Space Weather Modeling Research Articles

Space Weather Forecasts of Ground Level Space Weather in the UK: Evaluating Performance and Limitations

AbstractGeomagnetically Induced Currents (GICs) are a severe space weather hazard, driven through coupling between the solar wind and magnetosphere. GICs are rarely measured directly, instead the ground magnetic field variability is often used as a proxy. Recently space weather models have been developed to forecast whether the magnetic field variability (R) will exceed specific, extreme thresholds. We test an example machine learning‐based model developed for the northern United Kingdom. We evaluate its performance (discriminative skill and calibration) as a function of magnetospheric state, solar wind input and magnetic local time. We find that the model's performance is highest during active conditions, for example, geomagnetic storms, and lowest during isolated substorms and “quiet” intervals, despite these conditions dominating the training data set. Correspondingly, the performance is high when the solar wind conditions are elevated (i.e., high velocity, large total magnetic field strength, and the interplanetary magnetic field oriented southward), and at a minimum when the north‐south component of the magnetic field is highly variable or around zero. Regarding magnetic local time, performance is highest within the dusk and night sectors, and lowest during the day. The model appears to capture multiple modes of magnetospheric activity, including substorms and viscous interactions, but poorly predicts impulsive phenomena (i.e., storm sudden commencements) and longer timescale coupling processes. Future models of mid‐latitude magnetic field variability will need to effectively use longer time intervals of unpropagated (i.e., observations from L1) solar wind to more completely describe the magnetospheric conditions and response.

The Schatten current sheet

Space weather models endeavoring to connect remote observations to in-situ measurements at various locations in the heliosphere invariably require a coronal model to connect the photosphere magnetically to the inner heliosphere. The most famous and popular implementation of this connection is a potential field source surface (PFSS) model out to the source surface, typically located at 2.5 solar radii, combined with a Schatten current sheet (SCS) model. While the PFSS model is mostly understood, the SCS has been utilized in heliospheric physics for nearly 50 years with little understanding of it’s physical and mathematical underpinnings. In this overview article, I lay out the mathematical formalism of the SCS, describe how it differs from the PFSS, and summarize several techniques used to combine the PFSS and SCS to create a global coronal model from the photosphere to the inner heliosphere.

Graph-GIC: A smart and parallelized geomagnetically induced current modelling algorithm based on graph theory for space weather applications

Geomagnetically Induced Current (GIC) refers to the electromagnetic response of the Earth and its conductive modern infrastructures to space weather and would pose a significant threat to high-voltage power grids designed for the alternative current operation. To assess the impact of space weather on the power grid, one needs to calculate the GIC on a national or continental scale. In this study, we developed a smart and parallelized GIC modelling algorithm, Graph GIC. This algorithm deploys a graph representing a power grid in a single-line diagram, in which substations/transformers act as nodes and transmission lines as edges. With these denotations, a power grid and its electric parameters are mathematically represented with an adjacency matrix and an admittance matrix. We used sparse matrix and parallelisation techniques to expedite the intensive computation in cases of large-scale power grids. The Graph GIC was validated with a benchmark grid, applied to the GIC calculation of the 500 kV power grid of Guangdong, China, and conducted preliminary analysis on the grid’s susceptibility to geomagnetic storms. The Graph GIC algorithm has the advantage of an intuitive and highly scalable graph representation of a power grid at any scale. It achieves high-accuracy calculation and a speedup of about 18 times after parallelisation. This algorithm could be applied to assess the impact of space weather on a power grid up to continental scales and could be incorporated into global space weather modelling frameworks.

The Spatial Variation of Large‐ and Meso‐Scale Plasma Flow Vorticity Statistics in the High‐Latitude Ionosphere and Implications for Ionospheric Plasma Flow Models

AbstractThe ability to understand and model ionospheric plasma flow on all spatial scales has important implications for operational space weather models. This study exploits a recently developed method to statistically separate large‐scale and meso‐scale contributions to probability density functions (PDFs) of ionospheric flow vorticity measured by the Super Dual Auroral Radar Network (SuperDARN). The SuperDARN vorticity data are first sub‐divided depending on the Interplanetary Magnetic Field (IMF) direction, and the separation method is applied to PDFs of vorticity compiled in spatial regions of size 1° of geomagnetic latitude by 1 hr of magnetic local time, covering much of the high‐latitude ionosphere in the northern hemisphere. The resulting PDFs are fit by model functions using maximum likelihood estimation (MLE) and the spatial variations of the MLE estimators for both the large‐scale and meso‐scale components are presented. The spatial variations of the large‐scale vorticity estimators are ordered by the average ionospheric convection flow, which is highly dependent on the IMF direction. The spatial variations of the meso‐scale vorticity estimators appear independent of the senses of vorticity and IMF direction, but have a different character in the polar cap, the cusp, the auroral region, and the sub‐auroral region. The paper concludes by discussing the sources of the vorticity components in the different regions, and the consequences for the fidelity of ionospheric plasma flow models.

Impact of far-side structures observed by Solar Orbiter on coronal and heliospheric wind simulations

Context. Solar Orbiter is a new space observatory that provides unique capabilities to understand the heliosphere. In particular, it has made several observations of the far-side of the Sun and therefore provides unique information that can greatly improve space weather monitoring. Aims. In this study, we aim to quantify how the far-side data will affect simulations of the corona and the interplanetary medium, especially in the context of space weather forecasting. Methods. To do so, we focused on a time period with a single sunspot emerging on the far-side in February 2021. We used two different input magnetic maps for our models: one that includes the far-side active region and one that does not. We used three different coronal models typical of space weather modeling: a semi-empirical model (potential field source surface or PFSS) and two different magnetohydrodynamic models (Wind Predict and Wind Predict-AW). We compared all the models with both remote sensing and in situ observations in order to quantify the impact of the far-side active region on each solution. Results. We find that the inclusion of the far-side active region in the various models has a small local impact due to the limited amount of flux of the sunspot (at most 8% of the total map flux), which leads, for example, to coronal hole changes of around 7% for all models. Interestingly, there is a more global impact on the magnetic structure seen in the current sheet, with clear changes, for example, in the coronal hole boundaries visible in extreme ultra-violet (EUV) on the western limb, which is opposite to the active region and the limb most likely to be connected to Earth. For the Wind Predict-AW model, we demonstrate that the inclusion of the far-side data improves both the structure of the streamers and the connectivity to the spacecraft. Conclusions. In conclusion, the inclusion of a single far-side active region may have a small local effect with respect to the total magnetic flux, but it has global effects on the magnetic structure, and thus it must be taken into account to accurately describe the Sun-Earth connection. The flattening of the heliospheric current sheet for all models reveals that it causes an increase of the source surface height, which in return affects the open and closed magnetic field line distributions.

Multi-spacecraft study with the Icarus model. Modelling the propagation of CMEs to Mercury and Earth

Coronal mass ejections (CMEs) are the main drivers of the disturbances in interplanetary space. Earth-directed CMEs can be dangerous, and understanding the CME interior magnetic structure is crucial for advancing space weather studies. It is important to assess the capabilities of a numerical heliospheric model, as a firm understanding of the nature and extent of its limitations can be used to improve the model and the space weather predictions based on it. The aim of the present study is to test the capabilities of the recently developed heliospheric model Icarus and the linear force-free spheromak model that has been implemented in it. To validate the Icarus space weather modelling tool, two CME events were selected that were observed by two spacecraft located near Mercury and Earth, respectively. This enables us to test the heliospheric model computed with Icarus at two distant locations. The source regions for the CMEs were identified, and the CME parameters were determined and later optimised. Different adaptive mesh refinement levels were applied in the simulations to assess its performance by comparing the simulation results to in situ measurements. The first CME event erupted at 15:25 on July 9, 2013. The modelled time series were in good agreement with the observations both at MESSENGER and ACE. The second CME event started at 10:25 on February 16, 2014, and was more complicated, as three CME interactions occurred in this event. It was impossible to recover the observed profiles without modelling the other two CMEs that were observed, one before the main CME and one afterward. The parameters for the three CMEs were identified and the three CMEs were modelled in Icarus. For both CME studies, AMR level 3 was sufficient to reconstruct small-scale features near Mercury, while at Earth, AMR level 4 was necessary due to the radially stretched grid that was used. The profiles obtained at both spacecraft resemble the in situ measurements well. The current limitations of the space weather modelling tool result in an excessively small deceleration of the CME propagation during the CME--CME interaction as measured by MESSENGER and ACE.

Deep temporal convolutional networks for F10.7 radiation flux short-term forecasting

Abstract. F10.7, the solar flux at a wavelength of 10.7 cm (F10.7), is often used as an important parameter input in various space weather models and is also a key parameter for measuring the strength of solar activity levels. Therefore, it is valuable to study and forecast F10.7. In this paper, the temporal convolutional network (TCN) approach in deep learning is used to predict the daily value of F10.7. The F10.7 series from 1957 to 2019 are used. The data during 1957–1995 are adopted as the training dataset, the data during 1996–2008 (solar cycle 23) are adopted as the validation dataset, and the data during 2009–2019 (solar cycle 24) are adopted as the test dataset. The leave-one-out method is used to group the dataset for multiple validations. The prediction results for 1–3 d ahead during solar cycle 24 have a high correlation coefficient (R) of 0.98 and a root mean square error (RMSE) of only 5.04–5.18 sfu. The overall accuracy of the TCN forecasts is better than the autoregressive (AR) model (it only takes past values of the F10.7 index as inputs) and the results of the US Space Weather Prediction Center (SWPC) forecasts, especially for 2 and 3 d ahead. In addition, the TCN model is slightly better than other neural network models like the backpropagation (BP) neural network and long short-term memory (LSTM) network in terms of the solar radiation flux F10.7 forecast. The TCN model predicted F10.7 with a lower root mean square error, a higher correlation coefficient, and a better overall model prediction.

Accuracy of Global Geospace Simulations: Influence of Solar Wind Monitor Location and Solar Wind Driving

AbstractSome space weather models, such as the Space Weather Modeling Framework (SWMF) used in this study, use solar wind propagated from the first Lagrange point (L1) to the bow shock nose (BSN) to forecast geomagnetic storms. The SWMF is a highly coupled framework of space weather models that include multiple facets of the Geospace environment, such as the magnetosphere and ionosphere. The propagated solar wind measurements are used as a boundary condition for SWMF. The solar wind propagation method is a timeshift based on the calculated phase front normal (PFN) which leads to some uncertainties. For example, the propagated solar wind could have evolved during this timeshift. We use a data set of 123 geomagnetic storms between 2010 and 2019 run by the SWMF Geospace configuration to analyze the impact solar wind propagation and solar wind driving has on the geomagnetic indices. We look at the probability distributions of errors in SYM‐H, cross polar cap potential (CPCP), and auroral electrojet indices AL and AU. Through studying the median errors (MdE), standard deviations and standardized regression coefficients, we find that the errors depend on the propagation parameters. Among the results, we show that the accuracy of the simulated SYM‐H depends on the spacecraft distance from the Sun‐Earth line. We also quantify the dependence of the standard deviation in SYM‐H errors on the PFN and solar wind pressure. These statistics provide an insight into how the propagation method affects the final product of the simulation, which are the geomagnetic indices.

Identification and extraction of type II and III radio bursts based on YOLOv7

Solar radio bursts (SRBs) are extreme space weather events characterized by intense solar radio emissions that are closely related to solar flares. They represent signatures of the same underlying processes that are responsible for well-documented solar phenomena such as sunspots, solar flares, and coronal mass ejections (CMEs). The study of SRBs holds significant importance as it provides a means to monitor and predict solar flares and CMEs, enhancing our ability to forecast potential impacts on Earth’s communications and satellites. Typically, SRBs below several hundred megahertz can be categorized into five types (I–V), with type II and type III bursts being the most prevalent. This study introduces a novel approach based on the YOLOv7 model for the detection and classification of type II and type III SRBs. The proposed method effectively identifies and classifies various SRB types, achieving a mean average precision accuracy of 73.5%. A trained neural network was deployed for SRB detection in the Chashan Broadband Solar radio spectrograph at meter wavelength (CBSm) data, enabling the extraction of valuable SRB information for subsequent research. This demonstrates that even when we are dealing with extensive datasets, this method can automatically recognize outbursts and extract pertinent physical information. Although our experiments with the CBSm dataset currently rely on the daily spectrum, further advancements in CBSm backend data processing techniques are expected to enable near-real-time burst detection, which is a powerful tool for accurately assessing and analyzing SRBs, and significantly contribute to the field of space weather forecasting and protective measures. Furthermore, the applicability of this method to other stations within the Chinese Meridian Project II (e.g., Mingantu Spectral Radioheliograph and Daocheng Solar Radio Telescope) enhances the capability of space weather data fusion and model development. Therefore, this research represents a substantial contribution to the domain of space weather research, offering a valuable tool for the detection and classification of SRBs and thereby improving our ability to predict and mitigate the impacts of extreme space weather events on Earth’s technology and infrastructure.

A high‐order discontinuous Galerkin approach for physics‐based thermospheric modeling

AbstractThe accurate prediction of aerodynamic drag on satellites orbiting in the upper atmosphere is critical to the operational success of modern space technologies, such as satellite‐based communication or navigation systems, which have become increasingly popular in the last few years due to the deployment of constellations of satellites in low‐Earth orbit. As a result, physics‐based models of the ionosphere and thermosphere have emerged as a necessary tool for the prediction of atmospheric outputs under highly variable space weather conditions. This paper proposes a high‐fidelity approach for physics‐based space weather modeling based on the solution of the Navier–Stokes equations using a high‐order discontinuous Galerkin method, combined with a matrix‐free strategy suitable for high‐performance computing on GPU architectures. The approach consists of a thermospheric model that describes a chemically frozen neutral atmosphere in nonhydrostatic equilibrium driven by the external excitation of the Sun. A novel set of variables is considered to treat the low densities present in the upper atmosphere and to accommodate the wide range of scales present in the problem. At the same time, and unlike most existing approaches, radial and angular directions are treated in a nonsegregated approach. The study presents a set of numerical examples that demonstrate the accuracy of the approximation and validate the current approach against observational data along a satellite orbit, including estimates of established empirical and physics‐based models of the ionosphere‐thermosphere system. Finally, a one‐dimensional radial derivation of the physics‐based model is presented and utilized for conducting a parametric study of the main thermal quantities under various solar conditions.

A machine learning approach combined with wavelet analysis for automatic detection of Pc5 geomagnetic pulsations observed at geostationary orbits

Pc5 geomagnetic pulsations are ultra-low frequency (ULF) waves whose period lies within 150–600 s. Their detection is crucial for accurate space weather modelling, monitoring, and analysis. Identification of these pulsations using manual approaches is difficult and time-consuming because of the smaller magnitudes and the limited durations within which they occur. To overcome this challenge, we propose a robust Cascade Forward Neural Network (CFNN) model combined with the wavelet technique for automatically detecting Pc5 events from satellite datasets observed at geostationary orbits. The dataset used in this study is the magnetic field vector measurements retrieved from the Geostationary Operational Environmental Satellite-10 (GOES-10) from 2000 to 2009. Pc5 geomagnetic pulsations were extracted from the toroidal component of the field perturbation using a bandpass Butterworth filter. Continuous wavelet transform (CWT) analysis using the mother Morlet wavelet was utilized to validate the integrity of the extracted signal in the time–frequency domain. The extracted Pc5 signal was decomposed into details and approximations using Daubechies wavelet transform to separate the intelligible signal from the incoherent noise that left their trace on the magnetic field time series. The detail signal was subjected to denoising using the heuristic Stein Unbiased Risk Estimate (SURE) approach with soft thresholding to obtain the denoised Pc5 events. The denoised Pc5 events were utilized as the target in the machine learning whereas the inputs were obtained from 5-dimensional space transformation of the toroidal component of the time series. The developed algorithm performed well and demonstrated a detection accuracy of 80 % and a Root Mean Square Error (RMSE) of 0.13 nT indicating a high performance in detecting Pc5 events. For effective validation of the model, Pc5 events detected by our model were correlated with the Kp index and the amplitude of the Pc5 events observed at geostationary orbit. A good correlation was obtained in both cases, making our model a practical and preferred choice for Pc5 pulsation detection in contrast to conventional frequency analysis tools which take much time for signal processing on large data sets.

Data‐Driven Empirical Conductance Relations During Auroral Precipitation Using Incoherent Scatter Radar and All Sky Imagers

AbstractWe present empirical conductance relations that are derived from incoherent scatter radar observations and correlated with all sky imager observations to identify the morphology of the aurora. We use 75,461 events collected using the Poker Flat Incoherent Scatter Radar (PFISR) with associated all sky imagers observations spanning the years 2012–2016. In addition to classifying these events based on auroral morphology, we estimated the Hall and Pedersen conductance and the differential number flux from which the energy flux and the average energy can be calculated. The differential number flux was estimated using the maximum entropy inversion method described in Semeter and Kamalabadi (2005, https://doi.org/10.1029/2004RS003042), but now incorporating the Fang et al. (2010, https://doi.org/10.1029/2010GL045406) ionization model. The main results of this investigation are the power law equations that describe the median, 90th, and 10th percentile Hall and Pedersen conductance as a function of energy flux and average energy. These power law fits are performed for different auroral morphology including all events, discrete, diffuse, and pulsating auroral events. The median Pedersen conductance is found to be in good agreement with past empirical conductance specifications by Robinson et al. (1987, https://doi.org/10.1029/JA092iA03p02565); however, the median Hall conductance from the PFISR observations is found to be larger than the empirical Hall conductance formulas by Robinson et al. (1987, https://doi.org/10.1029/JA092iA03p02565). Pulsating aurora is found to be the most frequently occurring auroral morphology. Furthermore, pulsating aurora has an important contribution to Hall conductance since it has higher average energies than discrete aurora. The results from this investigation are applicable to space weather models and may enable better agreement between model‐data comparisons.

30-min decayless kink oscillations in a very long bundle of solar coronal plasma loops

The energy balance in the corona of the Sun is the key to the long-standing coronal heating dilemma, which could be potentially revealed by observational studies of decayless kink oscillations of coronal plasma loops. A bundle of very long off-limb coronal loops with the length of 736pm 80 Mm and a lifetime of about 2 days are found to exhibit decayless kink oscillations. The oscillations are observed for several hours. The oscillation amplitude is measured at 0.3–0.5 Mm, and the period at 28–33 min. The existence of 30-min periodicity of decayless kink oscillations indicates that the mechanism compensating the wave damping is still valid in such a massive plasma structure. It provides important evidence for the non-resonant origin of decayless kink oscillations with 2–6 min periods, i.e., the lack of their link with the leakage of photospheric and chromospheric oscillations into the corona and the likely role of the broadband energy sources. Magnetohydrodynamic seismology based on the reported detection of the kink oscillation, with the assistance of the differential emission measure analysis and a background coronal model provides us with a comprehensive set of plasma and magnetic field diagnostics, which is of interest as input parameters of space weather models.

The role of plasma β in global coronal models: Bringing balance back to the force

Context. COolfluid COrona uNstrUcTured (COCONUT) is a global coronal magnetohydrodynamic (MHD) model that was recently developed and will soon be integrated into the ESA Virtual Space Weather Modelling Centre (VSWMC). In order to achieve robustness and fast convergence to a steady state for numerical simulations with COCONUT, several assumptions and simplifications were made during its development, such as prescribing filtered photospheric magnetic maps to represent the magnetic field conditions in the lower corona. This filtering leads to smoothing and lower magnetic field values at the inner boundary (i.e. the solar surface), resulting in an unrealistically high plasma β (greater than 1 in a large portion of the domain). Aims. We aim to examine the effects of prescribing such filtered magnetograms in global coronal simulations and formulate a method for achieving more realistic plasma β values and improving the resolution of electromagnetic features without losing computational performance. Methods. We made use of the newly developed COCONUT solver to demonstrate the effects of the highly pre-processed magnetic maps set at the inner boundary and the resulting high plasma β on the features in the computational domain. Then, in our new approach, we shifted the inner boundary to 2 R⊙ from the original 1.01 R⊙ and preserved the prescribed highly filtered magnetic map. With the shifted boundary, the boundary density and pressure were also naturally adjusted to better represent the considered physical location. This effectively reduces the prescribed plasma β and leads to a more realistic setup. The method was applied on a magnetic dipole, a minimum (2008) and a maximum (2012) solar activity case, to demonstrate its effects. Results. The results obtained with the proposed approach show significant improvements in the resolved density and radial velocity profiles, and far more realistic values of the plasma β at the boundary and inside the computational domain. This is also demonstrated via synthetic white light imaging (WLI) and with the validation against tomography data. The computational performance comparison shows similar convergence to a limit residual on the same grid when compared to the original setup. Considering that the grid can be further coarsened with this new setup, as its capacity to resolve features or structures is superior, the operational performance can be additionally increased if needed. Conclusions. The newly developed method is thus deemed as a good potential replacement of the original setup for operational purposes, providing higher physical detail of the resolved profiles while preserving a good convergence and robustness of the solver.

Space Weather Modeling and Prediction for Intermediate Time-scales

Space Weather Modeling and Prediction for Intermediate Time-scales

Space Weather Observations, Modeling, and Alerts in Support of Human Exploration of Mars

Space Weather Observations, Modeling, and Alerts in Support of Human Exploration of Mars

Interplanetary scintillation observation and space weather modelling

Interplanetary scintillation (IPS) refers to random fluctuations in radio intensity of distant small-diameter celestial object, over time periods of the order of 1 s. The scattering and scintillation of emergent radio waves are ascribed to turbulent density irregularities transported by the ubiquitous solar wind streams. The spatial correlation length of density irregularities and the Fresnel radius of radio diffraction are two key parameters in determining the scintillation pattern. Such a scintillation pattern can be measured and correlated between multi-station radio telescopes on the Earth. Using the “phase-changing screen” scenario based on the Born approximation, the bulk-flow speed and turbulent spectrum of the solar wind streams can be extracted from the single-station power spectra fitting and the multi-station cross-correlation analysis. Moreover, a numerical computer-assisted tomography (CAT) model, iteratively fit to a large number of IPS measurements over one Carrington rotation, can be used to reconstruct the global velocity and density structures in the inner heliosphere for the purpose of space weather modelling and prediction. In this review, we interpret the underlying physics governing the IPS phenomenon caused by the solar wind turbulence, describe the power spectrum and cross correlation of IPS signals, highlight the space weather application of IPS-CAT models, and emphasize the significant benefits from international cooperation within the Worldwide IPS Stations (WIPSS) network.

Global Dynamical Network of the Spatially Correlated Pc2 Wave Response for the 2015 St. Patrick's Day Storm

AbstractWe show the global dynamics of spatial correlation of Pc2 wave activity can track the evolution of the 2015 St. Patrick's Day geomagnetic storm for an 8 hr time window around onset. The global spatially coherent response is tracked by forming a dynamical network from 1 s data using the full set of 100+ ground‐based magnetometer stations collated by SuperMAG and Intermagnet. The pattern of spatial coherence is captured by network parameters which in turn track the evolution of the storm. At onset interplanetary magnetic field (IMF) Bz > 0 and Pc2 power increases, we find a global response with stations correlated over both local and global distances. Following onset, whilst Bz > 0, the network response is confined to the day‐side. When IMF Bz < 0, there is a strong local response at high latitudes, consistent with the onset of polar cap convection driven by day‐side reconnection. The spatially coherent response as revealed by the network grows and is maximal when auroral (SuperMAG electrojet) and ring current (SuperMAG ring current) 1 min resolution geomagnetic indices peak, consistent with an active electrojet and ring‐current. Throughout the storm there is a coherent response both in stations located along lines of constant geomagnetic longitude, between hemispheres, and across magnetic local time. The network does not simply track average Pc2 wave power, it is characterized by network parameters which track the storm evolution. This is the first study to parameterize global Pc2 wave correlation and offers the possibility of statistical studies across multiple events and comparison with, and validation of, space weather models.

Validation of Ionospheric Specifications During Geomagnetic Storms: TEC and foF2 During the 2013 March Storm Event‐II

AbstractAssessing space weather modeling capability is a key element in improving existing models and developing new ones. In order to track improvement of the models and investigate impacts of forcing, from the lower atmosphere below and from the magnetosphere above, on the performance of ionosphere‐thermosphere models, we expand our previous assessment for 2013 March storm event (Shim et al., 2018, https://doi.org/10.1029/2018SW002034). In this study, we evaluate new simulations from upgraded models (the Coupled Thermosphere Ionosphere Plasmasphere Electrodynamics (CTIPe) model version 4.1 and the Global Ionosphere Thermosphere Model (GITM) version 21.11) and from the NCAR Whole Atmosphere Community Climate Model with thermosphere and ionosphere extension (WACCM‐X) version 2.2 including eight simulations in the previous study. A simulation from the NCAR Thermosphere‐Ionosphere‐Electrodynamics General Circulation Model version 2 (TIE‐GCM 2.0) is also included for comparison with WACCM‐X. TEC and foF2 changes from quiet‐time background are considered to evaluate the model performance on the storm impacts. For evaluation, we employ four skill scores: Correlation coefficient (CC), root‐mean square error (RMSE), ratio of the modeled to observed maximum percentage changes (Yield), and timing error (TE). It is found that the models tend to underestimate the storm‐time enhancements of foF2 (F2‐layer critical frequency) and TEC (Total Electron Content) and to predict foF2 and/or TEC better in North America but worse in the Southern Hemisphere. The ensemble simulation for TEC is comparable to results from a data assimilation model (Utah State University‐Global Assimilation of Ionospheric Measurements (USU‐GAIM)) with differences in skill score less than 3% and 6% for CC and RMSE, respectively.

Fast Reconstruction of 3D Density Distribution around the Sun Based on the MAS by Deep Learning

This study is the first attempt to generate a three-dimensional (3D) coronal electron density distribution based on the pix2pixHD model, whose computing time is much shorter than that of the magnetohydrodynamic (MHD) simulation. For this, we consider photospheric solar magnetic fields as input, and electron density distribution simulated with the MHD Algorithm outside a Sphere (MAS) at a given solar radius is taken as output. We consider 155 pairs of Carrington rotations as inputs and outputs from 2010 June to 2022 April for training and testing. We train 152 deep-learning models for 152 solar radii, which are taken up to 30 solar radii. The artificial intelligence (AI) generated 3D electron densities from this study are quite consistent with the simulated ones from lower radii to higher radii, with an average correlation coefficient 0.97. The computing time of testing data sets up to 30 solar radii of 152 deep-learning models is about 45.2 s using the NVIDIA TITAN XP graphics-processing unit, which is much less than the typical simulation time of MAS. We find that the synthetic coronagraphic images estimated from the deep-learning models are similar to the Solar Heliospheric Observatory (SOHO)/Large Angle and Spectroscopic Coronagraph C3 coronagraph data, especially during the solar minimum period. The AI-generated coronal density distribution from this study can be used for space weather models on a near-real-time basis.