Space Weather

In this work, a model for calculating the galactic cosmic rays (GCRs) radiation dose rate is developed. The model is based on a GCR modulation model, which is established by Shen and Qin, and the fluence-dose conversion coefficients (FDCCs) published by the International Commission on Radiological Protection (ICRP). With the model, the radiation absorbed dose rate of GCRs near the lunar surface over long time periods is calculated and compared with the observation data from the Cosmic Ray Telescope for the Effects of Radiation and the Lunar Lander Neutron and Dosimetry. First, the energy spectrum of GCRs at 1 AU in the ecliptic, where the lunar orbit is located, is computed using the GCR modulation model. Then, using the FDCCs of ICRP 123, the absorbed dose rates of 15 human organs/tissues at the lunar orbit position are calculated to represent the general absorbed dose rate of the body (in water). Furthermore, considering the albedo radiation (excluding neutrons) and using the water-silicon conversion coefficients, the total absorbed dose rates of GCRs near the lunar surface (in silicon) are calculated, it is shown that our modeling results generally agree with the observations from spacecraft. This work is useful for future manned space exploration to the Moon or other celestial bodies in the solar system.

Current magnetohydrodynamics (MHD) models largely rely on synoptic magnetograms, such as the ones produced by the Global Oscillation Network Group (GONG). Magnetograms are currently available mostly from the front side of the Sun, which significantly reduces the accuracy of MHD modeling. Extreme Ultraviolet (EUV) images can instead be obtained from other vantage points. To investigate the potential, we explore the possibility of using EUV information from the Atmospheric Imaging Assembly (AIA) to directly generate the input for the state-of-the-art 3D MHD model European Heliospheric FORecasting Information Asset (EUHFORIA). Toward this goal, we develop a method called Transfer-Solar-GAN which combines a conditional generative adversarial network with a transfer learning approach to overcome training data set limitations. The source domain data set is constructed from multiple pairs of the central portion of co-registered AIA and Helioseismic and Magnetic Imager (HMI) line of sight (LOS) full-disk images, while the target domain is constructed from pairs of portions of AIA and GONG sine-latitude synoptic maps that we call segments. We evaluate Transfer-Solar-GAN by comparing modeled and measured solar wind velocity and magnetic field density parameters at the L 1 Lagrange point and along the Parker Solar Probe (PSP) trajectory which were determined with EUHFORIA using both empirical GONG and artificial-intelligence (AI)-synthetic synoptic magnetograms as inputs. Our results demonstrate that the Transfer-Solar-GAN model can provide the necessary information to run solar physics models by EUV information. Our proposed model is trained with only 528 paired image segments and enforces a reliable data division strategy.

Signals recorded by two stations in the Brazilian region: [Fortaleza (3.74°S, 38.57°W) and Inconfidentes (22.31°S, 46.32°W)], receiving L1 transmissions from different geostationary satellites, were used to evaluate the amplitude scintillation index S 4 and several characteristics of scintillation events (continuous record with S 4 > 0.2) during nighttime hours (18:00 LT–02:00 LT) in the years 2014–2016. The effects from solar activity, season, and local time on the number of scintillation events per night, maximum scintillation, scintillation event duration, and spacing between consecutive events will be discussed. The results indicate that: (a) scintillation occurs from September to March in both links; (b) the most likely numbers of observed scintillation events per night were two or three, particularly during the first 2 years; (c) on average, the first scintillation event usually had larger maximum S 4 values when compared to those of the later ones along the night; (d) the first scintillation event had a longer mean duration than the succeeding ones in a given night; (e) the durations of scintillation events, regardless of their numbers per night and the location, decreased with local time; (f) the opposite dependence of spacings between consecutive events on local time was observed; (g) the cumulative distribution functions of the scintillation onset time indicated a strong dependence on the dip latitude of the station; and (h) early occurrences of onset times are directly related to the increased probability of the occurrence of multiple scintillation events.

The propagation of geoeffective fast halo coronal mass ejections (CMEs) from solar cycle 24 has been investigated using the European Heliospheric Forecasting Information Asset (EUHFORIA), ENLIL, Drag-Based Model (DBM) and Effective Acceleration Model (EAM) models. For an objective comparison, a unified set of a small sample of CME events with similar characteristics has been selected. The same CME kinematic parameters have been used as input in the propagation models to compare their predicted arrival times and the speed of the interplanetary (IP) shocks associated with the CMEs. The performance assessment has been based on the application of an identical set of metrics. First, the modeling of the events has been done with default input concerning the background solar wind, as would be used in operations. The obtained CME arrival forecast deviates from the observations at L1, with a general underestimation of the arrival time and overestimation of the impact speed (mean absolute error [MAE]: 9.8 ± 1.8–14.6 ± 2.3 hr and 178 ± 22–376 ± 54 km/s). To address this discrepancy, we refine the models by simple changes of the density ratio (dcld) between the CME and IP space in the numerical, and the IP drag (γ) in the analytical models. This approach resulted in a reduced MAE in the forecast for the arrival time of 8.6 ± 2.2–13.5 ± 2.2 hr and the impact speed of 51 ± 6–243 ± 45 km/s. In addition, we performed multi-CME runs to simulate potential interactions. This leads, to even larger uncertainties in the forecast. Based on this study we suggest simple adjustments in the operational settings for improving the forecast of fast halo CMEs.

The F 10.7 solar radiation flux is a well-known parameter that is closely linked to solar activity, serving as a key index for measuring the level of solar activity. In this study, the Variational Mode Decomposition (VMD) and Long Short-term Memory (LSTM) network are combined to construct a VMD-LSTM model for predicting F 10.7 values. The F 10.7 sequence is decomposed into several intrinsic mode functions (IMF) by VMD, then the LSTM neural network is utilized to forecast each IMF. All IMF prediction results are aggregated to obtain the final F 10.7 value. The data sets from 1957 to 2008 are used for training and the data sets from 2009 to 2019 are used for testing. The results show that the VMD-LSTM model achieves an annual average root mean square error of only 4.47 sfu and an annual average correlation coefficient (R) of 0.99 during solar cycle 24, which is significantly better than the accuracy of the LSTM model (W. Zhang et al., 2022, https://doi.org/10.3390/universe8010030), the AR model (Du, 2020, https://doi.org/10.1007/s11207-020-01689-x), and the BP model (Xiao et al., 2017, https://doi.org/10.11728/cjss2017.01.001). The VMD-LSTM model exhibits strong predictive capability for the F 10.7 index during solar cycle 24.

In this work, convolutional neural networks (CNN) are developed to detect and characterize sporadic E (E s ), demonstrating an improvement over current methods. This includes a binary classification model to determine if E s is present, followed by a regression model to estimate the E s ordinary mode critical frequency (foEs), a proxy for the intensity, along with the height at which the E s layer occurs (hEs). Signal-to-noise ratio (SNR) and excess phase profiles from six Global Navigation Satellite System (GNSS) radio occultation (RO) missions during the years 2008–2022 are used as the inputs of the model. Intensity (foEs) and the height (hEs) values are obtained from the global network of ground-based Digisonde ionosondes and are used as the “ground truth,” or target variables, during training. After corresponding the two data sets, a total of 36,521 samples are available for training and testing the models. The foEs CNN binary classification model achieved an accuracy of 74% and F1-score of 0.70. Mean absolute errors (MAE) of 0.63 MHz and 5.81 km along with root-mean squared errors (RMSE) of 0.95 MHz and 7.89 km were attained for estimating foEs and hEs, respectively, when it was known that E s was present. When combining the classification and regression models together for use in practical applications where it is unknown if E s is present, an foEs MAE and RMSE of 0.97 and 1.65 MHz, respectively, were realized. We implemented three other techniques for sporadic E characterization, and found that the CNN model appears to perform better.

Changes in the Earth's geomagnetic field induce geoelectric fields in the solid Earth. These electric fields drive Geomagnetically Induced Currents (GICs) in grounded, conducting infrastructure. These GICs can damage or degrade equipment if they are sufficiently intense—understanding and forecasting them is of critical importance. One of the key magnetospheric phenomena are Sudden Commencements (SCs). To examine the potential impact of SCs we evaluate the correlation between the measured maximum GICs and rate of change of the magnetic field (H′) in 75 power grid transformers across New Zealand between 2001 and 2020. The maximum observed H′ and GIC correlate well, with correlation coefficients (r 2) around 0.7. We investigate the gradient of the relationship between H′ and GIC, finding a hot spot close to Dunedin: where a given H′ will drive the largest relative current (0.5 A nT−1 min). We observe strong intralocation variability, with the gradients varying by a factor of two or more at adjacent transformers. We find that GICs are (on average) greater if they are related to: (a) Storm Sudden Commencements (SSCs; 27% larger than Sudden Impulses, SIs); (b) SCs while New Zealand is on the dayside of the Earth (27% larger than the nightside); and (c) SCs with a predominantly East-West magnetic field change (14% larger than North-South equivalents). These results are attributed to the geology of New Zealand and the geometry of the power network. We extrapolate to find that transformers near Dunedin would see 2000 A or more during a theoretical extreme SC (H′ = 4000 nT min−1).

We investigate the role of auroral particle precipitation in small-scale (below hundreds of meters) plasma structuring in the auroral ionosphere over the Arctic. In this scope, we analyze together data recorded by an Ionospheric Scintillation Monitor Receiver (ISMR) of Global Navigation Satellite System (GNSS) signals and by an All-Sky Imager located in Longyearbyen, Svalbard (Norway). We leverage on the raw GNSS samples provided at 50 Hz by the ISMR to evaluate amplitude and phase scintillation indices at 1 s time resolution and the Ionosphere-Free Linear Combination at 20 ms time resolution. The simultaneous use of the 1 s GNSS-based scintillation indices allows identifying the scale size of the irregularities involved in plasma structuring in the range of small (up to few hundreds of meters) and medium-scale size ranges (up to few kilometers) for GNSS frequencies and observational geometry. Additionally, they allow identifying the diffractive and refractive nature of fluctuations on the recorded GNSS signals. Six strong auroral events and their effects on plasma structuring are studied. Plasma structuring down to scales of hundreds of meters is seen when strong gradients in auroral emissions at 557.7 nm cross the line of sight between the GNSS satellite and receiver. Local magnetic field measurements confirm small-scale structuring processes coinciding with intensification of ionospheric currents. Since 557.7 nm emissions primarily originate from the ionospheric E-region, plasma instabilities from particle precipitation at E-region altitudes are considered to be responsible for the signatures of small-scale plasma structuring highlighted in the GNSS scintillation data.

This study evaluates the performance of deep learning approach in the prediction of the ionospheric total electron content (TEC) during magnetically quiet periods. Two deep learning techniques, long short-term memory (LSTM) and convolutional LSTM (ConvLSTM), are employed to predict TEC values 24 hr ahead in the vicinity of the Korean Peninsula (26.5°–40°N, 121°–134.5°E). The LSTM method predicts TEC at a single point based on time series of data at that point, whereas the ConvLSTM method simultaneously predicts TEC values at multiple points using spatiotemporal distribution of TEC. Both the LSTM and ConvLSTM models are trained using the complete regional TEC maps reconstructed by applying the Deep Convolutional Generative Adversarial Network–Poisson Blending (DCGAN-PB) method to observed TEC data. The training period spans from 2002 to 2018, and the model performance is evaluated using 2019 data. Our results show that the ConvLSTM method outperforms the LSTM method, generating more reliable TEC maps with smaller root mean square errors when compared to the ground truth (DCGAN-PB TEC maps). This outcome indicates that deep learning models can improve the prediction accuracy of TEC at a specific point by taking into account spatial information of TEC. We conclude that ConvLSTM is a reliable and efficient approach for the prompt ionospheric prediction.

A new one-dimensional variational (1D-Var) retrieval method for ionospheric GNSS radio occultation (GNSS-RO) measurements is described. The forward model implicit in the retrieval calculates the bending angles produced by a one-dimensional ionospheric electron density profile, modeled with multiple “Vary-Chap” layers. It is demonstrated that gradient based minimization techniques can be applied to this retrieval problem. The use of ionospheric bending angles is discussed. This approach circumvents the need for Differential Code Bias (DCB) estimates when using the measurements. This new, general retrieval method is applicable to both standard GNSS-RO retrieval problems, and the truncated geometry of EUMETSAT's Metop Second Generation (Metop-SG), which will provide GNSS-RO measurements up to about 600 km above the surface. The climatological a priori information used in the 1D-Var is effectively a starting point for the 1D-Var minimization, rather than a strong constraint on the final solution. In this paper the approach has been tested with 143 COSMIC-1 measurements. We find that the method converges in 135 of the cases, but around 25 of those have high “cost at convergence” values. In the companion paper (Elvidge et al., 2023), a full statistical analysis of the method, using over 10,000 COSMIC-2 measurements, has been made.

Culverwell et al. (2023, https://doi.org/10.1029/2023SW003572) described a new one-dimensional variational (1D-Var) retrieval approach for ionospheric GNSS radio occultation (GNSS-RO) measurements. The approach maps a one-dimensional ionospheric electron density profile, modeled with multiple “Vary-Chap” layers, to bending angle space. This paper improves the computational performance of the 1D-Var retrieval using an improved background model and validates the approach by comparing with the COSMIC-2 profile retrievals, based on an Abel Transform inversion, and co-located (within 200 km) ionosonde observations using all suitable data from 2020. A three or four layer Vary-Chap in the 1D-Var retrieval shows improved performance compared to COSMIC-2 retrievals in terms of percentage error for the F2 peak parameters (NmF2 and hmF2). Furthermore, skill in retrieval (compared to COSMIC-2 profiles) throughout the bottomside (∼90–300 km) has been demonstrated. With a single Vary-Chap layer the performance is similar, but this improves by approximately 40% when using four-layers.

Taiji is proposed as a space-based gravitational wave (GW) observatory consisting of three spacecraft in a heliocentric orbit meanwhile with the distance of 3 million kilometers ahead of the Earth at about 20°. Free-falling test masses (TMs) are a key component of the interferometer for space-based GW detection in the 0.1mHz–1 Hz frequency range. Exposure to energetic particles in the space environment can lead to charging of the TMs and thus cause additional electrostatic forces and Lorentz forces that limit the sensitivity of the interferometer and may affect the quality of the scientific data. This study aims to model the charging of TMs during Galactic cosmic rays and solar proton events (SPEs) using the Monte Carlo simulation toolkit meanwhile with constructing the sophisticated 3D spacecraft. The results show that the total net charging rates are 34.48 +e/s and 33.85 +e/s on TM1 and TM2 during the solar minimum, and 9.58 +e/s on TM1 and 9.65 +e/s on TM2 during the solar maximum. We confirm that no matter for solar minimum or solar maximum, protons contribute to the largest proportion of the TMs charging rate. Furthermore, charging for five typical SPEs is also investigated, and the maximum TMs charging rate reaches 76,674 +e/s, indicating that sporadic SPEs have a high risk for TMs charging. Finally, the charging rates of a TM imitation are tested on ground by the 30–50 MeV proton irradiation experiment, and the experimental results show good consistence with the simulation results with the error <10%.

In this paper, we constructed the Exospheric Temperature Models (ETM) on the basis of CHAMP and GRACE data using different empirical orthogonal functions (EOFs). The EOFs of the exospheric temperature can be derived either from satellite data directly or from the outputs of the Thermosphere Ionosphere Electrodynamics General Circulation Model (TIEGCM) and MSIS models by applying the Principal Component Analysis method. Then, the thermospheric mass densities calculated from ETM are used to compare with the observed data in order to evaluate the performance of different ETM models. It was found that all these three models can provide good specification of thermospheric density including day-night, seasonal, and latitudinal variations. However, the ETM based on CHAMP and GRACE data gives a better performance in modeling the Equatorial Thermospheric Anomaly and the Midnight Density Maximum features than the MSIS-ETM and TIEGCM-ETM. Specifically, independent SWARM-C data comparison showed that the Relative Deviations and corresponding Root-Mean-Square-Errors of our T exo models are less than 8.9% and 22.8%, much better than the MSIS-00 model.