3

Research, Observation, and Modeling Needs: Magnetosphere, Ionosphere, Thermosphere, and Mesosphere

Four of the workshop’s panels were focused on research, observation, and modeling needs for the magnetosphere and the ionosphere–thermosphere–mesosphere (ITM). These panels were the Observational and Modeling Needs for the Magnetosphere Panel, the Research Needs for the Ionosphere-Thermosphere Panel, the Observation and Modeling Needs for the Ionosphere-Thermosphere Panel, and the Research Needs for Cross-Scale and Cross-Regional Coupling Panel. As previously noted, these presentations specifically addressed three different parts of the statement of task:

• Examine trends in available and anticipated observations, including the use of constellations of small satellites, hosted payloads, ground-based systems, international collaborations and data buys,

• that are likely to drive future space weather architectures; review existing and developing technologies for both research and observations.
• Consider the adequacy and uses of existing relevant programs across the agencies, including NASA’s Living With a Star (LWS) program and its Space Weather Science Application initiative, the National Science Foundation’s (NSF’s) Geospace research programs, and NOAA’s Research to Operations (R2O) and Operations to Research (O2R) programs for reaching the goals described above.
• Consider how to incorporate data from NASA missions that are “one-off” or otherwise nonoperational into operational environments, and assess the value and need for real-time data (for example, by providing “beacons” on NASA research missions) to improve forecasting models.

The presentations are summarized in the following sections. Those summaries take different approaches depending on the panel. In some parts, the summary described what each panelist said in turn. In other cases, however, for panels in which there was significant overlap among the panelists’ key messages and the panel’s output became a group effort, the session summary is more of a synthesis and has fewer comments attributed to individual speakers.

THE MAGNETOSPHERE

The Magnetosphere Panel was moderated by committee member Terry Onsager. Its panelists were Christine Gabrielse of the Aerospace Corporation, Larry Kepko of NASA’s Goddard Space Flight Center, Matina Gkioulidou of the Applied Physics Laboratory of Johns Hopkins University, and Vania Jordanova of Los Alamos National Laboratory. The panelists were asked to address two key questions:

• What do we need to understand to enable predictive capability of the magnetospheric state and irregularities?
• What are the research needs required to make progress on that understanding?

Several of the panel members expressed that the magnetosphere needs to be measured and modeled as a system of systems with an emphasis on the mesoscale structures and processes that control its dynamics. Kepko noted that mesoscales are the messengers and connectors of energy and mass across regions (e.g., tail to inner magnetosphere) and across scales (kinetic to mesoscale to global). Mesoscales in the magnetosphere can be defined as roughly from 1,000 km to a few Earth radii, and about hundreds of meters to 1,000 km in the ionosphere. Measuring and modeling this cross-regional and cross-scale coupling is an important missing element for space weather (highlighted also in Session 2 discussion of cross-scale and cross-regional processes). It was pointed out that such an observational program requires a system-of-systems approach (see also Drew Turner talk in New Research Needs). In addition, data assimilation tools need to be combined with models that bridge the micro- and macroscale processes to address key problems, such as the rapid particle injection and acceleration during substorms, plasma wave generation and feedback on the particle populations, and magnetosphere–ionosphere coupling though particle precipitation.

Kepko emphasized, supported by several other panelists, that comprehensive measurements of Earth’s magnetosphere as a system of systems will require observational capabilities that are coordinated by design, not as an afterthought (Figure 3-1). Current data sets are sparse and do not adequately capture the mesoscales. Measurements have been carried out at small/kinetic scales and at global scales, but not at mesoscales. Resolving mesoscale processes will require coordinated satellite constellations combined with imaging components. Such an ambitious space program will need coordination with ground-based

observations as well as with international, interagency, and industrial partners (see also the New Architectures session in Chapter 6).

The Heliophysics System Observatory (HSO) is a valuable resource that provides coordinated observations in pursuit of system science. However, the current coordination is ad hoc and often accidental, and it is inadequate for resolving the mesoscale. Earth’s magnetosphere has an aging, inadequate fleet of satellites, and nothing is currently in the queue either for science or for space weather monitoring. The inadequacy of the available measurements is clear when considering the structures and dynamics indicated by global numerical models (Figure 3-1). The properties of the dynamic structures seen in the models cannot be measured with currently available three or four widely spaced satellites, and it is therefore not possible to adequately constrain or validate the models with existing data.

The various systems that make up the magnetosphere (bow shock, magnetosheath, magnetopause, plasmasphere, ring current, radiation belts, magnetotail, etc.) cannot be studied in isolation. The properties of these systems are “emergent”; that is, they emerge through interactions with the other systems. These interactions typically occur at the mesoscales. For example, when reconnection occurs on the dayside magnetopause, that information is communicated throughout the magnetosphere by flux transfer events, which are mesoscale structures. And when the magnetotail subsequently goes unstable and releases a substorm, that configuration change is transferred to the inner magnetosphere and ionosphere through mesoscale structures.

To achieve the required observing infrastructure, various panel members suggested taking a new coordinated international approach to provide multiple constellations of spacecraft in key regions to resolve the mesoscale structures. This would include mesoscale dynamics (e.g., flow burst and magnetopause flux transfer events), inner magnetosphere/magnetotail coupling, the cold plasma distribution, auroral coupling, ion outflow, and the coupling of kinetic processes to mesoscale structures. Geospace state variables, such as the upstream solar wind, auroral configuration, cross-polar-cap potential, the state of the radiation belts, polar cap open flux, solar irradiance, cold plasma density, auroral field–aligned currents, ionospheric convection maps, geomagnetic indices, and ionospheric total electron content need to be measured. The space-based measurements must be coordinated with ground-based measurements, and the full set of observations must in turn feed into and validate advanced numerical models and data assimilation. All of these activities should focus on studying the geospace holistically as a system and at the scale sizes we now know are driving the system dynamics.

One key region where mesoscale processes are particularly important is the transition region in the magnetotail between the stretched magnetic field in the plasma sheet and the dipolar magnetic field closer to Earth. As plasma flows rapidly earthward from the magnetotail reconnection site, it encounters the strong, rigid magnetic field near Earth, which causes considerable dynamics, energy conversion, and transport that occur over mesoscale (roughly 1 to 4 Earth radii) distances.

Previous satellites have sampled one or a few simultaneous locations within this region during dynamical time periods, but this has not been sufficient to resolve the mesoscale structures. Some members of the panel suggested it would be useful to have a grid of satellites with magnetometers, particle detectors, and potentially other instruments combined with ground-based and space-based imagers that can measure the two-dimensional structure and dynamics. The measurements available from such a collection would be essential in developing, driving, and validating current (research and operational) models.

Detailed knowledge is also needed of the near-Earth plasma ion composition. For example, the presence of N+ ions plays a key role in ionospheric outflow (Ilie 2021). However, information about the ion composition mostly comes from decades-old measurements. This observation gap inhibits the ability to accurately predict and characterize hazardous space weather events in the near-Earth space environment. As detailed in recent Heliophysics 2050 white papers, significant technological advances now enable more sophisticated mass spectrometers in smaller design packages than previously possible (Fernandes et al. 2021).

Another gap in current observations is the cold plasma distribution and variability. Both fundamental research and space weather predictions require the inclusion of the plasmaspheric dynamics, as the cold plasma affects many magnetospheric processes, such as the properties of the plasma waves controlling the behavior of the radiation belt electrons and the ring current ions. The cold particle populations with energies less than 100 eV are difficult to measure and therefore are the least studied, but they often dominate the plasma density. Recent Heliophysics 2050 white papers detailed the need to understand cross-scale dynamics of the plasmasphere (Goldstein et al. 2021) and the innovations needed to measure the cold plasma populations in space (Delzanno et al. 2021).

SmallSats were described as a promising approach to allow a larger number of satellites at an affordable cost. GTOSat (Blum et al. 2020), a NASA Goddard 6U CubeSat led by principal investigator Dr. Lauren Blum, is one such pathfinder. Originally planned for launch in summer 2022¹ to a low inclination geosynchronous transfer orbit, it will carry a magnetometer and a relativistic electron and proton detector that weighs only 1 kg. As this and other missions demonstrate lower-cost capabilities, the opportunities to obtain large numbers of simultaneous measurements within the magnetosphere will increase in the near future.

___________________

¹ At the time of publication, GTOSat had not been launched.

Two-dimensional images of the magnetosphere were described as an important complement to in situ space plasma measurements. Mesoscales, which are roughly one to a few Earth radii in size, are difficult to constrain with even three or four satellites. Images can provide the essential spatial context for multipoint, in situ spacecraft observations. Constellations of satellite measurements in the magnetosphere and ionosphere need to be coordinated with high-resolution two-dimensional images that resolve the mesoscales. For example, energetic neutral atom imaging, auroral imaging, and plasmasphere imaging can all monitor the dynamics of the geospace system in the key transition regions.

Recent studies using energetic neutral atom imagers on the TWINS satellites and all-sky imagers from the THEMIS network have documented flow channels coming Earthward from the magnetotail before a sudden storm commencement, indicating the importance of mesoscale injections to inner magnetosphere dynamics (Adewuyi et al. 2021). Although the TWINS mission was recently decommissioned and the THEMIS all-sky imagers may only operate for a few more years, Canada is implementing an all-sky camera array. In this situation there are important questions for heliophysics, including what the U.S. infrastructure contributions will be to two-dimensional imaging of the magnetosphere and how U.S. researchers can participate in internationally coordinated efforts. Ground-based radars were also noted as an important asset for imaging mesoscale dynamics in the magnetosphere–ionosphere system.

Numerical modeling was also described as an essential component of resolving the mesoscale structures and connecting the microscale, mesoscale, and global processes. Improving space weather predictions will require models that can connect the various regional models and incorporate the mesoscale processes and the cross-scale dynamics into the global model. For example, existing magnetohydrodynamic (MHD) models have demonstrated the importance of the mesoscale process in particle transport within the magnetosphere. However, the differentiation between adiabatic and non-adiabatic energization processes and their larger-scale consequences can only be done by coupling the kinetic scale physics with other models within the magnetospheric domain.

As geospace is a system of systems, models need to connect those systems in the critical mesoscale regimes. Efforts to develop such models need to be coordinated with the observations described above and will likely require new computational schemes. In addition to bridging the macro and micro scales, the modeling effort needs to be combined with data assimilation (see Chapter 5). Although scientists strive to use the highest resolution possible in numerical models, in the case of the complex global magnetosphere, mesoscale resolution may be the most practical near-term goal. Such models must include the full range of processes from the solar wind–magnetosphere interface to the ionosphere and various layers of the neutral atmosphere. The Center for Geospace Storms is one effort to model this multiscale atmosphere-geospace environment (see Figure 3-2).

As was brought up both in Session 2 and in the Session 3 Solar Wind panel, accurate measurements of the solar wind that hits the magnetosphere are essential for developing and running accurate space weather models. Solar wind measurements are used both as an outer boundary condition for magnetosphere models as well as the input to proxy models for internal boundary conditions, such as geostationary orbit state parameters for which observations are not always available. With appropriate space weather measurements, the models can provide a forecast with up to an hour lead time. Accurate forecasts of the magnetosphere–ionosphere system require equally accurate measurements of the solar wind that actually impacts the magnetosphere. It was pointed out that L1 measurements may not be sufficient due to a variety of processes creating variability between L1 and Earth’s bow shock (see also the discussions of the Solar Wind panel).

Artificial intelligence (AI) and neural networks were also highlighted as having important applications in magnetospheric modeling. A recent model by Claudepierre and O’Brien (2020) was shown to reproduce well the radiation belt electron fluxes throughout the inner magnetosphere using input data from low Earth orbit electron flux measurements and the K_p geomagnetic index.

The panel also highlighted a number of valuable concepts that have been documented in the recent Heliophysics 2050 white papers² and that complement the issues raised by the panel. More powerful computing technologies will lead to improved physics-based modeling capabilities in areas such as multi-fluid, hybrid, and Vlasov codes as well as implicit and kinetic solvers. These advances should provide opportunities to significantly improve space weather prediction and our understanding of the geospace system of systems. The increasing amounts of data will lead to novel combinations of physics-based models with machine learning, deep learning, and “physics emulator” approaches. AI has already been successfully applied for automatic event identification, feature detection and tracking, and uncertainty quantification. These advances would benefit from collaborations with other disciplines, and therefore space weather data sets should be promoted in computer science, engineering, and other related fields.

The panel also described specific examples where advances in magnetospheric observations and modeling would directly benefit the space weather user community. One example was the need to better constrain the radiation environment within the magnetosphere to enable more efficient engineering options for satellite design. The radiation environment is an important source of solar cell voltage degradation, and current statistical models (AE9/AP9) are insufficient to predict fluences on time scales of days, months, and even a couple of years.

___________________

² See NASA Science Mission Directorate Heliophysics Division, “Heliophysics 2050 Workshop Program,” Lunar and Planetary Institute and Universities Space Research Association, http://www.hou.usra.edu/meetings/helio2050/program.

It was shown that a model using data from the two van Allen probes was able to more accurately predict the observed voltage degradation on satellites than the statistical model.3 With additional satellites in the geostationary transfer orbit, the magnetic local time dependence of the radiation environment could also be modeled.

Another important user need for space weather information involves the atmospheric heating by energy input from the magnetosphere to the ionosphere and atmosphere (see Figure 3-3). The subsequent upwelling of the atmosphere increases satellite drag and impacts spacecraft lifetime as well as efforts to manage space traffic and collision avoidance. Existing models that are driven by solar wind parameters can capture much of the large-scale global effects, but the mesoscale phenomena are poorly characterized by current models and can be highly dynamic.

Finally, the panelists identified two concerns: The first was the need to further develop the workforce needed to advance space weather capabilities. Although space weather is a discipline with high societal importance, there seems to be a lack of career opportunities that can attract people with the right expertise

to enter and to stay in the field (also see the Diversity in Workforce session in Chapter 1). The second concern was the current lack of information sharing between space weather users and the scientific community. Many users are unwilling to share proprietary data, which can make it difficult for researchers to understand what information is needed to address the space weather impacts. Furthermore, it is expected that as the technological applications advance, the need for space weather information will increase, but many companies may not yet realize how they are or will be impacted by space weather. Improving communication between researchers and users of space weather information could accelerate the progress in understanding space weather and its utility.

3 Gabrielse et al., 2022, “Radiation Belt Daily Average Electron Flux Model (RB-Daily-E) from the Seven-Year Van Allen Probes Mission and Its Application to Interpret GPS On-orbit Solar Array Degradation,” submitted to the workshop.

IONOSPHERE AND THERMOSPHERE

Research Needs

The Ionosphere and Thermosphere Research Needs Panel addressed the state of the ionosphere–thermosphere and its irregularities. The five panelists were Seebany Datta-Barua of the Illinois Institute of Technology, Charles Carrano of Boston College, Jonathan Snively of Embry-Riddle Aeronautical University, Sean Bruinsma of the Space Geodesy Office of the French space agency CNES, and Greg Ginet of the Massachusetts Institute of Technology. The panelists were asked to address two key questions:

• What do we need to understand to enable predictive capability of the thermospheric and ionospheric state and irregularities?
• What are the research needs required to make progress on that understanding?

The panel’s answer to the first question was that predictive capability with quantified uncertainties requires understanding of how the state of the ionosphere evolves (due to driving from above and from below) to create the irregularities, how the irregularities affect signal propagation, and what are the impacts of the changed signal propagation on the system performance. To make progress on that understanding (second question above), the panel identified global instrument network data sets that are routinely calibrated and readily available; coordination strategies across communities that share data (such as the Global Navigation Satellite System [GNSS] for Earth science and heliophysics); models with flexible interfaces and products that can be used with other models; and modeling strategies and community-adopted assessments validation metrics.

Datta-Barua began the session with a general description of the complexities of the ionospheric state, with dependencies on latitude (high-latitude, mid-latitude, and low-latitude bands) as well as scale (small ~100 m to to mesoscale ~1,000 km). The inputs to this system come from the Sun and the magnetosphere as well as from neutral dynamics, acoustic–gravity waves, traveling ionospheric disturbances (TIDs), and turbulence, all of them interacting with the geomagnetic field.

Turning to the operational applications, Datta-Barua focused on safety-critical Global Positioning System (GPS) services, such as aircraft landing systems. Aviation users require both positioning accuracy and the assurance that uncertainties due to ionospheric ranging errors are within acceptable limits, or a prompt alert to switch to other navigation methods to maintain their safety.

Better predictive capabilities would require estimates with quantified uncertainties of safety-critical parameters, Datta-Barua said. Furthermore, new methods—either deterministic or stochastic—are needed for forecasting the onset of instabilities. The forecast updates are needed on time scales of minutes to hours (as opposed to the currently available daily updates) and with spatial scales extending from global to tens of meters.

Making progress on understanding the underlying physics, Datta-Barua said, will require continuous support for geodetic GNSS (Global Navigation Satellite System) networks and upgrading these with

multi-frequency sensors. Another means to sample the irregularity structures would be via space-borne measurements. However, the data records are only useful for the statistical approaches if they cover long time periods and are available and accessible to the community.

Carrano spoke about ionospheric radio wave scintillation (i.e., rapid changes in the wave propagation caused by ionospheric electron density variations that can affect the performance of various technological systems, including satellite communication and GNSS). Scintillations can be measured using ground-to-satellite and satellite-to-satellite transmissions, but gaining real-time specification and forecasts will require a fusion of ground- and space-based data sources, including the use of existing sensors and receiver networks as well as an improved theoretical understanding of the ionospheric state and the driving mechanisms that cause irregularities to develop. This in turn will require support for sensors and networks, modeling efforts, and system impact models.

Ginet addressed the effects of ionospheric disturbances on the propagation of high-frequency (HF) radio waves (i.e., 3-30 MHz). Because HF waves can reflect from the bottom of the ionosphere, they can be used for long-distance communications (e.g., AM radios) and over-the-horizon radar (OTHR) applications. However, improving the performance of these systems, which are limited by the knowledge of the ionospheric state, will require a better ability to measure and predict the state of the ionosphere. In particular, global networks measuring the bottom-side electron density profiles over the entire planet would bring major advancements.

There are several possible ways to realize observational systems that would address these issues, each of which has its own issues: A GNSS ground-based and remote occultation total electron content (TEC) network may not have the spatial and temporal resolution needed in propagation applications and data assimilation models. A network of standard and low-power HF sounders across the globe might be feasible, but it is not clear whether oblique networks would work. Oceanographic HF radars operate in the right frequency bands, but to be routinely usable would require standardization of waveform parameters and a network of ionospheric monitoring links to be established using passive receivers. In summary, Ginet said, the HF community needs better bottom-side electron density profiles and parameters over the entire planet, maps of sporadic E-layers, and validated models using operationally relevant metrics (e.g., path loss, group delay, angle-of-arrival). The needs for improved predictive capability include systematic validation of ionospheric models and establishing data integration centers where this validation can occur.

Snively discussed the effects of lower thermospheric dynamics on the bottom of the ionosphere. Atmospheric inputs include thermospheric density and compositional fluctuations as well as strong and variable shears over a wide range of scales. The ionospheric E- and F-regions can be modulated by lower thermosphere neutral waves as well as by acoustic gravity waves, which may include neutral fluctuations. Snively also pointed out that the TEC and HF measurements are valuable inputs not only for ionospheric but also for thermospheric dynamics. Multi-scale, multi-system, physics-based models of thermosphere–ionosphere dynamics in both small scales and across the scales and systems as well as continuous measurements of the most relevant parameters are required for progress in physical understanding and in predictions.

Snively summarized that key observations include high-resolution neutral composition and winds, particularly in the lower thermosphere 100-200 km, and quantification of dynamic processes in the thermosphere that modulate the bottom-side ionosphere. Physics-based modeling of ion-neutral coupling should continue and connect to impacts on signal propagation. Models at different scales should mesh well, be data-constrained, and leverage high-resolution data sets, which will require big-data handling strategies for heterogeneous data sets where the coverage is distributed over space and time.

Bruinsma addressed the large-scale state of the thermosphere–ionosphere system. Thermospheric density is a key parameter for satellite orbit analysis as it is directly proportional to the atmospheric drag force causing spacecraft orbital decay. Currently, the accuracy of orbit calculations is limited by the performance

of the thermospheric specification models, which rely on simple empirical models of the atmospheric density. The models are limited by the use of proxies (such as the F10.7 flux) instead of using the actual extreme ultraviolet flux (EUV; see also the discussion of continuous measurements of the solar irradiance in the EUV), by the lack of an accurate description of the neutral composition, and by the inconsistent quality and sparse distribution of data. Keeping track of the increasing number of low Earth orbiting satellites will require precision orbit determination and conjunction analysis, satellite lifetime estimation, and mission analysis.

Improving predictive capabilities in the lower thermosphere–ionosphere, in particular for atmospheric drag calculations, will require improved models in the transition region at altitudes between 100 and 200 km that can capture the net energy input from the solar EUV, the solar wind and interplanetary magnetic field, and the magnetosphere, ionosphere, and thermosphere at sufficiently high temporal cadence.

Bruinsma summarized that observations of composition and density with calibration and processing standards, of sustained measurements in the EUV (of the He II) line, and of solar observations made from the L1 and L5 Lagrange points will be needed. Data assimilation testing schemes, ensemble modeling, and systematic assessment of models (e.g., under the Community Coordinated Modeling Center [CCMC]) will continue to be needed. Lastly, Bruinsma pointed out how important investments in people, workforce expertise and education will be.

Modeling Needs

Another panel addressed observation and modeling needs for the ionosphere–thermosphere system. Its six panelists were Naomi Maruyama of the University of Colorado, Matt Zettergren of Embry-Riddle Aeronautical University, Katrina Bossert of Arizona State University, Bill Lotko of the University Corporation for Aeronautical Research, Larisa Goncharenko of the Massachusetts Institute of Technology, and Hanli Liu of the University Corporation for Aeronautical Research. The panelists were asked to address these key questions:

• What are the advances in modeling and observations needed to improve the understanding of the Sun–Earth system that generates space weather?
• What are the biggest challenges in ionosphere–thermosphere–mesosphere science in the coming decades?
• What observations are needed to test our understanding and our ability to nowcast/forecast the system?
• To what extent is space weather in the ionosphere–thermosphere driven from below?

An overarching theme from this panel was that the whole system—the ionosphere coupled from above and from below—needs to be understood for predicting space weather. The ionosphere between the neutral atmosphere and the magnetosphere is forced from below and from above, and the relative contributions of this forcing depend on location and geomagnetic activity as well as the past history of forcing (“preconditioning”). At high latitudes and during large disturbances, forcing from above dominates, while at middle and low latitudes forcing from below is dominant most of the time. It is as yet uncertain how seasonal and solar cycle variations or climate change shift the relative importance of forcing from below and above or how the upper atmosphere affects the geo-effectiveness of external forcing.

A number of panelists made the point that although multi-scale and multi-step physical processes are important in understanding the ITM system and space weather, they are not well observed or well modeled. ITM variability contains significant energy that isn’t captured by the mean state of the system: the large-scale background state can drive small-scale irregularities, and large tides can trigger smaller waves

(especially in the polar regions). On the other hand, small-scale turbulence can lead to large-scale changes. Multi-satellite observations are required to resolve the evolution of different waves and their impacts in the ionosphere–thermosphere system.

The single most important factor regulating the magnetosphere–ionosphere interaction is the ionospheric conductivity tying the neutral atmosphere to the ionospheric electrodynamics. The conductivity magnitude and spatial distribution need to be measured at mesoscales (10-100 km) and as a function of time in order to improve the predictive, global geospace models. While it is straightforward to get the conductivity for a given thermospheric state, estimating the precipitation-induced conductivity is more complicated. Since measuring the global conductivity distribution is challenging, the panel suggested combining space-based and ground-based multispectral imaging covering both the global ionosphere and mesoscales down to 10 km resolution.

Ionospheric outflows are a key process through which the ionosphere influences the magnetosphere: They affect magnetic reconnection rates, global dynamics, tail and substorm dynamics, and the ring current and the radiation belts. The observing solutions proposed by the panel included multipoint measurements along flux tubes, global ultraviolet (UV) imaging to determine the ion outflow source and fluxes, and using the upcoming Geospace Dynamics Constellation (GDC) measurements, which, unfortunately, are not optimized for that particular science and will provide only some information. The discussion following the panel presentations touched on imaging solutions, recognizing that some species, such as H⁺, cannot be imaged optically, while the imaging of others may present significant challenges.

Gravity waves and tides are another means by which the atmosphere affects the variability of the ionosphere and the plasmasphere. Atmospheric waves are an important source of energy flux from the lower atmosphere and encompass structures in different spatial and temporal scales, such as TIDs, wave dissipation, the Madden-Julian oscillation, and the quasi-biennial oscillation. Understanding how gravity waves affect the circulation and compositional structure of the thermosphere, in terms of both day-to-day variability and smaller-scale waves, requires altitude profile measurements of wind, temperature, and density at a resolution capable of following the evolution of waves with altitude and latitude/longitude. These processes are important for the models, as the thermosphere is highly sensitive to the meridional system, and small changes in thermospheric composition or temperature changes can lead to major changes higher up. Furthermore, climatological measurements of thermospheric parameters (e.g., tides and winds) are required to understand the multi-scale interactions in the ionosphere–thermosphere system. Finally, global-scale thermosphere measurements and models are important in understanding the interaction of forces from below and above and their role in modulating the ITM system during quiet and storm times. Open questions include whether auroral-generated gravity waves produce a different impact than other gravity waves, whether the system predictability changes during storms, and how the lower atmosphere processes affect the fidelity of whole-atmosphere models.

Electric fields and currents connect Earth’s atmosphere with the magnetosphere through the ionosphere (i.e., R1 and R2 field-aligned currents) at a range of temporal and spatial scales. The energy injected into the atmosphere as Joule heating is important (as was demonstrated, for instance, by the failure of nearly 40 SpaceX satellites after a geomagnetic storm), but we lack both understanding and models of these nonlinear connections over the full range of temporal and spatial scales. However, we do know that preconditioning of the upper atmosphere affects the geoeffectiveness of external forcing; for example, neutral winds increase the variability of the penetration of electric fields. An observational solution would be to monitor the electric currents, as with the approach pioneered by the Active Magnetosphere and Planetary Electrodynamics Response Experiment (AMPERE) and the European Space Agency’s Swarm satellite constellation; more spacecraft would improve the spatial coverage and temporal resolution.

A common theme from the panel was that forcing from below is an important aspect of space weather from the plasmasphere out to the magnetosphere. Quantifying this forcing requires high-resolution,

four-dimensional observations of the thermospheric parameters (e.g., temperature, O, N₂, O₂ density, vector winds); a further need is to understand the biases between past and current observations and between observations using different techniques. These measurements should also be coordinated with Earth sciences to uncover how climate change may affect the ionosphere–thermosphere coupling (e.g., through increasing temperatures producing convective instabilities [e.g., cyclones, thunderstorms, etc.]), which launch gravity waves, which in turn lead to changes in the upper atmosphere tides and gravity waves.

A new modeling approach to multi-scale physics-based models is needed that can describe that coupling and provide feedback on cross-scale resolution (Figure 3-4). The development of such models will require sustainable support for the research, computational infrastructures, a dense observation network for model validation, and collaboration between the science community and private companies.

Progress will require sustained investment in local and general-purpose open-source codes that address issues in the current licensing practices. Of specific techniques, it was noted that adaptive mesh refinement (AMR) schemes may be underused. Regarding system predictability, Liu commented that it does not matter “where you put your butterfly”; global data assimilation will be required to manage the error growth.

CROSS-SCALE AND CROSS-REGION COUPLING

Committee member Endawoke Yizengaw moderated a panel on cross-scale and cross-region coupling. The panelists were Josh Semeter of Boston University; Astrid Maute of the University Corporation for Atmospheric Research; Joe Huba of Syntek Technologies; Seth Claudepierre of the University of California, Los Angeles; and Jonathan Rae of Northumbria University in Newcastle upon Tyne, England. The panelists were asked to address the following questions:

• What do we need to understand about cross-scale and cross-region coupling to enable predictive capability of the state of the magnetosphere/ionosphere–thermosphere–mesosphere?
• What are the research needs required to make progress on that understanding?

Semeter began by proposing definitions for cross-scale and cross-region coupling. Cross-scale coupling occurs when dynamics at one spatio-temporal scale define boundary conditions for another scale governed by different physics (e.g., processes coupling MHD and kinetic scales). Cross-region coupling takes place between geophysical domains demarcated by a change in physical description (e.g., between solar wind and magnetosphere or ring current and plasmasphere). Examples of cross-scale coupling from an energy perspective (see Table 3-1) include tail reconnection at substorm onset driven by ion scale physics

and leading to global magnetotail reconfiguration and energy release; or Alfvén waves transferring energy between the magnetosphere and the ITM system. Many such cross-scale or cross-regional coupling processes are not yet adequately included in the predictive models.

Recognizing the challenge of sampling the geospace system with sufficient resolution, Semeter outlined a scheme to use the ionospheric “projection screen” to deduce information about outer space processes. This would be accomplished by combining ground-based network data to search for cross-scale and cross-region dynamics, which would involve data fusion and inversion models as well as new information theory and AI/machine learning methods and collaboration with computer scientists, engineers, statisticians, and data scientists.

Maute focused on mesoscale electric fields and precipitation (10s to 100s of kilometers) that are critical for accurate estimations of the Joule heating and precipitation; lacking the smaller scales can lead to underestimations by up to 50 percent during active times. Furthermore, the F-region neutral winds respond within minutes to auroral forcing down to 100-km scales and can reduce the local Joule heating. As future needs, Maute identified regionally self-consistent observations of ion drift, particle precipitation, field-aligned currents (FACs), and E- and F-region neutral winds, which are tied to a self-consistent picture by empirical models and data assimilation. On a larger scale, she continued, ionospheric electrodynamics and the geomagnetic fields must be coupled at appropriate scales across physical domains.

Huba discussed cross-scale coupling and irregularities in the ITM using an example of a hierarchy of irregularities in the ionosphere (Figure 3-5) ranging spatially from tens of centimeters to hundreds of kilometers and temporally from milliseconds to hours. This vast range of scales requires the use of different physical equations, from kinetic to fluid theory.

Huba explained that while the majority of current electrodynamics models assume equipotential magnetic field lines, recent work has extended the description to three-dimensional electrodynamics with varying potential along the magnetic field, which allows new instabilities to develop in the models. The community still lacks self-consistent electrodynamic models that would combine the different processes of the low- and high-latitude regions. Such global models would be particularly important for storm times, capturing both the storm time penetration electric fields imposed on the ionosphere by the magnetosphere and the ionospheric storm-time dynamo electric fields. Making progress, he said, will require including scale sizes down to 100s of meters, sub-gridding for small-scale physics, and embedding particle-in-cell (PIC) codes in the global MHD codes.

Claudepierre identified three open questions that need to be addressed to improve the predictive capability of the magnetosphere/ITM system: the role that mesoscale injections of plasma sheet particles play in inner magnetosphere energetic particle dynamics; the significance of energetic particle precipitation (EPP) as a facilitator of magnetosphere–ionosphere coupling; and an improved understanding of global scale cold plasma evolution, energization, and motion.

The ring current formation is an example of the first issue: Is it built up through enhanced global convection, through a series of localized injections, or some combination of both? Estimation of the efficiency of these processes requires a number of assumptions that are not well constrained by either observations or models. This knowledge gap limits the current ability to predict ring current evolution and, thus, the development of geomagnetic storms.

Claudepierre made the following observations as part of a broad strategy to close knowledge gaps: Coordinated multipoint observations distributed widely over the ITM system should include measurements of the injections over a broad region of the nightside plasma sheet. These should be combined with global imaging of the nightside region and concurrent measurements of the ring current and radiation belt particles. In addition, these measurements need to be coordinated with a robust modeling program, such as MHD test-particle simulations or global MHD with embedded PIC simulations. A particular need is

to improve the accuracy of global geomagnetic field models, as the mapping from the magnetosphere to the ionosphere plays a crucial role in the coupling studies. Concerning implementation, Claudepierre said that the HSO has not been strategically planned from a systems science perspective and that multipoint measurements could be realized by adding sensors onboard operational missions such as the polar-orbiting environmental satellites (such as the Polar Operational Environmental Satellites [POES] system) or the Defense Meteorological Satellite Program (DMSP), which have traditionally provided such measurements. It will be important both to ensure the long-term continuity of such data and to establish new programs.

Rae began by noting that cross-scale and cross-region coupling in space plasmas covers a vast range of spatial and temporal scales as well as of plasma characteristics present throughout the heliosphere, from the solar corona to the ionosphere. In particular, he said, it is not possible for any single simulation to cover all of the associated energy ranges.

Solar wind energy input can only account for 50-60 percent of the magnetic perturbations in the ionosphere, with the other 40-50 percent likely coming from the magnetosphere associated with substorm activity. This second component is important, as substorms power the radiation belts and the ring current and deposit energy into the ionosphere and drive geomagnetically induced currents, which are a space weather hazard. A further problem with predicting magnetosphere–ionosphere coupling and space weather impacts, Rae said, is the lack of understanding of the radiation belts. It is important to understand how the magnetotail feeds the radiation belts, including cold plasma (i.e., on the order of eV), reconnection, substorms, wave propagation, and wave–particle interactions; warm plasma (keV) for wave–particle interactions, wave propagation, and spacecraft charging; and relativistic plasma (MeV) for single-event upsets. Multi-point measurements are needed to verify the data-driven models.

For the future, new point measurements in the key regions of geospace will be needed to both drive and test models, with continuous ground-based measurements providing key global measurements. Realistic models of plasma behavior for substorms, cold plasma, warm plasma, and relativistic plasma are needed; they would build on these new data. And it will be important to determine what can be predicted deterministically and what is best described by probabilistic models.

REFERENCES