on a near-real-time basis by deep learning. For this, we apply the Pix2PixCC deep-learning model to three-dimensional (3D) distributions of these parameters: synoptic maps of the photospheric magnetic field as an input and the magnetohydrodynamic algorithm outside a sphere (MAS) results as an output. To generate the 3D structure of the solar coronal parameters from 1 to 30 solar radii, we train and evaluate 152 distinct deep-learning models. For each parameter, we consider the data of 169 Carrington rotations from 2010 June to 2023 February: 132 for training and 37 for testing. The key findings of our study are as follows: First, our deep-learning models successfully reconstruct the 3D distributions of coronal parameters from 1 to 30 solar radii with an average correlation coefficient of 0.98. Second, during the solar active and quiet periods, the AI-generated data exhibits consistency with the target MAS simulation data. Third, our deep-learning models for each parameter took a remarkably short time (about 16 s for each parameter) to generate the results with an NVIDIA Titan XP GPU. As the MAS simulation is a regularization model, we may significantly reduce the simulation time by using our results as an initial configuration to obtain an equilibrium condition. We hope that the generated 3D solar coronal parameters can be used for the near-real-time forecasting of heliospheric propagation of solar eruptions.
Application of Deep Learning to Solar and Space Weather Data Yong-Jae Moon, Harim Lee, Jihyeon Son, Suk-Kyung Sung, Kangwoo Yi, Hyun-Jin Jeong, Eunsu Park, Eun-Young Ji, Il-Hyun Cho, Bendict Lawrance, Daye Lim, Gyungin Shin, Sujin Lee, Sumiaya Rahman, Taeyoung Kim Proceedings of the International Astronomical Union, 2023 In this review, we introduce our recent applications of deep learning to solar and space weather data. We have successfully applied novel deep learning methods to the following applications: (1) generation of solar farside/backside magnetograms and global field extrapolation based on them, (2) generation of solar UV/EUV images from other UV/EUV images and magnetograms, (3) denoising solar magnetograms using supervised learning, (4) generation of UV/EUV images and magnetograms from Galileo sunspot drawings, (5) improvement of global IRI TEC maps using IGS TEC ones, (6) one-day forecasting of global TEC maps through image translation, (7) generation of high-resolution magnetograms from Ca II K images, (8) super-resolution of solar magnetograms, (9) flare classification by CNN and visual explanation by attribution methods, and (10) forecasting GOES solar X-ray profiles. We present major results and discuss them. We also present future plans for integrated space weather models based on deep learning.
Generation of Solar Coronal White-light Images from SDO/AIA EUV Images by Deep Learning Bendict Lawrance, Harim Lee, Eunsu Park, Il-Hyun Cho, Yong-Jae Moon, Jin-Yi Lee, Shanmugaraju A, Sumiaya Rahman Astrophysical Journal, 2022 Low coronal white-light observations are very important to understand low coronal features of the Sun, but they are rarely made. We generate Mauna Loa Solar Observatory (MLSO) K-coronagraph like white-light images from the Solar Dynamics Observatory/Atmospheric Imaging Assembly (SDO/AIA) EUV images using a deep learning model based on conditional generative adversarial networks. In this study, we used pairs of SDO/AIA EUV (171, 193, and 211 Å) images and their corresponding MLSO K-coronagraph images between 1.11 and 1.25 solar radii from 2014 to 2019 (January to September) to train the model. For this we made seven (three using single channels and four using multiple channels) deep learning models for image translation. We evaluate the models by comparing the pairs of target white-light images and those of corresponding artificial intelligence (AI)–generated ones in October and November. Our results from the study are summarized as follows. First, the multiple channel AIA 193 and 211 Å model is the best among the seven models in view of the correlation coefficient (CC = 0.938). Second, the major low coronal features like helmet streamers, pseudostreamers, and polar coronal holes are well identified in the AI-generated ones by this model. The positions and sizes of the polar coronal holes of the AI-generated images are very consistent with those of the target ones. Third, from AI-generated images we successfully identified a few interesting solar eruptions such as major coronal mass ejections and jets. We hope that our model provides us with complementary data to study the low coronal features in white light, especially for nonobservable cases (during nighttime, poor atmospheric conditions, and instrumental maintenance).
Investigations of slow and fast coronal mass ejections and their associated activities in solar cycle 24 Indian Journal of Radio and Space Physics, 2015
