A storm-time statistical enhancement of NRLMSIS 2.1 thermospheric density during the May 2024 extreme geomagnetic storm

Temporal variation and spatial distribution of the thermospheric density can change significantly during geomagnetic storms. These variations in thermospheric density enhance atmospheric drag, posing risks to Low-Earth-Orbit (LEO) spacecraft. Therefore, studying the characteristics of intense storm-time thermospheric density perturbations and orbit decay is crucial for practical applications. In this study, neutral density was simulated for the strongest magnetic storm events of solar cycles 24, 23, and 22, corresponding to minimum Dst indices of −234 nT (2015 St. Patrick’s Day storm), −442 nT (20 November 2003 storm), and −598 nT (1989 Quebec blackout storm). Four representative thermospheric models (DTM-2020, JB 2008, NRLMSIS 2.0, and TIEGCM 2.0) were employed to evaluate their performance during extreme geomagnetic storms by comparing simulated densities with satellite observations from Swarm, CHAMP, and GRACE during the November 2003 and March 2015 storm events. The results indicate that the errors of all models exhibit larger errors in the main and recovery phases, with a bias toward underestimation of density during the main phase. It is important to note that no thermospheric model is perfect and each model has its own limitations, especially dealing with extreme space weather events. Although JB2008 performs relatively well, it does not maintain the best performance across all phases, and its predictions still deviate from observations by at least 20%. Therefore, combining multiple model outputs is recommended for extreme cases. Furthermore, these thermospheric models were coupled with the High-Precision Orbit Propagator (HPOP) to examine the orbital decay of the China Space Station (CSS, ∼380–400 km altitude) during these events. The effects of drag on CSS during the strongest magnetic storm events in the 24th, 23rd and 22nd solar cycles were simulated. The orbital decay is about 233%, 300% and 266% higher than that in the quiet period, respectively. The reults of this study might serve as a reference for spacecraft for possible upcoming extreme magnetic storm events.

Machine learning (ML) has been applied to space weather problems with increasing frequency in recent years, driven by an influx of in-situ measurements and a desire to improve modeling and forecasting capabilities throughout the field. Space weather originates from solar perturbations and is comprised of the resulting complex variations they cause within the numerous systems between the Sun and Earth. These systems are often tightly coupled and not well understood. This creates a need for skillful models with knowledge about the confidence of their predictions. One example of such a dynamical system highly impacted by space weather is the thermosphere, the neutral region of Earth’s upper atmosphere. Our inability to forecast it has severe repercussions in the context of satellite drag and computation of probability of collision between two space objects in low Earth orbit (LEO) for decision making in space operations. Even with (assumed) perfect forecast of model drivers, our incomplete knowledge of the system results in often inaccurate thermospheric neutral mass density predictions. Continuing efforts are being made to improve model accuracy, but density models rarely provide estimates of confidence in predictions. In this work, we propose two techniques to develop nonlinear ML regression models to predict thermospheric density while providing robust and reliable uncertainty estimates: Monte Carlo (MC) dropout and direct prediction of the probability distribution, both using the negative logarithm of predictive density (NLPD) loss function. We show the performance capabilities for models trained on both local and global datasets. We show that the NLPD loss provides similar results for both techniques but the direct probability distribution prediction method has a much lower computational cost. For the global model regressed on the Space Environment Technologies High Accuracy Satellite Drag Model (HASDM) density database, we achieve errors of approximately 11% on independent test data with well-calibrated uncertainty estimates. Using an in-situ CHAllenging Minisatellite Payload (CHAMP) density dataset, models developed using both techniques provide test error on the order of 13%. The CHAMP models—on validation and test data—are within 2% of perfect calibration for the twenty prediction intervals tested. We show that this model can also be used to obtain global density predictions with uncertainties at a given epoch.

On 03 February 2022, SpaceX launched 49 Starlink satellites, 38 of which re‐entered the atmosphere on or about 07 February 2022 due to unexpectedly high atmospheric drag. We use empirical model (NRLMSIS, JB08, and HASDM) outputs as well as solar extreme ultraviolet occultation and high‐fidelity accelerometer data to show that thermospheric density was at least 20%–30% higher at 210 km relative to the 9 days prior to the launch due to consecutive geomagnetic storms related to solar eruptions from NOAA AR12936 on 29 January 2022. We model the orbital altitude and in‐track position of a Starlink‐like satellite in a low‐drag configuration at 200 km during minor (G1) and extreme (G5) geomagnetic storms to show that an extreme storm would have at least a factor of two higher impact, with cumulative in‐track errors on the order of 10,000 km after a 5‐day duration extreme storm. Comparison of the JB08 and NRL MSIS models relative to the HASDM model during modeled historical minor and extreme geomagnetic storms shows that in‐track errors on the order of 100 km per day at 250 km, decreasing to cumulative errors on the order of 1 km per day at 550 km during geomagnetic storms. We conclude that full‐physics, data assimilative, coupled models of the magnetosphere and upper atmosphere, as well as new operational satellite missions providing “nowcasting” data to launch controllers, space traffic coordinators, and satellite operators, are needed to prevent similar—or worse—orbital system impacts during future geomagnetic storms.

In this study, long-term driver-response relationships of thermospheric mass density variations are derived from 12 year of GRACE accelerometer measurements. In order to obtain long-term variations, diurnal and annual variations have been modeled and removed from the main modes. The global distribution of the averaged thermospheric air mass density shows a clear alignment with vertical magnetic field component. Two asymmetric cells located in each eastern polar side show a southern enhancement and a northern attenuation. The latitudinal variation between the northern and southern polar-regions is positive during all time-span, and strongly controlled by the solar radiation. The reconstructed solution has been averaged and parameterized in terms of the solar-flux F10.7, and geomagnetic Am indices. The results show a considerable magnetospheric forcing on the long-term global thermospheric mass density, which has not been reported previously. With this first results and our planned research, we expect soon to provide an accurate air mass density model for the thermospheric research community.

The charge to the space science community is to improve specification and forecast of the low Earth orbit (LEO) space environment to provide reliable collision avoidance and risk assessment analyses for space traffic management and reduce the number of false conjunction warnings. The challenge is that the prediction of LEO object trajectories is severely limited. This is due primarily to poorly captured variability in neutral density estimates during space weather events, resulting in large and variable position errors of all resident space objects across LEO. To improve operations in LEO, the specification and forecast of the thermosphere neutral mass density must improve.Most recently launched LEO satellites are equipped with global navigation satellite system (GNSS) devices, making them excellent sources of continuous orbit ephemeris to enable precision orbit determination (POD). Many are also equipped with attitude and vehicle knowledge to allow for the construction of an accurate force model. Combining these “data of opportunity” from LEO satellites with POD processing tools offers the possibility of extracting thermospheric mass density information regularly and globally from the multitude of GNSS-equipped LEO satellites in operation today.This talk explores this possible trove of LEO space environment data by investigating methods and providing specificity to the level of data information required to provide useful mass density outcomes. The ICESAT-2 spacecraft is used as a test vehicle for this type of analysis. The NASA GSFC GEODYN POD software is employed to produce precise science orbits for the ICESAT-2 spacecraft. These precise science orbits are then used to extract mass density estimates along specified orbital arcs. Simulation is also employed to address the potential errors of system requirements and how they can influence the thermospheric mass density estimate from LEO spacecraft. The future GDC mission will also be highlighted as a much-needed LEO space environment resource for direct multi-point measurements of the thermospheric gas and ionospheric plasma. These direct neutral measurements will enable a careful validation of POD-extracted densities. By fully characterizing all free-stream parameters, GDC is also a well-equipped constellation for studying the gas-surface interactions critical for drag research.

Extreme space weather can be defined as an event with variability in solar, ionospheric, thermospheric or magnetospheric parameters that cause significant degradation of certain technological systems or pose a serious risk to human health. Space weather effects/impacts can be divided into five broad groups: 1) ionospheric effects; 2) ionizing radiation/plasma effects on s/c electronics; 3) ionizing radiation effects on human health; 4) atmospheric uplifting and satellite drag; 5) geomagnetically induced currents (GICs). Extreme space weather can be traced to their solar and interplanetary origins, including solar flares, interplanetary coronal mass ejections and shock waves, and solar energetic particles. By fitting the tail of the cumulative distributions of space weather parameters, it is possible to estimate the probability of extremes. Historical examples show that before the space age, the most common examples of extreme space weather were extreme geomagnetic storms with intense GICs in ground-based infrastructure including power grids. Advances in technologies are expanding the area of extreme space weather. With the number of space objects increasing dramatically each year, satellite drag effects are becoming a concern. With thousands of aircraft maneuvering in response to a geomagnetic storm, it could theoretically be possible to achieve extreme space weather conditions for an event that would otherwise be considered moderate.

In this paper, globally-averaged, thermospheric total mass density, derived from the orbits of ∼5000 objects at 250, 400, and 550 km that were tracked from 1967 to 2006, has been used to quantitatively study the annual asymmetry of thermospheric mass density and its mechanism(s). The results show that thermospheric mass density had a significant annual asymmetry, which changed from year to year. The annual asymmetry at the three altitudes varied synchronously and its absolute value increased with altitudes. The results suggest that there is an annual asymmetry in solar EUV radiation that is caused by the difference in the Sun-Earth distance between the two solstices and the random variation of solar activity within a year. This change in radiation results in an annual change in the thermospheric temperature and thus the scale height of the neutral gas, and is the main cause of the annual asymmetry of thermospheric mass density. The annual asymmetry of mass density increases with altitude because of the accumulating effect of the changes in neutral temperature and scale height in the vertical direction.

We review impacts of satellite drag and describe past, current and future capabilities designed to meet evolving operational requirements. Historically, thermospheric research has been data starved. Thus, from the early space age to the end of the 20th century little progress was made in satellite-drag modeling. This condition improved greatly with the development of empirical assimilative models and recent availability of comprehensive drag measurements. The resurgence in orbital drag analyses to specify thermospheric densities has been particularly useful for addressing input requirements of assimilation models as well as their development and validation. With the new Jacchia-Bowman 2006 model the status of empirical modeling improved significantly. It builds on the expanded satellite drag database and incorporates improved estimates of solar flux changes as well as semiannual and local time variations of the thermosphere. However, magnetic storm representations of Jacchia-Bowman 2006 are similar to those used in other current models. Satellite-borne accelerometers and optical sensors now provide complementary spatial and temporal capabilities that permit monitoring the thermosphere over a wide range of altitudes under most solar and geomagnetic conditions. Long-standing shortfalls during periods of high geomagnetic activity are now being attacked with these data and through new analyses of solar wind and IMF measurements, correlations with magnetosphere-based magnetic indices and emerging theoretical tools. These advances in understanding thermospheric coupling during magnetic storms will be incorporated into empirical model upgrades. The analyses of new data sets joined with on-going research on physical thermosphere-ionosphere- magnetosphere coupling processes support the pursuit of our ultimate goal, an assimilative and predictive operational model of thermospheric neutral densities.

A candidate auroral report in the Bamboo Annals, indicating a possible extreme space weather event in the early 10th century BCE

Given the infrequency of extreme geomagnetic storms, it is significant to note the concentration of three extreme geomagnetic storms in 1941, whose intensities ranked fourth, twelfth, and fifth within the aa index between 1868–2010. Among them, the geomagnetic storm on 1941 March 1 was so intense that three of the four Dst station magnetograms went off scale. Herein, we reconstruct its time series and measure the storm intensity with an alternative Dst estimate (Dst*). The source solar eruption at 09:29–09:38 GMT on February 28 was located at RGO AR 13814 and its significant intensity is confirmed by large magnetic crochets of ∣35∣ nT measured at Abinger. This solar eruption most likely released a fast interplanetary coronal mass ejection with estimated speed 2260 km s−1. After its impact at 03:57–03:59 GMT on March 1, an extreme magnetic storm was recorded worldwide. Comparative analyses on the contemporary magnetograms show the storm peak intensity of minimum Dst* ≤ −464 nT at 16 GMT, comparable to the most and the second most extreme magnetic storms within the standard Dst index since 1957. This storm triggered significant low-latitude aurorae in the East Asian sector and their equatorward boundary has been reconstructed as 38.°5 in invariant latitude. This result agrees with British magnetograms, which indicate an auroral oval moving above Abinger at 53.°0 in magnetic latitude. The storm amplitude was even more enhanced in equatorial stations and consequently casts caveats on their usage for measurements of the storm intensity in Dst estimates.

Fluctuations in thermospheric neutral density affect the operational stability and lifetime of low Earth orbit (LEO) satellites. Solar activity and geomagnetic disturbances induce substantial density variations in the thermosphere, thereby impacting critical satellite operations such as orbit control, attitude maneuvers, and collision avoidance. However, existing empirical models fail to accurately capture these localized thermospheric density oscillations. To date, effective methods for the high‐precision prediction of LEO satellite orbital decay under varying geomagnetic conditions remain underdeveloped. This study proposes a machine learning‐enhanced method for predicting orbital decay at specified altitudes within the LEO region by making use of Gravity Recovery and Climate Experiment Level‐1B observations and integrating along‐track high‐precision thermospheric density, aerodynamic coefficients, and satellite mass parameters. During the 9–11 May 2024 storm event, along‐track thermospheric density surged, resulting in a 48‐hr semi‐major‐axis decay of approximately 168 m before stabilizing at around 83 m thereafter. For the 24 August 2005 interplanetary coronal mass ejection (ICME) case, the cumulative decay (−45.4 m) showed close alignment with the observed orbital data (−40.4 m). When independently tested across 113 ICME events, the random forest model accounted for 85% of the variance in orbital decay, achieving a test R 2 of 0.749 during all geomagnetically periods in 2005. The results demonstrate that our proposed approach delivers significantly improved prediction accuracy of satellite orbital decay across varying geomagnetic conditions compared with empirical models. This work provides novel insights into thermospheric disturbance impacts on satellite orbits and offers essential theoretical support for LEO mission planning and orbital management.

While flagship missions such as CHAMP and GOCE have shown us with accelerometer measurements that the thermospheric density in Low Earth Orbit (LEO) can increase by more than 200% during enhanced geomagnetic activity, current empirical models, such as those of the MSISE and Jacchia families, as well as the Drag Temperature Model, fail to reproduce this behavior, limiting the ability to perform orbit prediction and space situational awareness. Several methods have been employed to address this dilemma. One is the High-Accuracy Satellite Drag Model (HASDM), which uses its Dynamic Calibration Atmosphere to employ differential correction across 75 spherical calibration satellites to generate correction parameters to the density that are related to 10.7 cm solar radio flux and ap (Storz et al. 2005). Doornbos et al. 2008 has implemented a method that estimates height-dependent scale factors to the densities from empirical models with respect to densities directly derived from two- line element sets (TLEs). HASDM’s reliance on Space Surveillance Network observations limit its accessibility and detail, and Doornbos’ methods are limited by the fact that TLEs are mean elements; densities derived from them are subject to errors due to smoothing over an entire orbit. In addition, the method of deriving densities from TLEs was initially done only to provide inputs to the SGP4 orbital propagator, which was initially developed without consideration of solar radiation pressure on the trajectory of modeled spacecraft. We present a method to generate new model densities during geomagnetic storms by using an in-house orbital propagator, the Spacecraft Orbital Characterization Kit (SpOCK). This method estimates and applies scale factors to F10.7 and a p to minimize orbit propagation errors with TLEs. The method is tested on a variety of satellites, including CHAMP, GOCE, and the CubeSats of the QB50 and FLOCK constellations. This method proposes to grant insight into storm-time thermospheric density enhancement by modeling the effects of storms on the drag of numerous LEO spacecraft, increasing our understanding of thermospheric dynamics and granting us improved tools for space traffic management and thermospheric research.

High-resolution thermospheric mass density (TMD) measurements from Low Earth Orbit (LEO) Satellites are valuable to accurately estimate the short-term atmosphere abrupt disturbances, triggered by magnetospheric forcing. A good characterization of TMD variation ahead of the arrival geomagnetic storms can benefit LEO operations and crucial for both orbit propagation and collision avoidance. In this contribution, we will reveal the most probable feature of TMD variation during the initial stage of solar cycle 25, at the same time, we proved Wygant function as a better geomagnetic events indicator.In this study, GRACE-FO 10s accelerometer-derived TMD measurements were employed and normalized at altitude of 505km by the NRLMSISE-00 (Naval Research Laboratory Mass Spectrometer and Incoherent Scatter Radar Exosphere 2000) empirical atmosphere model to investigate the status of solar cycle 25 between September 1 and December 31, 2020. With the high-inclination orbit global coverage, three magnetic latitude regions were separated and divided into day and nighttime using magnetic local times (MLT). 4-month enhancing disturbances observations suggest solar activities will shift from its relatively quiet condition to a much more active behavior, which reveal unexpected dependencies on the temporal and spatial characteries. Our detailed analysis shows that (1) TMD spreads from high latitudes to low latitudes and as same as time lag, (2) TMD enhancement in the Southern hemisphere is more intense than in the Northern one, reaching peak value around 15:00 MLT; geomagnetic activities cause TMD to increase up to 0.86×10-13 kg/m3 at night side, 3.4×10-13 kg/m3 at day side, and (3) the TMD enhancement was symmetric in both N- and S- hemispheres before the equinox. In general, thermospheric mass density analysis reveals the significant impact of solar and geomagnetic activities, providing the most relevant and probable characteristic of the TMD disturbances driven by solar wind.Additionally, we try to use different geomagnetic indices for a complete description of geomagnetic storms and their phases. The S10.7 index is used as a proxy for solar irradiation. These indicators show high correlation with the TMD variation during recurrent geomagnetic activities. What’s more, the cross-correlation analysis reflects a high correlation of to the Wygant function EWAV found both at three latitude bins.Even thought our study is considered a minor to moderate geomagnetic storm of the upcoming solar cycle 25 maximum, the high-speed stream injection into the thermosphere still caused thermosphere expansion that significantly enhanced the neutral density in the LEO environment. Therefore, all these findings provide a possibility to improve our understanding of LEO orbital drag.

Here we provide a selection of extreme geomagnetic storms of the last century based on NOAA classification which lead to the energetic particle precipitation (EPP). EPP of such geomagnetic storms can cause power outages, communication failures, and navigation problems as well as impact on the environment and the ozone level. Studies of historical extreme geomagnetic storms together with EPP for large space weather events in the space era can help to reconstruct the parameters of extreme events of past centuries.

The increase in the number of objects in Low‐Earth Orbit has heightened the demand for high‐accuracy orbital prediction models driven by dependable measurements of thermospheric mass density (TMD). Given the added cost and complexity burden of equipping satellites with high precision accelerometers, recent attention has focused on alternative techniques for observing TMD such as “GNSS accelerometry,” which involves harnessing spacecraft as instruments themselves to quantify thermospheric density vis‐à‐vis orbital decay. This work demonstrates how the Energy Dissipation Rate (EDR) technique utilizes the change in spacecraft orbital energy to recover density measurements at cadences ranging from a single orbital period down to as small as a quarter of such periods. After presenting a framework for applying the EDR method to the elliptical orbit of the Communications/Navigation Outage Forecasting System (C/NOFS) satellite, “effective” TMD measurements integrated over a continuous “orbit arc” are recovered for C/NOFS during January 2011. The merits of the EDR method, especially in its heightened sensitivity to solar/geomagnetic activity, are underscored by investigating a minor geomagnetic storm on 7 January 2011 and contrasting the results with those obtained from processing Two‐Line Element sets (TLEs) or the output from NRLMSISE‐00 and HASDM. Furthermore, this study introduces the novel application of fractional‐orbit average EDR integration tailored for satellites with eccentric orbits, demonstrating its efficacy in offering nuanced insights into thermospheric conditions. The results demonstrate the ability of physics‐based techniques and readily accessible data sets to estimate thermospheric density and provide insight into aeronomy and space weather science.