Space Weather

Low latitude ionosphere experiences complex dynamical and electrodynamical processes, which make the spatiotemporal variations of the corresponding electron density complicated and therefore influence trans-ionosphere radio communications. The monitoring of low latitude dynamical drivers, such as neutral wind and ionospheric electric field, is essential for both dynamic mechanism investigations and applications. The Sanya Incoherent Scatter Radar Tristatic System (SYISR-TS) was proposed with the main objective of low latitude ionospheric monitoring and investigation and has been successfully developed over the past decade. The system consists of the Sanya (18.3°N, 109.6°E) trans-receiving main station with key parameters of ∼1,600 m2 antenna aperture, >4 MW peak power, <120 K system noise temperature, and ∼46 dBi normal gain, and Danzhou (19.5°N, 109.1°E) and Wenchang (19.6°N, 110.8°E) receiving only stations with key parameters of ∼790 m2 antenna aperture, <130 K system noise temperature, and ∼43 dBi normal gain. Three stations form a quasi-equilateral triangle at Hainan Island and use Global Navigation Satellite System satellite common view technique to achieve the time synchronization with the uncertainty of the timing and time synchronization less than 50 and 10 ns, respectively. Initial collaborative satellite tracking and ionospheric common volume experiments among three stations have confirmed the detection ability of SYISR-TS and the feasibility of achieving its scientific goals in the future.

During geomagnetic storms, the capabilities of current climate models in predicting ionospheric behavior are notably limited. A data assimilation tool, Estimating Model Parameters Reverse Engineering (EMPIRE), implements a Kalman filter to ingest electric density rate correcting the background electric potential and neutral wind. For the baseline setup, or case (1), EMPIRE ingests electron density global map output from the Ionospheric Data Assimilation 4-Dimensional (IDA4D) algorithm. In this work, a new augmentation method is evaluated in which ion drift measurements are also assimilated into EMPIRE. The ion drift measurements used in the new augmentation method are obtained from Super Dual Auroral Radar Network (SuperDARN) sites in the mid-to-high latitude region of the northern hemisphere. Cases (2) and (3) are set up for evaluating the impacts from ingesting different types of observations: SuperDARN fit and grid data, respectively. Six independent data sources are used as validation data sets to compare outcomes with or without ingesting ion drifts. One is the vector ion velocities derived from the Millstone Hill Incoherent Scatter Radar (MHISR) and a second is the vertical drift from Arecibo site. The other four are SuperDARN ion velocity grid data from Saskatoon, Kapuskasing, Christmas Valley West, and Hokkaido East. Results show improvements in performance at mid-latitudes by augmenting electron density rates with 3D spatially distributed line-of-sight ion drift measurements, with negligible improvements to low and high latitude estimations. The lack of improvement at high-latitudes is attributed to the increase in EMPIRE ion drift error poleward of 60° magnetic.

Based on two meteor radars in mid-latitudes of China, the mid-latitude lower thermospheric neutral wind responses to the 2015 St. Patrick's Day great storm are investigated. The AE and PCN indices presented the similar quasi-5-hour oscillations during the storm. Interestingly, the analogous and close-correlated storm-time quasi-5-hour oscillations were also observed in both the meridional wind differences at 90–102 km derived from meteor radars. The meridional wind disturbances in the lower thermosphere also showed the extension toward the lower latitudes. It has been found that the enhanced equatorward wind disturbances at 250 km estimated by the Horizontal Wind Model-14 and Fabry-Perot Interferometer (FPI) emerged accordingly with the increases of AE and PCN with a time delay. And the enhancements of equatorward (poleward) wind disturbances at 250 km were accompanied by the increments of equatorward (poleward) wind disturbances at 94 km with a time lag of a few hours. It is thus suggested that the multiple intensified Joule heating events with quasi-5-hour time intervals were triggered by the successive substorm expansions during the storm. Then the Joule heating events led to the vertical wind and temperature disturbances in the mid-latitude lower thermosphere via disturbing the thermospheric meridional circulation, which consequently induced the quasi-5-hour meridional wind disturbances therein.

Prior to use in operational systems, it is essential to validate ionospheric models in a manner relevant to their intended application to ensure satisfactory performance. For Over-the-Horizon radars (OTHR) operating in the high-frequency (HF) band (3–30 MHz), the problem of model validation is severe when used in Coordinate Registration (CR) and Frequency Management Systems (FMS). It is imperative that the full error characteristics of models is well understood in these applications due to the critical relationship they impose on system performance. To better understand model performance in the context of OTHR, we introduce an ionospheric model validation technique using the oblique ground backscatter measurements in soundings from the Super Dual Auroral Radar Network (SuperDARN). Analysis is performed in terms of the F-region leading edge (LE) errors and assessment of range-elevation distributions using calibrated interferometer data. This technique is demonstrated by validating the International Reference Ionosphere (IRI) 2016 for January and June in both 2014 and 2018. LE RMS errors of 100–400 km and 400–800 km are observed for winter and summer months, respectively. Evening errors regularly exceeding 1,000 km across all months are identified. Ionosonde driven corrections to the IRI-2016 peak parameters provide improvements of 200–800 km to the LE, with the greatest improvements observed during the nighttime. Diagnostics of echo distributions indicate consistent underestimates in model NmF2 during the daytime hours of June 2014 due to offsets of −8° being observed in modeled elevation angles at 18:00 and 21:00 UT.

Solar Energetic Particles (SEPs) form a critical component of Space Weather. The complex, intertwined dynamics of SEP sources, acceleration, and transport make their forecasting very challenging. Yet, information about SEP arrival and their properties (e.g., peak flux) is crucial for space exploration on many fronts. We have recently introduced a novel probabilistic ensemble model called the Multivariate Ensemble of Models for Probabilistic Forecast of Solar Energetic Particles (MEMPSEP). Its primary aim is to forecast the occurrence and physical properties of SEPs. The occurrence forecasting, thoroughly discussed in a preceding paper (MEMPSEP-I by Chatterjee et al., 2024a, https://doi.org/10.1029/2023sw003568), is complemented by the work presented here, which focuses on forecasting the physical properties of SEPs. The MEMPSEP model relies on an ensemble of Convolutional Neural Networks, which leverage a multi-variate data set comprising full-disc magnetogram sequences and numerous derived and in-situ data from various sources (MEMPSEP-III by Moreland et al., 2024, https://doi.org/10.1029/2023SW003765). Skill scores demonstrate that MEMPSEP exhibits improved predictions on SEP properties for the test set data with SEP occurrence probability above 50%, compared to those with a probability below 50%. Results present a promising approach to address the challenging task of forecasting SEP physical properties, thus improving our forecasting capabilities and advancing our understanding of the dominant parameters and processes that govern SEP production.

We introduce a new multivariate data set that utilizes multiple spacecraft collecting in-situ and remote sensing heliospheric measurements shown to be linked to physical processes responsible for generating solar energetic particles (SEPs). Using the Geostationary Operational Environmental Satellites (GOES) flare event list from Solar Cycle (SC) 23 and part of SC 24 (1998–2013), we identify 252 solar events (>C-class flares) that produce SEPs and 17,542 events that do not. For each identified event, we acquire the local plasma properties at 1 au, such as energetic proton and electron data, upstream solar wind conditions, and the interplanetary magnetic field vector quantities using various instruments onboard GOES and the Advanced Composition Explorer spacecraft. We also collect remote sensing data from instruments onboard the Solar Dynamic Observatory, Solar and Heliospheric Observatory, and the Wind solar radio instrument WAVES. The data set is designed to allow for variations of the inputs and feature sets for machine learning (ML) in heliophysics and has a specific purpose for forecasting the occurrence of SEP events and their subsequent properties. This paper describes a data set created from multiple publicly available observation sources that is validated, cleaned, and carefully curated for our ML pipeline. The data set has been used to drive the newly-developed Multivariate Ensemble of Models for Probabilistic Forecast of SEPs (MEMPSEP; see MEMPSEP-I (Chatterjee et al., 2024, https://doi.org/10.1029/2023SW003568) and MEMPSEP-II (Dayeh et al., 2024, https://doi.org/10.1029/2023SW003697) for accompanying papers).

The Sun continuously affects the interplanetary environment through a host of interconnected and dynamic physical processes. Solar flares, Coronal Mass Ejections (CMEs), and Solar Energetic Particles (SEPs) are among the key drivers of space weather in the near-Earth environment and beyond. While some CMEs and flares are associated with intense SEPs, some show little to no SEP association. To date, robust long-term (hours-days) forecasting of SEP occurrence and associated properties (e.g., onset, peak intensities) does not effectively exist and the search for such development continues. Through an Operations-2-Research support, we developed a self-contained model that utilizes a comprehensive data set and provides a probabilistic forecast for SEP event occurrence and its properties. The model is named Multivariate Ensemble of Models for Probabilistic Forecast of Solar Energetic Particles (MEMPSEP). MEMPSEP workhorse is an ensemble of Convolutional Neural Networks that ingests a comprehensive data set (MEMPSEP-III by Moreland et al. (2024, https://doi.org/10.1029/2023SW003765)) of full-disc magnetogram-sequences and in situ data from different sources to forecast the occurrence (MEMPSEP-I—this work) and properties (MEMPSEP-II by Dayeh et al. (2024, https://doi.org/10.1029/2023SW003697)) of a SEP event. This work focuses on estimating true SEP occurrence probabilities achieving a 2.5% improvement in reliability and a Brier score of 0.14. The outcome provides flexibility for the end-users to determine their own acceptable level of risk, rather than imposing a detection threshold that optimizes an arbitrary binary classification metric. Furthermore, the model-ensemble, trained to utilize the large class-imbalance between events and non-events, provides a clear measure of uncertainty in our forecast.

In this paper, we demonstrate the applicability of the data-driven solar energetic particle (SEP) model, SOlar-wind with FIeld-lines and Energetic-particles (SOFIE), to simulate the acceleration and transport processes of SEPs and make forecast of the energetic proton flux at energies ≥10 MeV that will be observed near 1 AU. The SOFIE model is built upon the Space Weather Modeling Framework developed at the University of Michigan. In SOFIE, the background solar wind plasma in the solar corona and interplanetary space is calculated by the Stream-Aligned Aflvén Wave Solar-atmosphere Model(-Realtime) driven by the near-real-time hourly updated Global Oscillation Network Group solar magnetograms. In the background solar wind, coronal mass ejections (CMEs) are launched by placing an force-imbalanced magnetic flux rope on top of the parent active region, using the Eruptive Event Generator using Gibson-Low model. The acceleration and transport processes are modeled by the Multiple-Field-Line Advection Model for Particle Acceleration. In this work, nine SEP events (Solar Heliospheric and INterplanetary Environment challenge/campaign events) are modeled. The three modules in SOFIE are validated and evaluated by comparing with observations, including the steady-state background solar wind properties, the white-light image of the CMEs, and the flux of solar energetic protons, at energies of ≥10 MeV.

In this study, the thermospheric mass density (TMD) features observed by the CHAllenging Minisatellite Payload between 2002 and 2010 were extracted using deep learning (DL) technology; the TMD features were then mapped and modeled with the Interplanetary environment information (IEI), solar radiation, and geomagnetic indices. The DL model was used to simulate the TMD features during Day of Year (DOY) 222–241 in 2014, a period that experienced complex solar-terrestrial environmental variations. We explore the TMD features under different solar-terrestrial environmental conditions and discuss the effects of various inputs by comparing the DL simulation results with satellite observations from Gravity Recovery and Climate Experiment-A and Swarm-A, as well as the simulation results from Jacchia-Bowman 2008, Naval Research Laboratory Mass Spectrometer Incoherent Scatter radar model 2.1, and Drag Temperature Model 2013. These results show that the DL model can better capture the TMD features after adding IEI. Part of these TMD features, including the high-latitude TMD enhancement during the space hurricane event (DOY 232, 2014) and global TMD variations under complex solar-terrestrial environmental disturbances (DOY 222–225, 2014), cannot be well described by the geomagnetic indices. The DL model indicates that the east-west component of the interplanetary magnetic field (IMF By) has a great impact on TMD variations, and its modulation is different from the typical energy injection process during storms. Our results emphasize the crucial influence of IEI on TMD under both geomagnetic disturbances and quiet conditions.

This paper investigates time-lag effects of ionospheric response to two severe geomagnetic storms (Kp = 8) on the degradation of kinematic precise point positioning (PPP) solutions, utilizing over 5500 Global Navigation Satellite Systems (GNSS) stations distributed worldwide. Focusing on these two severe geomagnetic storms that occurred during solar cycle 24, the study employs an open-source positioning software package, namely RTKLIB, to derive the PPP solutions. The findings reveal significant variations in time lags across different magnetic latitudes. These variations are driven by ionospheric responses to a southward interplanetary magnetic field and subsequent decreases in the SMY-H index during the 2015 St. Patrick's Day Storm and the 2017 September 7–8 Storm. Specifically, at high latitudes, PPP degradation primarily manifests during the main phase of the storm, resulting in delays spanning from several minutes to 1–2 hr after the sudden onset of the storm. In contrast, mid- and low latitudes exhibit a wider range of delays extending up to tens of hours. Notably, rapid positioning degradation is observed predominantly at the magnetic local time noon and midnight sectors. The study discusses these time lag effects concerning the intensity of various ionospheric disturbances triggered by the interactions among the solar wind, magnetosphere, and ionosphere during geomagnetic storms. The insights obtained from this research have the potential to be integrated into physics-based and machine-learning models to enhance forecasting capabilities of space weather impacts.

No abstract is available for this article.

Abduallah et al. (2024b, https://doi.org/10.1029/2023sw003824) proposed a novel approach using a deep neural network model, which includes a graph neural network and a bidirectional LSTM layer, named SYMHnet, to forecast the SYM-H index one and 2 hr in advance. Additionally, the network also provides an uncertainty quantification of the predictions. While the approach is innovative, there are some areas where the model's design and implementation may not align with best practices in uncertainty quantification and predictive modeling. We focus on discrepancies in the input and output of the model, which can limit the applicability in real-world forecasting scenarios. This comment aims to clarify these issues, offering detailed insights into how such discrepancies could compromise the model's interpretability and reliability, thereby contributing to the advancement of predictive modeling in space weather research.

This study presents an innovative approach to improving the accuracy and reducing the error convergence time of static Precise Point Positioning (PPP) in Global Positioning System (GPS) navigation. The research focuses on the impact of the high spatial and temporal resolution of a regional ionospheric data assimilation model on PPP over Taiwan. The study further evaluates the performance of both static PPP with the ionospheric information using commonly used models such as Klobuchar and International Reference Ionosphere (IRI), as well as a global ionospheric data assimilation model. Compared to the default IRI, the data assimilated IRI model can improve the overall ionospheric total electron content by approximately 83%. Additionally, it can significantly reduce horizontal positioning errors and shorten the error convergence time more than 52% for static PPP, even during geomagnetic storm events. The study concludes that the high resolution of a regional ionospheric data assimilation model can enhance the accuracy and reduce the error convergence time of PPP navigation and positioning. This research provides valuable insights for future studies in this field, especially in the development of more precise and efficient models for correcting ionospheric delay in GPS navigation.

In this study, based on solar energetic particle (SEP) events classification and a solution of the diffusion equation, we present an efficient system, HITSEP, to predict the intensities in different energy channels (P4 15.0–44.0 MeV, P5 40.0–80.0 MeV, and P6 80.0–165.0 MeV) of energetic proton events observed by GOES spacecraft. The system can predict the rising phase (especially the peak time and peak intensity) of the energetic proton events using only a small amount of data at the beginning of the solar energetic proton events. Among the events that meet the conditions for the use of our prediction system from 2003 to 2017, for P4, P5, and P6 channels, the median Warning Times are 3.70, 2.52, and 1.69 hr; the median Error of the Intensity for events are 0.43, 0.23, 0.34 orders of magnitude; the median Error of the Peak Time for events are 2.53, 0.55, 0.43 hr, respectively. Our system is based on physical mechanisms and has a high accuracy in forecasting the peak intensity with a strict definition of the error. The HITSEP system has huge potential to apply in the space weather forecast. The application of the HITSEP system in space weather forecasting is very promising.

Equatorial Plasma Bubbles (EPBs) can generate ionospheric scintillation at GHz frequencies used in the Global Navigation Satellite System (GNSS). Emerging at any longitude following sunset and typically moving eastward, monitoring the EPBs is essential for space weather services. Using three GNSS receivers positioned at the same latitude (∼0°N) but separated in longitudes (∼9°, ∼16°, and ∼25°) and the 47 MHz Equatorial Atmosphere Radar (EAR) in Indonesia, our study delineates the zonal extent of eastward-traveling post-sunset EPB inducing ionospheric GNSS scintillation. Typically, the scintillation occurrences detected by a ground receiver concentrate between 19 and 01 local time (LT), with a peak incidence observed at 21 LT. Furthermore, an experiment combining EAR observations with GNSS receiver data allowed for the determination of the linear change in the speed of eastward-traveling EPB inducing scintillation during this time period. Interestingly, the longitudinal range of eastward-traveling EPBs increased with higher solar flux (F10.7) levels. Our findings suggest that EPB can induce scintillation up to a longitudinal distance of approximately 25° from the onset location at sunset to the eastern midnight region, particularly in F10.7 ranging from 90 to 150 solar flux units. Moreover, experiments using longitudinally separated GNSS receivers indicated that scintillations during 19–01 LT originate from post-sunset EPBs within a longitudinal range extending 25° to the west. In conclusion, our research provides valuable insight into the ability of eastward-traveling EPB to induce GNSS scintillation within a longitudinal range of 25°, thereby enhancing EPB and scintillation monitoring and prediction in regional space weather services.

Regression models (LEEMYR: Low Energy Electron MLT geosYnchronous orbit Regression) predict hourly 4.1–30 keV electron flux at geostationary orbit (GOES-16) using solar wind, IMF, and geomagnetic index parameters. Multiplicative interaction and polynomial terms describe synergistic and nonlinear effects. We reduce predictors to an optimal set using stepwise regression, resulting in models with validation comparable to a neural network. Models predict 1, 3, 6, 12, and 24 hr into the future. Validation correlations are as high as 0.78 (4.1 and 11 keV, 1 hr prediction) and Heidke Skill scores (HSS) up to 0.66. A 3 hr ahead prediction is more practical, with slightly lower validation correlation (0.75) and HSS (0.61). The addition of location (MLT: magnetic local time) as a covariate, including multiplicative interaction terms, accounts for location-dependent flux differences and variation of parameter influence, and allows prediction over the full orbit. Adding a substorm index (SME) provides minimal increase in validation correlation (0.81) showing that other parameters are good proxies for an unavailable real time substorm index. Prediction intervals on individual values provide more accurate assessments of model quality than confidence intervals on the mean values. An inverse N-weighted least squares approach is impractical as it increases false positive warnings. Physical interpretations are not possible as spurious correlations due to common cycles are not removed. However, SME, Bz, Kp, and Dst are the highest correlates of electron flux, with solar wind velocity, density, and pressure, and IMF magnitude being less well correlated.

Recent studies have shown that the measurements of Langmuir Probes (LPs) onboard ESA's Swarm mission overestimate ion densities on the nightside by up to 50%. The overestimation is due to the assumption of oxygen-only plasma for ion density calculations, which is often violated at mid-latitudes on the nightside. In this study, we present a calibration model that resolves the nighttime overestimation by Swarm LPs. Using observations by Swarm FacePlate (FP) as a reference, we develop a neural network (NN) model that adjusts LP data to the FP measurements. The model incorporates dependence on solar and geomagnetic conditions, parameterized by the P10.7 and Hp30 indices, location, day of the year and local time. Our model reveals a distinct double-crest pattern in nighttime density overestimation by LPs, centered at ∼30° quasi-dipole latitude in both hemispheres. This overestimation intensifies during low solar activity and shows strong seasonal dependence. During solstices, the crests are more pronounced in the local winter hemispheres, while during equinoxes the crests are weaker and exhibit hemispheric symmetry. This morphology aligns with the presence of light ions diffusing downward from the plasmasphere. Validating the LP data in conjunctions with Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) observations showed a much stronger agreement after applying the developed correction: for Swarm B, nighttime correlation with COSMIC increased from 0.74 to 0.93. The NN-calibrated LP data set has numerous applications in ionospheric research, and the developed model can provide useful insights into the ion composition in the topside ionosphere.

Specification and forecast of ionospheric parameters, such as ionospheric electron density (Ne), have been an important topic in space weather and ionospheric research. Neural networks (NNs) emerge as a powerful modeling tool for Ne prediction. However, heavy manual adjustments are time consuming to determine the optimal NN structures. In this work, we propose to use neural architecture search (NAS), an automatic machine learning method, to mitigate this problem. NAS aims to find the optimal network structure through the alternate optimization of the hyperparameters and the corresponding network parameters within a pre-defined hyperparameter search space. A total of 16-year data from Millstone Hill incoherent scatter radar (ISR) are used for the NN models. One single-layer NN (SLNN) model and one deep NN (DNN) model are both trained with NAS, namely SLNN-NAS and DNN-NAS, for Ne prediction and compared with their manually tuned counterparts (SLNN and DNN) based on previous studies. Our results show that SLNN-NAS and DNN-NAS outperformed SLNN and DNN, respectively. These NN predictions of Ne daily variation patterns reveal a 27-day mid-latitude topside Ne variation, which cannot be reasonably represented by traditional empirical models developed using monthly averages. DNN-NAS yields the best prediction accuracy measured by quantitative metrics and rankings of daily pattern prediction, especially with an improvement in mean absolute error more than 10% compared to the SLNN model. The limited improvement of NAS is likely due to the network complexity and the limitation of fully connected NN without the time histories of input parameters.

Electron fluxes in the keV energy range can cause significant spacecraft surface charging, which in turn can affect the functioning of spacecraft components. In this paper, the geostationary electron fluxes measured by the satellites GOES 13-18 in the energy range 2–200 keV are analyzed in order to look for their dependence on solar wind conditions. For this purpose, a range of solar wind parameters, IMF parameters and geomagnetic indices are examined, to look for the parameters which most significantly affect the electron flux. The analysis includes fluxes in the lower energy range of 2–40 keV, measured by GOES 16-18, which have not been analyzed before. The measured electron fluxes are averaged over all directions, and high-pass filtered to isolate variations shorter than 1 month. The analysis concentrates of the dawn sector, where variations are largest. A number of solar wind parameters and magnetic indices are analyzed concurrently with the electron flux data, to look for the most significant correlations between them. Most parameters have the highest correlation with electron flux when shifted in time by a certain delay. In addition to the different solar wind parameters and magnetic indices, combinations of different parameters are also examined for their best correlation with the electron flux. The most significant driving parameters are found to be the auroral electrojet index, combined with either the solar wind plasma velocity or the plasma density. The relative contribution of each of these parameters depends on electron energy, and differs between periods of high and low flux.

Low Earth orbit (LEO) radio occultation|radio occultations (RO) constellations can provide global electron density profiles (EDPs) to better specify and forecast the ionosphere-thermosphere (I-T) system. To inform future RO constellation design, this study uses comprehensive Observing System Simulation Experiments (OSSEs) to assess the ionospheric specification impact of assimilating synthetic EDPs into a coupled I-T model. These OSSEs use 10 different sets of RO constellation configurations containing 6 or 12 LEO satellites with base orbit parameter combinations of 520 or 800 km altitude, and 24° or 72° inclination. The OSSEs are performed using the Ensemble Adjustment Kalman Filter implemented in the data assimilation (DA) Research Testbed and the Thermosphere-Ionosphere-Electrodynamics General Circulation Model (TIEGCM). A different I-T model is used for the nature run, the Whole Atmosphere Model-Ionosphere Plasmasphere Electrodynamics (WAM-IPE), to simulate the period of interest is the St. Patrick's Day storm on March 13–18, 2015. Errors from models and EDP retrieval are realistically accounted for in this study through distinct I-T models and by retrieving synthetic EDPs through an extension Abel inversion algorithm. OSSE assessment, using multiple metrics, finds that greater EDP spatial coverage leading to improved specification at altitudes 300 km and above, with the 520 km altitude constellations performing best due to yielding the highest observation counts. A potential performance limit is suggested with two 6-satellite constellations. Lastly, close examination of Abel inversion error impacts highlights major EDP limitations at altitudes below 200 km and dayside equatorial regions with large horizontal gradients and low electron density magnitudes.