The Origin of Solar Energetic Particle Events: Coronal Acceleration versus Shock Wave Acceleration

We review evidence that led to the view that acceleration at shock waves driven by coronal mass ejections (CMEs) is responsible for large particle events detected at 1 AU. It appears that even if the CME bow shock acceleration is a possible model for the origin of rather low energy ions, it faces difficulties on account of the production of ions far above 1 MeV: (i) although shock waves have been demonstrated to accelerate ions to energies of some MeV nucl^–1 in the interplanetary medium, their ability to achieve relativistic energies in the solar environment is unproven; (ii) SEP events producing particle enhancements at energies 100 MeV are also accompanied by flares; those accompanied only by fast CMEs have no proton signatures above 50 MeV. We emphasize detailed studies of individual high energy particle events which provide strong evidence that time-extended particle acceleration which occurs in the corona after the impulsive flare contributes to particle fluxes in space. It appears thus that the CME bow shock scenario has been overvalued and that long lasting coronal energy release processes have to be taken into account when searching for the origin of high energy SEP events.     

Solar Sources of Heliospheric Energetic Electron Events—Shocks or Flares?

Electrons with near-relativistic (E≳30 keV, NrR) and relativistic (E≳0.3 MeV) energies are often observed as discrete events in the inner heliosphere following solar transient activity. Several acceleration mechanisms have been proposed for the production of those electrons. One candidate is acceleration at MHD shocks driven by coronal mass ejections (CMEs) with speeds ≳1000 km s⁻¹. Many NrR electron events are temporally associated only with flares while others are associated with flares as well as with CMEs or with radio type II shock waves. Since CME onsets and associated flares are roughly simultaneous, distinguishing the sources of electron events is a serious challenge. On a phenomenological basis two classes of solar electron events were known several decades ago, but recent observations have presented a more complex picture. We review early and recent observational results to deduce different electron event classes and their viable acceleration mechanisms, defined broadly as shocks versus flares. The NrR and relativistic electrons are treated separately. Topics covered are: solar electron injection delays from flare impulsive phases; comparisons of electron intensities and spectra with flares, CMEs and accompanying solar energetic proton (SEP) events; multiple spacecraft observations; two-phase electron events; coronal flares; shock-associated (SA) events; electron spectral invariance; and solar electron intensity size distributions. This evidence suggests that CME-driven shocks are statistically the dominant acceleration mechanism of relativistic events, but most NrR electron events result from flares. Determining the solar origin of a given NrR or relativistic electron event remains a difficult proposition, and suggestions for future work are given.     

The Evolving Sigmoid: Evidence for Magnetic Flux Ropes in the Corona Before, During, and After CMES

It is generally accepted that the energy that drives coronal mass ejections (CMEs) is magnetic in origin. Sheared and twisted coronal fields can store free magnetic energy which ultimately is released in the CME. We explore the possibility of the specific magnetic configuration of a magnetic flux rope of field lines that twist about an axial field line. The flux rope model predicts coronal observables, including heating along forward or inverse S-shaped, or sigmoid, topological surfaces. Therefore, studying the observed evolution of such sigmoids prior to, during, and after the CME gives us crucial insight into the physics of coronal storage and release of magnetic energy. In particular, we consider (1) soft-X-ray sigmoids, both transient and persistent; (2) The formation of a current sheet and cusp-shaped post-flare loops below the CME; (3) Reappearance of sigmoids after CMEs; (4) Partially erupting filaments; (5) Magnetic cloud observations of filament material.     

CME Theory and Models

This chapter provides an overview of current efforts in the theory and modeling of CMEs. Five key areas are discussed: (1) CME initiation; (2) CME evolution and propagation; (3) the structure of interplanetary CMEs derived from flux rope modeling; (4) CME shock formation in the inner corona; and (5) particle acceleration and transport at CME driven shocks. In the section on CME initiation three contemporary models are highlighted. Two of these focus on how energy stored in the coronal magnetic field can be released violently to drive CMEs. The third model assumes that CMEs can be directly driven by currents from below the photosphere. CMEs evolve considerably as they expand from the magnetically dominated lower corona into the advectively dominated solar wind. The section on evolution and propagation presents two approaches to the problem. One is primarily analytical and focuses on the key physical processes involved. The other is primarily numerical and illustrates the complexity of possible interactions between the CME and the ambient medium. The section on flux rope fitting reviews the accuracy and reliability of various methods. The section on shock formation considers the effect of the rapid decrease in the magnetic field and plasma density with height. Finally, in the section on particle acceleration and transport, some recent developments in the theory of diffusive particle acceleration at CME shocks are discussed. These include efforts to combine self-consistently the process of particle acceleration in the vicinity of the shock with the subsequent escape and transport of particles to distant regions.     

The Pre-CME Sun

The coronal mass ejection (CME) phenomenon occurs in closed magnetic field regions on the Sun such as active regions, filament regions, transequatorial interconnection regions, and complexes involving a combination of these. This chapter describes the current knowledge on these closed field structures and how they lead to CMEs. After describing the specific magnetic structures observed in the CME source region, we compare the substructures of CMEs to what is observed before eruption. Evolution of the closed magnetic structures in response to various photospheric motions over different time scales (convection, differential rotation, meridional circulation) somehow leads to the eruption. We describe this pre-eruption evolution and attempt to link them to the observed features of CMEs. Small-scale energetic signatures in the form of electron acceleration (signified by nonthermal radio bursts at metric wavelengths) and plasma heating (observed as compact soft X-ray brightening) may be indicative of impending CMEs. We survey these pre-eruptive energy releases using observations taken before and during the eruption of several CMEs. Finally, we discuss how the observations can be converted into useful inputs to numerical models that can describe the CME initiation.     

Multi-Wavelength Observations of CMEs and Associated Phenomena

This chapter reviews how our knowledge of CMEs and CME-associated phenomena has been improved, since the launch of the SOHO mission, thanks to multi-wavelength analysis. The combination of data obtained from space-based experiments and ground based instruments allows us to follow the space-time development of an event from the bottom of the corona to large distances in the interplanetary medium. Since CMEs originate in the low solar corona, understanding the physical processes that generate them is strongly dependant on coordinated multi-wavelength observations. CMEs display a large diversity in morphology and kinematic properties, but there is presently no statistical evidence that those properties may serve to group them into different classes. When a CME takes place, the coronal magnetic field undergoes restructuring. Much of the current research is focused on understanding how the corona sustains the stresses that allow the magnetic energy to build up and how, later on, this magnetic energy is released during eruptive flares and CMEs. Multi-wavelength observations have confirmed that reconnection plays a key role during the development of CMEs. Frequently, CMEs display a rather simple shape, exhibiting a well known three-part structure (bright leading edge, dark cavity and bright knot). These types of events have led to the proposal of the ‘`standard model’' of the development of a CME, a model which predicts the formation of current sheets. A few recent coronal observations provide some evidence for such sheets. Other more complex events correspond to multiple eruptions taking place on a time scale much shorter than the cadence of coronagraph instruments. They are often associated with large-scale dimming and coronal waves. The exact nature of these waves and the physical link between these different manifestations are not yet elucidated. We also discuss what kind of shocks are produced during a flare or a CME. Several questions remain unanswered. What is the nature of the shocks in the corona (blast-wave or piston-driven?) How they are related to Moreton waves seen in Hα? How they are related to interplanetary shocks? The last section discusses the origin of energetic electrons detected in the corona and in the interplanetary medium. “Complex type III-like events,”which are detected at hectometric wavelengths, high in the corona, and are associated with CMEs, appear to originate from electrons that have been accelerated lower in the corona and not at the bow shock of CMEs. Similarly, impulsive energetic electrons observed in the interplanetary medium are not the exclusive result of electron acceleration at the bow shocks of CMEs; rather they have a coronal origin.     

Modelling of Solar Wind, CME Initiation and CME Propagation

Simulations of coronal mass ejections (CMEs) evolving in the interplanetary (IP) space from the Sun up to 1 AU are performed in the framework of ideal magnetohydrodynamics (MHD) by the means of a finite-volume, explicit solver. The aim is to quantify the effect of the background solar wind and of the CME initiation parameters, such as the initial magnetic polarity, on the evolution and on the geo-effectiveness of CMEs. First, three different solar wind models are reconstructed using the same numerical grid and the same numerical scheme. Then, different CME initiation models are considered: Magnetic foot point shearing and magnetic flux emergence. For the fast CME evolution studies, a very simple CME model is considered: A high-density and high-pressure magnetized plasma blob is superposed on a background steady state solar wind model with an initial velocity and launch direction. The simulations show that the initial magnetic polarity substantially affects the IP evolution of the CMEs influencing the propagation velocity, the shape, the trajectory (and thus, the geo-effectiveness).     

Numerical Magnetohydrodynamic (MHD) Modeling of Coronal Mass Ejections (CMEs)

The Coronal Mass Ejection (CME) is arguably the most important discovery of solar eruptive phenomena in the 20th century. It is now also recognized that CMEs have great impact on the Earth's environment by inducing geomagnetic storms. Thus, development of simulation models to understand the physical mechanisms of CME initiation and propagation has become a challenge in the solar MHD community. In this paper we shall summarize chronologically the development of the theoretical analyses, and the successes and failures of the numerical magnetohydrodynamic (MHD) simulations of coronal mass ejections (CMEs) during the past two decades. The chronological development of numerical simulation models and the evolution of the numerical methods to treat this class of problems are presented. The most appropriate way to model CMEs is to have (i) a realistic pre-event coronal atmosphere, and (ii) realistic driving mechanisms. Details of the progress and assessment of the theoretical and modeling efforts for the understanding of the physics of the CME initiation and propagation will be presented, and the numerical methods to construct these simulation models will be discussed.     

The Role of Magnetic Reconnection in CME Acceleration

Observations carried out from the coronagraphs on board space missions (LASCO/SOHO, Solar Maximum and Skylab) and ground-based facilities (HAO/Mauna Loa Observatory) show that coronal mass ejections (CMEs) can be classified into two classes based on their kinematics evolution. These two classes of CMEs are so-called fast and slow CMEs. The fast CME starts with a high initial speed that remains more or less constant; it is also called the constant-speed CME. On the other hand, the slow CME starts with a low initial speed, but shows a gradual acceleration; it is also called the accelerated and slow CME. Low and Zhang [Astrophys. J. 564, L53–L56, 2002] suggested that these two classes of CMEs could be a result of a difference in the initial topology of the magnetic fields associated with the underlying quiescent prominences. A normal prominence magnetic field topology will lead to a fast CME, while an inverse quiescent prominence results in a slow CME, because of the nature of the magnetic reconnection processes. In a recent study given by Wu et al. [Solar Phys. 225, 157–175, 2004], it was shown that an inverse quiescent prominence magnetic topology also could produce a fast CME. In this study, we perform a numerical MHD simulation for CMEs occurring in both normal and inverse quiescent prominence magnetic topology. This study demonstrates three major physical processes responsible for destabilization of these two types of prominence magnetic field topologies that can launch CMEs. These three initiation processes are identical to those used by Wu et al. [Solar Phys. 225, 157–175, 2004]. The simulations show that both fast and slow CMEs can be initiated from these two different types of magnetic topologies. However, the normal quiescent prominence magnetic topology does show the possibility for launching a reconnection island (or secondary O-line) that might be thought of as a “CME’’.     

Global Properties of Solar Flares

This article broadly reviews our knowledge of solar flares. There is a particular focus on their global properties, as opposed to the microphysics such as that needed for magnetic reconnection or particle acceleration as such. Indeed solar flares will always remain in the domain of remote sensing, so we cannot observe the microscales directly and must understand the basic physics entirely via the global properties plus theoretical inference. The global observables include the general energetics—radiation in flares and mass loss in coronal mass ejections (CMEs)—and the formation of different kinds of ejection and global wave disturbance: the type II radio-burst exciter, the Moreton wave, the EIT “wave”, and the “sunquake” acoustic waves in the solar interior. Flare radiation and CME kinetic energy can have comparable magnitudes, of order 10³² erg each for an X-class event, with the bulk of the radiant energy in the visible-UV continuum. We argue that the impulsive phase of the flare dominates the energetics of all of these manifestations, and also point out that energy and momentum in this phase largely reside in the electromagnetic field, not in the observable plasma.     

CME-Driven Solar Wind Disturbances at High Heliographic Latitudes

We have identified 20 coronal mass ejections, or CMEs, in the solar wind in the Ulysses data obtained between S30° and S75° during the second polar orbit. Unlike CME-driven disturbances observed at high latitudes during Ulysses’ first polar orbit, these disturbances had plasma and magnetic field characteristics similar to those observed in the ecliptic plane near 1 AU when one allows for evolution with heliocentric distance. Here we provide a brief overview of CME observations at high latitudes both close to and far from the Sun, with emphasis on the recent Ulysses measurements on the rising portion of solar cycle 23. This revised version was published online in August 2006 with corrections to the Cover Date.     

A Twin-CME Scenario for Ground Level Enhancement Events

Ground Level Enhancement (GLEs) events are extreme Solar Energetic Particle (SEP) events. Protons in these events often reach ～GeV/nucleon. Understanding the underlying particle acceleration mechanism in these events is a major goal for Space Weather studies. In Solar Cycle 23, a total of 16 GLEs have been identified. Most of them have preceding CMEs and in-situ energetic particle observations show some of them are enhanced in ICME or flare-like material. Motivated by this observation, we discuss here a scenario in which two CMEs erupt in sequence during a short period of time from the same Active Region (AR) with a pseudo-streamer-like pre-eruption magnetic field configuration. The first CME is narrower and slower and the second CME is wider and faster. We show that the magnetic field configuration in our proposed scenario can lead to magnetic reconnection between the open and closed field lines that drape and enclose the first CME and its driven shock. The combined effect of the presence of the first shock and the existence of the open close reconnection is that when the second CME erupts and drives a second shock, one finds both an excess of seed population and an enhanced turbulence level at the front of the second shock than the case of a single CME-driven shock. Therefore, a more efficient particle acceleration will occur. The implications of our proposed scenario are discussed.     

A Comparison of Ground Level Event e/p and Fe/O Ratios with Associated Solar Flare and CME Characteristics

Solar energetic particle (SEP) events reaching rigidities >1 GV are observed at 1?AU as ground-level events (GLEs). They are considered to be extreme cases of gradual SEP events, produced by shocks driven by wide and fast CMEs that are usually associated with long-duration (>1 hour) soft X-ray (SXR) flares. However, some large gradual SEP events, including GLEs, are associated with flares of short-duration (<1 hour) timescales comparable to those of flares seen with impulsive, low-energy SEP events with enhanced charge states, heavy-element abundances, and e/p ratios. The association of some GLEs with short-duration SXR events challenges us to understand the GLE event-to-event variation with SXR durations and whether it truly reflects the nature of the particle acceleration processes or simply the characteristics of the solar regions from which large, fast CMEs arise. We examine statistically the associated flare, active region (AR), and CME characteristics of ～40?GLEs observed since 1976 to determine how the GLE e/p and Fe/O ratios, each measured in two energy ranges, depend on those characteristics. The abundance ratios trend weakly to lower, more coronal, and less scattered values with increasing flare timescales, thermal and nonthermal peak fluxes, and measures of source AR sizes. These results and the wide range of solar longitude connections for GLEs with high abundance ratios argue against a significant role for flare effects in the GLEs. We suggest that GLE SEPs are accelerated predominately in CME-driven shocks and that a coupling of flare size and timescales with CME properties could explain the SEP abundance correlations with flare properties.     

Magnetic Flux Emergence, Activity, Eruptions and Magnetic Clouds: Following Magnetic Field from the Sun to the Heliosphere

We present an overview of how the principal physical properties of magnetic flux which emerges from the toroidal fields in the tachocline through the turbulent convection zone to the solar surface are linked to solar activity events, emphasizing the effects of magnetic field evolution and interaction with other magnetic structures on the latter. We compare the results of different approaches using various magnetic observables to evaluate the probability of flare and coronal mass ejection (CME) activity and forecast eruptive activity on the short term (i.e. days). Then, after a brief overview of the observed properties of CMEs and their theoretical models, we discuss the ejecta properties and describe some typical magnetic and composition characteristics of magnetic clouds (MCs) and interplanetary CMEs (ICMEs). We review some individual examples to clarify the link between eruptions from the Sun and the properties of the resulting ejecta. The importance of a synthetic approach to solar and interplanetary magnetic fields and activity is emphasized.     

Coronal mass ejections at high heliographic latitudes: Ulysses

Nine coronal mass ejections (CMEs) have been detected in the solar wind by the Ulysses plasma experiment between 31° and 61° South. One of these events, which was also a magnetic cloud, was directly associated with an event observed by the soft X-ray telescope on Yohkoh in which large magnetic loops formed in the solar corona directly beneath Ulysses. This association suggests that the flux rope topology of the magnetic cloud resulted from reconnection between the legs of neighboring magnetic loops within the rising CME. The average CME speed (740 km s^–1) at these latitudes was comparable to that of the normal solar wind there and is much greater than average CME speeds observed either in the solar wind in the ecliptic plane or in the corona close to the Sun. We suggest that the same basic acceleration process applies to both slow CMEs and the normal solar wind at any latitude.     

Relationship of coronal transients to interplanetary shocks: 3D aspects

More than 1000 coronal mass ejections (CMEs) caused by different types of coronal transients have been analyzed up to now, based on the images from white light coronagraphs on board the OSO 7, Skylab, P78-1, and SMM spacecraft. In many cases, the CME images lead us to the impression of loop-like, more planar structures, similar to those of prominence structures often seen in H pictures. There is increasing evidence, though, for a three-dimensional bubble- or cloud-like structure of CMEs. In several cases, CMEs directed toward the earth (or away from it) were identified, as their outer fronts emerged on all sides of the coronagraph's occulting disk, thus suggesting a bubble-like appearance.There now appears to be unanimity about the crucial role that magnetic reconnection plays during the transient process. Recently, direct evidence was found for the pinch-off of CMEs, both from optical observations and from in situ measurements of isolated magnetic clouds' following transient shock waves. However, the detailed sequence of events during the generation of a CME is still unclear.Interplanetary shock waves associated with the CMEs are usually restricted in latitudinal extent to about the angular width of the optically observed CMEs. They may be somewhat less restricted in longitudinal extent. A nearly 1 1 association between CMEs and shock waves measured in situ from spacecraft (Helios 1 and 2, IMP 7 and 8, ISEE 3, Pioneer Venus) can be established, provided the CME and the spacecraft were in the same longitudinal and latitudinal range and the CME speed exceeds 400 km s^–1. Around the past solar activity minimum all CMEs observed were centered at solar latitudes of less than 60°. Around solar maximum, a significant fraction of CMEs also originated from the polar regions. Thus, there is a good chance that the Ulysses spaceprobe will encounter many shocks caused by both low- and high-latitude CMEs, when it finally starts its journey over the Sun's poles.     

What Are Special About Ground-Level Events?

Ground level events (GLEs) occupy the high-energy end of gradual solar energetic particle (SEP) events. They are associated with coronal mass ejections (CMEs) and solar flares, but we still do not clearly understand the special conditions that produce these rare events. During Solar Cycle 23, a total of 16 GLEs were registered, by ground-based neutron monitors. We first ask if these GLEs are clearly distinguishable from other SEP events observed from space. Setting aside possible difficulties in identifying all GLEs consistently, we then try to find observables which may unmistakably isolate these GLEs by studying the basic properties of the associated eruptions and the active regions (ARs) that produced them. It is found that neither the magnitudes of the CMEs and flares nor the complexities of the ARs give sufficient conditions for GLEs. It is possible to find CMEs, flares or ARs that are not associated with GLEs but that have more extreme properties than those associated with GLEs. We also try to evaluate the importance of magnetic field connection of the AR with Earth on the detection of GLEs and their onset times. Using the potential field source surface (PFSS) model, a half of the GLEs are found to be well-connected. However, the GLE onset time with respect to the onset of the associated flare and CME does not strongly depend on how well-connected the AR is. The GLE onset behavior may be largely determined by when and where the CME-driven shock develops. We could not relate the shocks responsible for the onsets of past GLEs with features in solar images, but the combined data from the Solar TErrestrial RElations Observatory (STEREO) and the Solar Dynamics Observatory (SDO) have the potential to change this for GLEs that may occur in the rising phase of Solar Cycle 24.     

On the Coronal Magnetic Field Configuration and Solar Flare/CME Process

The coronal magnetic field configuration is important for understanding the energy storage and release processes that account for flares and/or CMEs. Here we present a model which is based on the work for potential magnetic field problems that only applies the condition at infinity with the boundary condition on the solar surface specified. We also discuss some recent progress on general force-free field models. For some event analyses, we have employed MDI/SOHO longitudinal magnetogram insected into the synoptic magnetogram to obtain whole boundary condition over the solar surface. Globally, the extrapolated global magnetic field structures effectively demonstrate the case for the disk signature of the radio CMEs and the evolution of the radio sources during the CME/flare processes.     

Properties of Ground Level Enhancement Events and the Associated Solar Eruptions During Solar Cycle 23

Solar cycle 23 witnessed the most complete set of observations of coronal mass ejections (CMEs) associated with the Ground Level Enhancement (GLE) events. We present an overview of the observed properties of the GLEs and those of the two associated phenomena, viz., flares and CMEs, both being potential sources of particle acceleration. Although we do not find a striking correlation between the GLE intensity and the parameters of flares and CMEs, the solar eruptions are very intense involving X-class flares and extreme CME speeds (average ～2000?km/s). An M7.1 flare and a 1200?km/s CME are the weakest events in the list of 16 GLE events. Most (80?%) of the CMEs are full halos with the three non-halos having widths in the range 167 to 212?degrees. The active regions in which the GLE events originate are generally large: 1290?msh (median 1010?msh) compared to 934?msh (median: 790?msh) for SEP-producing active regions. For accurate estimation of the CME height at the time of metric type?II onset and GLE particle release, we estimated the initial acceleration of the CMEs using flare and CME observations. The initial acceleration of GLE-associated CMEs is much larger (by a factor of 2) than that of ordinary CMEs (2.3?km/s² vs. 1?km/s²). We confirmed the initial acceleration for two events for which CME measurements are available in the inner corona. The GLE particle release is delayed with respect to the onset of all electromagnetic signatures of the eruptions: type?II bursts, low frequency type?III bursts, soft X-ray flares and CMEs. The presence of metric type?II radio bursts some 17?min (median: 16?min; range: 3 to 48?min) before the GLE onset indicates shock formation well before the particle release. The release of GLE particles occurs when the CMEs reach an average height of ～3.09?R _s (median: 3.18?R _s ; range: 1.71 to 4.01?R _s ) for well-connected events (source longitude in the range W20–W90). For poorly connected events, the average CME height at GLE particle release is ～66?% larger (mean: 5.18?R _s ; median: 4.61?R _s ; range: 2.75–8.49?R _s ). The longitudinal dependence is consistent with shock accelerations because the shocks from poorly connected events need to expand more to cross the field lines connecting to an Earth observer. On the other hand, the CME height at metric type?II burst onset has no longitudinal dependence because electromagnetic signals do not require magnetic connectivity to the observer. For several events, the GLE particle release is very close to the time of first appearance of the CME in the coronagraphic field of view, so we independently confirmed the CME height at particle release. The CME height at metric type?II burst onset is in the narrow range 1.29 to 1.8?R _s , with mean and median values of 1.53 and 1.47?R _s . The CME heights at metric type?II burst onset and GLE particle release correspond to the minimum and maximum in the Alfvén speed profile. The increase in CME speed between these two heights suggests an increase in Alfvénic Mach number from?2 to?3. The CME heights at GLE particle release are in good agreement with those obtained from the velocity dispersion analysis (Reames in Astrophys. J. 693:812, 2009a; Astrophys. J. 706:844, 2009b) including the source longitude dependence. We also discuss the implications of the delay of GLE particle release with respect to complex type?III bursts by ～18?min (median: 16?in; range: 2 to 44?min) for the flare acceleration mechanism. A?similar analysis is also performed on the delay of particle release relative to the hard X-ray emission.     

GeV Particle Acceleration in Solar Flares and Ground Level Enhancement (GLE) Events

Ground Level Enhancement (GLE) events represent the most energetic class of solar energetic particle (SEP) events, requiring acceleration processes to boost ?1?GeV ions in order to produce showers of secondary particles in the Earth’s atmosphere with sufficient intensity to be detected by ground-level neutron monitors, above the background of cosmic rays. Although the association of GLE events with both solar flares and coronal mass ejections (CMEs) is undisputed, the question arises about the location of the responsible acceleration site: coronal flare reconnection sites, coronal CME shocks, or interplanetary shocks? To investigate the first possibility we explore the timing of GLE events with respect to hard X-ray production in solar flares, considering the height and magnetic topology of flares, the role of extended acceleration, and particle trapping. We find that 50% (6 out of 12) of recent (non-occulted) GLE events are accelerated during the impulsive flare phase, while the remaining half are accelerated significantly later. It appears that the prompt GLE component, which is observed in virtually all GLE events according to a recent study by Vashenyuk et al. (Astrophys. Space Sci. Trans. 7(4):459–463, 2011), is consistent with a flare origin in the lower corona, while the delayed gradual GLE component can be produced by both, either by extended acceleration and/or trapping in flare sites, or by particles accelerated in coronal and interplanetary shocks.