* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download A Review of ULF Interactions with Radiation Belt Electrons
Survey
Document related concepts
Electromagnetism wikipedia , lookup
First observation of gravitational waves wikipedia , lookup
Aharonov–Bohm effect wikipedia , lookup
Elementary particle wikipedia , lookup
Introduction to gauge theory wikipedia , lookup
Relativistic quantum mechanics wikipedia , lookup
History of subatomic physics wikipedia , lookup
Electromagnetic radiation wikipedia , lookup
Wave–particle duality wikipedia , lookup
Matter wave wikipedia , lookup
Atomic theory wikipedia , lookup
Theoretical and experimental justification for the Schrödinger equation wikipedia , lookup
Transcript
GM01069_Ch12.qxd 11/9/06 8:52 PM Page 1 A Review of ULF Interactions with Radiation Belt Electrons Scot R. Elkington Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder, Colorado, USA Energetic particle fluxes in the outer zone radiation belts can vary over orders of magnitude on a variety of timescales. Power at ULF frequencies, on the order of a few millihertz, have been associated with changes in flux levels among relativistic electrons comprising the outer zone of the radiation belts. Power in this part of the spectrum may occur as a result of a number of processes, including internallygenerated waves induced by plasma instabilities, and externally generated processes such as shear instabilities at the flanks or compressive variations in the solar wind. Changes in the large-scale convective motion of the magnetosphere are another important class of externally-driven variations with power at ULF wavelengths. The mechanism for interaction between ULF variations and the radiation belts may result in (or require) pitch angle scattering, or may conserve the first two adiabatic invariants of particle motion. Of the latter class of interactions, radial diffusion describes the result when ULF variations lead to stochastic motion among the particle populations, and has been studied extensively as a description of radial transport within the belts. Rates of radial diffusion depend strongly on the characteristics of the driving ULF waves. This work is intended as a nonexhaustive review of radiation belt interactions with ULF waves, outlining the current theories and methods in studying the interaction, and describing pertinent wave properties. 1. THE RADIATION BELTS where perturbations have characteristic frequencies of a few millihertz. The outer zone radiation belts are largely field-aligned structures, with equatorial crossing distances of approximately 3–7 Earth radii (RE). This region of space is of particular significance due to the large number of spacecraft operating at these altitudes, and global society’s increasing reliance on space-based platforms for communications, navigation, weather prediction, and a variety of other economic and geopolitical purposes. Spacecraft operating within the belts have sometimes been found susceptible to “anomalies” related to changes in the local radiation environment [e.g., Wrenn, 1995; Baker et al., 1994, 1998a]. For this reason, considerable effort has been put into observing and understanding the basic physics driving changes in the outer zone belts. Energetic electron fluxes in the outer zone show substantial variation over a variety of time scales. There is a noted association between solar wind speed and relativistic electron enhancements [Paulikas and Blake, 1979; Li et al., The outer zone radiation belts consist of energetic electrons trapped in the geomagnetic field. The dynamics of the belts are dictated by the global and local electric and magnetic fields which guide the individual particles’ motion. The plasmas and magnetic fields comprising the geomagnetic environment support a large variety of waves resulting from a range of internal and external processes. This paper reviews what is currently understood of the interactions between energetic electrons in the outer zone and electric and magnetic field variations occurring on ULF time scales, Solar Eruptions and Energetic Particles Geophysical Monograph Series 165 This paper is not subject to U.S. copyright. Published in 2006 by the American Geophysical Union 10.1029/165GM12 1 GM01069_Ch12.qxd 2 11/9/06 8:52 PM Page 2 A REVIEW OF ULF/RADIATION BELT INTERACTIONS 1997; Baker et al., 1997]. Rapid enhancement of the belts, on time scales of a few minutes, has been observed during sudden commencement of storms [Blake et al., 1992b,a]. More frequently discussed are the flux variations associated with the main phase and recovery phase of geomagnetic storms, which may occur over a range of a few hours to 1–2 days [Li et al., 1998; Reeves et al., 2003]. These variations occur outside of the flux dropout and recovery normally associated with the large scale changes in the geomagnetic field during storms, the so-called “Dst effect” [e.g., Kim and Chan, 1997]. While storms are often associated with enhancements in the radiation belts, Reeves et al. [2003] showed that they can in fact produce a variety of outcomes in terms of enhancing or depleting the belts. Indeed, there are a variety of processes capable of affecting the belts in terms of the transport, heating, and loss of particles [e.g., Friedel et al., 2002; O’Brien et al., 2003], and the outcome of any substantial geomagnetic activity is a result of the complex interplay among the range of active processes. The motion of the trapped particles comprising the belts can be decomposed into three essential types, consisting of a gyromotion about the field line, a periodic motion parallel to the field lines as the particles bounce between magnetic mirror points, and a gradient-curvature drift about the Earth. Time scales associated with each type of motion for particles trapped in the radiation belts typically differ by 1–2 orders of magnitude, ranging from kHz time scales associated with the gyromotion, to mHz time scales corresponding to the drift. Particle dynamics may be ordered in terms of an “adiabatic invariant” associated with each of these types of motion [Northrop, 1963; Schulz, 1996], which will be conserved so long as the fields that guide the particle motion do not change significantly on temporal or spatial scales commensurate with the particle motion. The invariant associated with the gyromotion is given by M = p2⊥/2m0B, where p is the relativistic particle momentum, m0 the rest mass, and B the magnetic field measured at the guiding center of the gyromotion. The second invariant is based on the parallel motion evaluated along the guiding center trajectory of the particle bounce motion, and the third invariant results in a form equal to magnetic flux enclosed by the guiding center drift path about the Earth. Typically the third invariant is restated as L*, a quantity related to the inverse of the magnetic flux enclosed by the particle drift [Roederer, 1970]. In the special case of a dipole magnetic field, L* is often denoted simply as “L”, and is numerically equivalent to the radial distance (measured in RE) at which a drifting particle crosses the equatorial plane and frequently. Because the dipole L is so directly related to the L* as defined by Roederer [1970], the use of the variable L in the remainder of this review will refer to the radial dimension defined through the third invariant, rather than the simple radial distance commonly implied by the dipole limit. The widely differing time scales associated with each motion mean that field variations that lead to violation of one invariant may not necessarily perturb the other invariants. In the case of variations on drift time scales that violate the third invariant while conserving the first and second invariants, the resulting radial motion and change in guiding center field strength will lead to a change in the bounce-averaged perpendicular momentum, and hence, energy. In general, the energy change of an electron conserving the first invariant is given by dW ∂B , = qE ⋅ v d + M dt ∂t (1) where q is the particle charge, E is the background electric field, and vd is the total particle drift velocity [Northrop, 1963]. 2. MAGNETOSPHERIC ULF VARIATIONS There are a variety of geomagnetic disturbances that may produce power at ULF frequencies. Field line resonances represent a particular class of ULF variations. Waveforms that have quasisinusoidal signals that exist for several periods are classified as continuous, and are further broken down into five subcategories, Pc-1–5, depending upon their frequency [Jacobs et al., 1964]. Pc-5 waves have frequencies in the range of a few mHz. A simple mathematical description of these waves was developed by Dungey [1967], and uses the idea of standing Alfvén waves on geomagnetic field lines. In the “toroidal mode” of this theory, magnetospheric shells oscillate coherently with perturbations in the azimuthal direction, and is obtained in the low-m limit where m is the global azimuthal mode structure of the oscillations. Toroidal ULF waves have an associated induced electric field in the radial direction. The mode in which the magnetic field oscillations are constrained to the meridional plane is called poloidal, and is obtained in the high-m limit. Poloidal waves are additionally characterized by compressional magnetic perturbations, and electric fields induced in the azimuthal direction. Numerous ground- and space-based observations as well as modeling efforts have confirmed the basics of this description [Hughes, 1994, and references therein]. The broader class of ULF variations may be either broadband or quasisinusoidal in nature, and may originate in a variety of internal (magnetospheric) or external (solar winddriven) processes. As an example of internal generation mechanisms, Hasegawa [1969] developed a theory for the generation of ULF waves by invoking a mirror instability in the cold plasma population drifting in the inner magnetosphere. Southwood et al. [1969] similarly considered a drift instability as a possible internal source of the ULF waves. GM01069_Ch12.qxd 11/9/06 8:52 PM Page 3 ELKINGTON More recently, Takahashi et al. [1985b] proposed anisotropy in the perpendicular ring current as source region. Shear flow instabilities along the magnetopause provide one means of generating ULF energy in the magnetosphere [Cahill and Winckler, 1992; Mann et al., 1999]. Miura [1992] used a two-dimensional magnetohydrodynamic (MHD) simulation to show that the velocity shear along the magnetopause along the tail flanks is unstable to the KelvinHelmholtz effect regardless of the sonic Mach number of the magnetosheath flow. This research showed that a ULF wave could be generated by energy and momentum transfered into the LLBL. Mathie and Mann [2000b] used observations of Pc-5 waves with Common Azimuthal Phase Speeds (CPS) as a diagnostic for the generation of ULF waves by a shear flow instability along the magnetopause. Figure 1, taken from their paper, schematically shows these waves occurring preferentially in the morning and afternoon sectors. The speed of the “ripple” generated by the shear instability along the flanks couples into the field line resonances present on high latitude fields, with each resonance having the same phase speed due to the common source. This figure implies an equal likelihood for wave occurrence on the dawn and dusk sides of the magnetosphere; the work of Greenstadt et al. [1981], by contrast, suggests that because the dawnside of magnetopause forms a quasiparallel shock configuration, a larger seed population of perturbations is present here, resulting in more ULF waves on the dawnside than on the duskside. A second category of external driver for the generation of ULF waves within the magnetosphere is related to the transmission of variations in solar wind pressure through the 3 magnetosphere, allowing the flow of energy in the form of compressional waves [Kivelson and Southwood, 1988; Lysak and Lee, 1992] which may couple with toroidal- or poloidalmode field line resonances in the inner magnetosphere [Mann et al., 1995]. Spacecraft observations of strong dynamic pressure variations upstream of the Earth’s bow shock were shown to cause wavelets along the magnetopause and drive quasi-periodic variations within the magnetosphere [Sibeck et al., 1989a,b]. Wright and Rickard [1995] proposed that discrete variations in the solar wind driving would create field line resonances with independent phase speeds (IPS) across a wide range of local times, with maximum being centered around noon as illustrated in Figure 1. Kepko et al. [2002] and Kepko and Spence [2003] completed a spectral analysis for a series of events during which ULF waves were seen at a series of discrete frequencies in the Pc-5 range. A subsequent analysis of the solar wind density driving the magnetosphere during the same intervals revealed significant wave power at the same set of discrete frequencies as observed in the geomagnetic pulsations. They proposed that the driving of the ULF waves was not accomplished by excitation of a cavity mode, but by the slow modulation of the entire magnetospheric cavity as it changes size to maintain pressure balance under slowly varying driving conditions. A third type of magnetospheric field variation that may have power at ULF frequencies is related to changes in the large-scale convective motion of the magnetosphere as a result of changes in the global magnetospheric reconnection rate and the flow of solar wind around the magnetopause. In the Earth’s frame, the quasi-steady convective motion of fields and plasmas may be described in terms of a global potential electric field, generally pointing from dawn to dusk. The magnitude of Earth’s convection electric field may respond directly to variations in the direction and strength of the interplanetary magnetic field and the solar wind electric field [Ridley et al., 1997, 1998; Ruohoniemi and Greenwald, 1998]. While not often considered a “ULF wave” in the sense of the compressional variations and field line resonances described above, ULF variations in the Earth’s convection electric field occur on a global scale, and may have a direct impact on trapped energetic particles in the magnetosphere. This impact is described in more detail in Section 4. 3. OBSERVATIONS: THE ULF/RADIATION BELT CONNECTION Figure 1. Schematic [Mathie and Mann, 2000b] showing the shear generation of ULF waves along the dawn and dusk flanks, while ULF waves generated by stationary pulses in the solar wind are predominant in the noon sector. Much of the recent attention on the connection between ULF waves and radial transport has resulted from radiation belt flux increases occurring on time scales ranging from a few hours to 1–2 days, commonly observed following onset of geomagnetic storms. For example, in a comparison of the May 27, 1996, magnetic cloud event with that of January 10-11, 1997, GM01069_Ch12.qxd 4 11/9/06 8:52 PM Page 4 A REVIEW OF ULF/RADIATION BELT INTERACTIONS Baker et al. [1998c] found that large-amplitude oscillations in the Pc-5 frequency range were associated with the relativistic electron event of the 1997 storm, while the 1996 storm, which did not exhibit extensive Pc-5 activity, had no comparable increase in electron fluxes. Likewise, an analysis of the May 15, 1997, electron event also showed large increases in Pc-5 activity [Baker et al., 1998b]. Nakamura et al. [2002] used in situ wave measurements from Equator-S in combination with particle measurements from POLAR during the March 10, 1998, storm. Observations indicated a rapid enhancement of electrons in association with increased ULF wave activity during the main phase of the storm, during which the electrons exhibited a pitch angle distribution strongly peaked toward 90°. A subsequent, more gradual period of electron acceleration was observed, this having a more isotropic pitch angle distribution. Green and Kivelson [2001] examined ULF wave power and electron response for nine storms in 1997, noting an association between ULF wave power and electron fluxes in most of the cases. Mathie and Mann [2000a] carried out an epoch study of 17 storms in 1995, examining ULF waves observed by ground magnetometers and electron flux measurements from geosynchronous. Here storms were subcategorized according to energetic electron response during the storm, which indicated that significant electron flux increases were only observed in response to ULF wave power which is sustained at high levels over a number of days following storm onset. A study by Rostoker et al. [1998] examined a 90 day period in 1994, comparing GOES-7 >2 MeV electron flux with ULF wave activity observed by a ground magnetometer at Gillam in Manitoba, Canada. The results are shown in Figure 2, and visually indicate a strong correlation between levels of observed ULF activity and energetic electron flux. Throughout the period under study, increases in energetic electron fluxes are seen to be preceded by increases ULF wave activity by a period of 1–2 days. More quantitative studies of the correlation between ULF waves and energetic electron dynamics have been undertaken. O’Brien et al. [2001] used a cross correlation to determine which parameters in the solar wind and magnetosphere might most influence energetic electron fluxes. Similar to the results epoch study conducted by Mathie and Mann [2000a], they found sustained solar wind speeds in excess of 450 km/s to be a strong external indicator of increasing magnetospheric electron fluxes, and long-duration Pc-5 activity during the recovery phase of a storm to be the best discriminator between those storms that produced relativistic electrons and those that did not. This study was extended by O’Brien et al. [2003] to compare electron flux increases with both ULF waves and VLF chorus emissions to suggest that ULF waves are generally the Figure 2. ULF wave power observed by a ground magnetometer plotted together with energetic electron fluxes observed at geosynchronous orbit. Throughout the 90 day period there is a strong association between energetic electron fluxes and ULF wave activity (from Rostoker et al. [1998]). GM01069_Ch12.qxd 11/9/06 8:52 PM Page 5 ELKINGTON geoeffective driver of radiation belt dynamics outside of 5 RE geocentric. More recently, Mann et al. [2004] used an array of magnetometers to determine a rank-order correlation between intervals of high solar wind speed, increased ULF wave power, and subsequent increases in MeV electron fluxes, for a period spanning a complete solar cycle (1990–2001). This work indicated that the vsw and ULF wave power had highest correlation with MeV electron flux late in the declining phase of the solar cycle, when electron fluxes were highest. 4. ULF/PARTICLE INTERACTIONS: THEORY AND MODELING There have been a range of possible interactions suggested to explain the observed relation between ULF waves and energetic electron fluxes. They may be broadly categorized according to whether the mechanism features or requires violation of the first or second adiabatic invariants, or whether the interaction acts through violation of the drift invariant only. 4.1. ULF Interactions with Pitch Angle Scattering Based on the observations of Rostoker et al. [1998] depicted in Figure 2, Liu et al. [1999] proposed a “magnetic pumping” acceleration mechanism, whereby acceleration was accomplished by ULF waves in combination with pitch angle scattering via whistler-mode turbulence. In this mechanism, the electrons are acted on by a global, symmetric (m = 0) magnetic compression, with attendant induced electric field. Since symmetric compressions by themselves cannot energize particles [Fälthammar, 1965], the mechanism required scattering by whistler waves in order to break particle adiabaticity. The ULF waves provide the energization through the E ⋅ vd component of equation (1); in a symmetric perturbation, the time integral of this function is zero. However, since the drift velocity is pitch angle-dependent [Roederer, 1970], random scattering by the whistler waves can lead to a net change in energy over time. Acceleration of the particles in aggregate is attributed to the “zero energy barrier”, where an ensemble of particles undergoing a random walk in energy space will experience a net acceleration because particles cannot be decelerated past a “zero energy”, a point where no more energy can be lost; the potential for energy acceleration, on the other hand, is unbounded. This mechanism predicts the most efficient acceleration at small pitch angles, and is most effective when the wave period and scattering time scales are approximately equal. Summers and Ma [2000] offered an alternate mechanism for accelerating particles via interaction with ULF waves, including pitch angle scattering as a fundamental element of the interaction. In the process of “transit time acceleration”, 5 electrons with a particle velocity exceeding the local Alfvén velocity interact gyroresonantly with oblique fast-mode MHD turbulence. This interaction is a resonant form of Fermi acceleration, so named because the interaction requires λ||/v|| ≈ T, where v|| is the parallel particle velocity and T and λ|| are the wave period and parallel wavelength, respectively. Under geomagnetically active conditions, Summers and ma [2000] suggest that transit-time acceleration can lead to heating of keV electrons to MeV energies in timescales of a few hours. 4.2. ULF Interactions Conserving the First and Second Invariants Since the minutes-long periods associated with ULF waves are commensurate with electron drift timescales in the radiation belts, but orders larger than the gyro or bounce timescales, many theories on the relation between electron fluxes and ULF waves assume only violation of the third invariant. During the January 9-11, 1997, geomagnetic storm, Hudson et al. [1999, 2000] noted large-scale ULF waves occurring concurrently with a rapid rise in relativistic electron fluxes. The waves were largely monochromatic, with a dominant toroidal polarization. The observation of these waves, in conjunction with the energetic electron response, led to the suggestion that electrons could be accelerated via a drift-resonant interaction with global toroidal-mode field line resonances. The mechanism was quantified by Elkington et al. [1999, 2003], and depends essentially on the noonmidnight asymmetry that results from the dynamic interaction of the solar wind with the Earth’s magnetosphere. The basic mechanism is depicted in the leftmost panel of Figure 3. Here an equatorial electron drifting in a compressed dipole experiences an interaction with a low mode number global toroidal wave [e.g., Takahashi and McPherron, 1984]. On the dawn side, the electron has a component of motion in the radially outward direction; at dusk, radially inward. This permits energization through interaction with the radial electric field associated with toroidal-mode field line resonances via the first term in equation (1). Those electrons drifting with a frequency ωd satisfying ω = (m ± 1)ωd (2) where ω is the wave mode number and frequency, respectively, will see an electric field of constant sign throughout their motion, and thus will experience a continual acceleration. Further observational evidence for electron acceleration via toroidally-polarized field line resonances was reported in Tan et al. [2004] for an event occurring in August 1991. Since the largest component of a drifting electron’s motion is in the azimuthal direction, interaction with waves with an GM01069_Ch12.qxd 6 11/9/06 8:52 PM Page 6 A REVIEW OF ULF/RADIATION BELT INTERACTIONS associated azimuthal electric field is likely to be of generally more importance to radiation belt dynamics than interaction relying on the radial motion of the electrons as described above. Any of the wave generation mechanisms described in Section 2 may include azimuthal components of the electric field. The idealized toroidal mode described above represents a low-mode number limit; for finite mode number the Alfvénic disturbances will occur in mixed modes with both toroidal and poloidal signatures and radial and azimuthal electric fields [Ukhorskiy et al., 2005; Loto’ aniu et al., 2006]. The resonant acceleration for an electron drifting in a dipole magnetic field and interacting with an azimuthal electric field is depicted in the center frame of Figure 3. The resonance condition for this azimuthally symmetric interaction is given by ω = mωd. (3) An asymmetric resonance exists also for the poloidal component of a ULF disturbance, and is depicted in the rightmost panel of Figure 3. Here the resonance condition is the same as for the toroidal mode discussed above. A particle drifting in an asymmetric magnetic field will be subject to the combined effects of each of the resonances outlined in the figure, the degree of interaction depending on the wave power at frequencies corresponding to the asymmetric or symmetric resonance conditions described in equations (2) or (3). As an aside, we note that the internally-generated poloidal mode ULF waves discussed in Section 2 [Southwood et al., 1969; Takahashi et al., 1985b] are typically high-m in character [Takahashi et al., 1985a]. Since wave power in the ULF spectrum is typically a decreasing function of frequency [Bloom and Singer, 1995; Brautigam et al., 2005; Elkington et al., 2004], the resonance conditions (2) and (3) indicate that these internally-generated sources of wave power will be of less importance to overall radiation belt dynamics. The use of Poincaré plots to map the dynamics of a particle in phase space provides a powerful tool for analyzing the resonant interactions described above [Elkington et al., 2003; Degeling et al., 2006; Loto’aniu et al., 2006]. An example is shown in Figure 4, showing the resonant interaction of an electron with a poloidal-mode ULF wave. The plot indicates the dynamics in a space consisting of the third invariant L and Figure 3. Sketches indicating resonant interaction for electrons with the toroidal component of a global field line resonance in a magnetic field compressed by the solar wind (a), and for electrons interacting with the poloidal component of the ULF waves in a dipole (b) and compressed (c) magnetic field. Top panels in each case show resonant fields at particular locations along representative electron orbits; bottom panels indicate the drift velocities, fields, and instantaneous energy gain throughout a resonant trajectory [Elkington et al., 2003]. GM01069_Ch12.qxd 11/9/06 8:52 PM Page 7 ELKINGTON Figure 4. Poincaré plot indicating the resonant interaction of an electron with the poloidal component of a ULF wave [Loto’aniu et al., 2006]. the corresponding drift angle variable, here the particle drift phase measured relative to the phase of the propagating resonant ULF wavefront. Particles move through this space on trajectories constrained by the indicated lines; those with trajectories inside the separatrix of the resonant island centered at ∼3.5 MeV are resonant, and are able to experience energy changes dictated by the extent of their possible motion along the energy (or L) axis. 4.3. ULF Variations and Radial Diffusion Magnetospheric power at ULF frequencies may be either largely monochromatic, or it may be broadband in nature [e.g., Brautigam et al., 2005]. For the general case of finite bandwidth variations, the drift resonant interactions described above will lead to stochastic motion among individual particles in the third invariant, L. This can be seen in Figure 5, where the separate interaction of two independent toroidal-mode waves at differing frequencies leads to diffusive motion in the regions of resonant island overlap. In the case that the frequency spectrum approaches a continuum over some range, as shown in the bottom panel of Figure 5, the motion becomes fully stochastic in energy, or equivalently, L. A distribution of particles undergoing this type of stochastic motion in L can be shown to satisfy a diffusion equation of the form ∂f ∂ DLL ∂f f = L2 − , ∂t ∂L L2 ∂L τ L (4) where f is the phase space density averaged over all phase angles, a function of M, L, and t [Roederer, 1970; Schulz and Lanzerotti, 1974], and τL is the characteristic particle lifetime. The radial diffusion coefficient DLL describes the average rate of radial transport; analytic forms for the diffusion 7 coefficient may be derived starting with either the drift equations describing the influence of the perturbing waves [Fälthammar, 1965, 1968; Schulz and Eviatar, 1969; Schulz and Lanzerotti, 1974; Ukhorskiy et al., 2005], or starting from a Hamiltonian formulation [Brizard and Chan, 2001]. In general, the diffusion coefficient will be an explicit function of L, the wave power and mode structure of the driving waves, and will contain either an implicit or explicit dependence on particle energy. Further, the relevant form of the diffusion coefficient will depend on whether the waves are electromagnetic in nature, or whether the wave can be ascribed to variations in a large-scale potential field [Cornwall, 1968]. Diffusion resulting from interaction with the symmetric resonance (3) may take one of two forms, depending on the nature of the waves. If the field perturbations driving the diffusion results from variations in a potential electric field, or the magnetic field perturbations in equation (1) are otherwise zero, the diffusion coefficient for particles drifting with frequency ωd takes the form E , Sym DLL = 1 L6 ∑ PmE ( L, mω d ). 8 BE2 RE2 m (5) Here BE and RE are the Earth’s dipole moment and radius, respectively, and PmE is the power spectral density of the perturbing electric fields at the resonant frequency mωd, where m describes the global azimuthal structure of the waves. Variations in the dawn-dusk electric field associated with global magnetospheric convection is one example of such field perturbations [Cornwall, 1968]; the ideal toroidal field line resonance acting on an equatorially-bound particle (where the associated magnetic perturbation has a null) provides a second example [Loto’aniu et al., 2006]. If, however, the perturbation is electromagnetic in nature, where magnetic and electric perturbations are related by Faraday’s law, then the diffusion can be described by the expression B , Sym DLL = M2 L4 ∑ m 2 PmB ( L, mω d ). 8q γ 2 BE2 RE4 m 2 (6) In this expression, γ is the Lorentz factor, M the relativistic first adiabatic invariant, and PmB is the power spectral density of the compressional wave magnetic field at frequency mωd. The electric diffusion coefficient (5) has an L6 dependence, plus the L dependence in P Em. The L dependence of the magnetic diffusion coefficient is more complicated since γ also depends on L. In the ultra-relativistic limit γ 2 ∝ L−3, so for radiation belt electrons L4/γ 2 is approximately proportional to L7, not including the L dependence implicit in P Bm. Note that equation (6) does not exhibit the same explicit L10 GM01069_Ch12.qxd 8 11/9/06 8:52 PM Page 8 A REVIEW OF ULF/RADIATION BELT INTERACTIONS Figure 5. Poincaré plot showing the onset of radial diffusion for electrons interacting with toroidal-mode ULF waves spanning a range of frequencies [Elkington et al., 2003]. dependence as that of Fälthammar [1965], which is nonrelativistic, has only m = 1 terms, and assumes an explicit radial dependence in the perturbing fields. The effects of both the magnetic perturbation and the induced electric perturbation contribute to the diffusion described by equation (6). The asymmetric diffusion coefficients have different resonance frequencies, dictated by equation (2), and are also proportional to the square of the global asymmetry in the Earth’s magnetic field [Elkington et al., 2003; Perry et al., 2005; Ukhorskiy et al., 2005], here described by the factor b1(r)/BE. For variations in a potential electric field, we find 2 E , Asym LL D 2 b = 2 2 1 L12 9 BE RE BE × ∑ m 2 [ PmE ( L,(m + 1)ω d ) + PmE ( L, (m − 1)ω d )], (7) m whereas if the perturbations are electromagnetic in nature, we get 2 B , Asym LL D b 2M 2 = 2 2 2 4 1 L10 9q γ BE RE BE × ∑ m 2 [ PmB ( L,(m + 1)ω d ) + PmB ( L,(m − 1)ω d )]. (8) m The asymmetric modes have steeper L dependence, indicating they may be more important at larger L-shells. Specifically, the electric asymmetric diffusion coefficient has an explicit L12 dependence, and in the ultrarelativistic limit B , Asym the magnetic asymmetric diffusion coefficient, DLL , has 13 approximately an L dependence. In each case there may be an additional L dependence in the b1(r)/BE term and in the power spectral density, P B,E m . GM01069_Ch12.qxd 11/9/06 8:52 PM Page 9 ELKINGTON The total diffusion coefficient is the sum of equations (5) to (8), including asymmetric and symmetric diffusion coefficients and all m modes: Asym , m −1 Sym , m Asym , m +1 DLL = ∑ ( DLL + DLL + DLL ) (9) m Note that in expressions (5) through (8), the stochastic motion of the particles is a result of the intrinsic randomness of superposed disturbance fields with a range of differing frequencies and corresponding phases, as seen by the particles. However, the needed stochasticity may arise as a result of extrinsic variations of the driver as well. For example, Ukhorskiy et al. [2005] conducted a study that included an analysis of the effects of monochromatic electric waves occurring over a limited extent in local time in an asymmetric magnetic field, but driven by a randomly-varying external driver. Similar to equation (5) and equation (7), an L6 dependence is found for azimuthal electric fields in a dipole, and an L12 dependence is found for radial electric fields effective as a result of the asymmetry in the magnetic field. Degeling et al. [2006] likewise considered waves occurring in a limited local time extent, here using a wave model describing a scenario of a fast compressional wave with a single discrete frequency coupling to a field line resonance at a particular L. As depicted in Figure 6, they show that particles may undergo a transition from ordered to stochastic behavior if acted upon by monochromatic waves with an amplitude that depends on local time, without the need for a random external driver. Figure 6. Poincaré plot showing the onset of radial diffusion when acted on by monochromatic ULF waves with increasing azimuthal asymmetry (from Degeling et al. [2006]). 9 This change in dynamics is attributed to the appearance and eventual overlap of additional resonance islands produced by the spectral components of the azimuthal wave structure. 5. EMPIRICAL QUANTIFICATION OF THE EFFECTS OF ULF WAVES There has been considerable effort to determine diffusion rates based on observational evidence, using a number of different techniques. A summary of a few of these results is indicated in Table 1, taken from Elkington et al. [2003]; more extensive compilations can be found in Lanzerotti et al. [1970], Tomassian et al. [1972], Holzworth and Mozer [1979], Riley and Wolf [1992], and elsewhere. Two of the works listed in the table, Lanzerotti and Morgan [1973] and Holzworth and Mozer [1979], assumed that diffusion rates were correctly given by equations (5) or (6) and used measured magnetic and electric field spectra to arrive at the relevant diffusion coefficients. The remainder of those listed in the table arrived at diffusion rates without an assumed L-dependence. The characteristic time scales exhibited in Table 1 range from ∼1 day at L = 6 to several years for electrons in the inner belt. Except in the case of Lanzerotti and Morgan [1973], who found that the diffusion coefficient at L = 4 could vary by orders of magnitude with magnetic activity, the observations used in these analyses were generally made during quiet geomagnetic periods. Two of the works listed in Table 1, Newkirk and Walt [1968b] and Selesnick et al. [1997], made an attempt to determine the radial dependence of outer zone diffusion coefficients based on particle measurements. The approach used by Selesnick et al. [1997], for example, examined electron dynamics over a 3-month period in 1996. Assuming a radial diffusion coefficient of form DLL = D0 (L/4)n and a characteristic particle lifetime of form τ = τ0 (L/4)−m, they were able to fit observations from three clear electron injections to a timedependent radial diffusion equation with losses [Schulz and lanzerotti, 1974] and deduce a radial diffusion coefficient varying as L(11.7 ± 1.3), where the error bars were based on statistical rather than systematic considerations. Newkirk and Walt [1968b] similarly found a radial diffusion coefficient of L(10 ± 1). In addition to the works listed in Table 1, two recent works deduce relevant diffusion time scales based on field measurements seen by CRRES, a spacecraft in an equatorial geotransfer orbit that precessed over a range of local times during its lifetime. Brautigam et al. [2005] assumed all the observed electric power was in potential electric fluctuations (rather than electromagnetic), and parameterized their observations according to local time and geomagnetic activity measured by Kp. By applying the results to the diffusion GM01069_Ch12.qxd 10 11/9/06 8:52 PM Page 10 A REVIEW OF ULF/RADIATION BELT INTERACTIONS Table 1. Examples of of empirically-determined radial diffusion coefficients (from Elkington et al.[2003]) DLL(day−1) L −7 2.0 × 10 4.0 × 10−6 10−8 L(10 ± 1) a 4 – 8 × 10−10 L10 a 2.7 × 10−5 M −0.5L7.9 10(0.75ΚFR – 10.2) b (2.23 ± 0.67)ωd−1.1 ± 0.15 c ∼0.2 – 5.0 W or M range Reference L = 1.20 L = 1.16 1.76 ≤ L ≤ 5.0 3.0 ≤ L ≤ 5.0 1.7 ≤ L ≤ 2.6 L=4 L = 6.0 L = 5.3, L = 6.1 W > 1.6 MeV Newkirk and Walt [1968a] W > 1.6 MeV W > 0.5 MeV 13.3 – 27.4 MeV/G 350 – 750 MeV/G 50 keV – 1.2 Mev 100 keV – 1.6 Mev Newkirk and Walt [1968b] Lanzerotti et al. [1970] Tomassian et al. [1972] Lanzerotti and Morgan [1973] Holzworth and Mozer [1979] Chiu et al. [1988] 3.0 ≤ L ≤ 6.0 3 – 8 MeV Selesnick et al. [1997] 11.7 ± 1.3 L 2.1 × 10−3 4 aD based on an assumed L10 radial dependence. 0 bBased on equation (6) and measured magnetic power cBased spectrum; KFR ≡ Fredricksburg magnetic index (see also Brautigam and Albert [2000]). on equation (5) and measured electric power spectrum; ωd in h−1. E that varied by coefficient (5), they found a Kp-dependent DLL up to two orders of magnitude between Kp = 1 and Kp = 6 depending on the L shell. Ukhorskiy et al. [2005] similarly used CRRES electric field measurements in their analysis of the diffusive effects of monochromatic ULF variations. Diffusion resulting from these monochromatic waves at L ∼ 6 had characteristic time scales on the order of a day. In spite of the strong explicit L dependence of any of the the forms (5) through (8), at times, radial diffusion may act at low L on very fast time scales. For example, Loto’aniu et al. [2006] combined ground magnetometer measurements of ULF waves with in situ measurements of energetic electrons during the October 31, 2001, storm, calculating diffusion E by mapping the wave observations from the coefficients D LL ground to the equatorial plane using the methods of Ozeke and Mann [2004]. They suggested that radial diffusion time scales were sufficient to account for the two-order of magnitude, ∼24 hour increase in >2 MeV electrons within the slot region ( L 2 − 2.5) . 6. QUANTIFYING DIFFUSION: MODELING EFFORTS As alluded in previous sections, numerical modeling can provide considerable insight into the nature of radial diffusion, both in terms of the basic physical characteristics of the diffusive process, as well as its overall applicability in radiation belt dynamics. Test particle simulations using idealized geomagnetic and wave fields has provided a method of examining the effects of ULF waves on radiation belt dynamics in a controlled way. Elkington et al. [1999, 2003] examined the dynamics of equatorially-trapped particles in a simplified magnetic field model of a compressed dipole. Applying an electric field model of global-scale ULF waves, they were able to examine the nature of the asymmetric resonances (2) for waves with both toroidal and poloidal polarizations, and take a first look at the nature of radial diffusion that results from these higherorder resonances. Perry et al. [2005] extended this work in an examination of the effect of the poloidal mode on particles with motion out of the equatorial plane. The ULF wave model used in the Perry et al. effort was based on a vector potential, and included the effects of the compressional magnetic perturbations necessary to examine both terms in equation (1). This work also incorporated variations of wave power with frequency and radial location, and concluded that radial diffusion was most effective for particles with large equatorial pitch angles and at higher L values. Test particle simulations by Ukhorskiy et al. [2005] and Degeling et al. [2006], described in Section 4.3, examined the physical effects of waves occurring over a finite local time. Ukhorskiy et al. [2005] determined the rates of diffusion expected from externally-driven random waves of a single frequency, while Degeling et al. [2006] determined that finite local time effects by themselves could induce radial diffusion. A useful aspect of these controlled simulations is their ability to determine relevant time scales for a given wave interaction, important both as a means of verifying the functional forms of the resulting diffusion, and as a means of determining the relative effects of diffusion for different particle populations in different regions of the magnetosphere. An example from Ukhorskiy et al. [2005] is shown in Figure 7. Here an ensemble of particles starting from a particular L value is allowed to interact with waves occurring in the noon sector of the magnetosphere. The dynamics of individual particles are indicated by the grey lines; in aggregate, the mean spread of the particles in time is given by the dotted line. The diffusion coefficient may be easily determined GM01069_Ch12.qxd 11/9/06 8:52 PM Page 11 ELKINGTON Figure 7. Calculation of rates of radial diffusion of an ensemble of particles as considered by Ukhorskiy et al. [2005]. Grey lines indicate the motion of individual particles in the ensemble, dotted line indicates the mean-squared deviation of the ensemble, and the solid line is a fit which allows calculation of DLL. from the slope of a line fit to this mean spread, using the definition DLL ≤ ∆L2 > / 2t [Schulz and Lanzerotti, 1974]. Test particle simulations carried out using physical field models driven by realistic solar wind conditions do not offer the same controlled environment for examining diffusion as the empirical models described above, but do allow examination of the effect of low frequency waves under more realistic conditions. Using MHD/particle simulations of the magnetosphere, Hudson et al. [1999, 2000, 2001] and Elkington et al. [2002] examined the effects of ULF waves during a series of geomagnetic storms, and suggested that the resulting diffusive effects were an important driver of particles during the main phase of storms. Direct solution of the radial diffusion equation (4) provides another means of looking at the global effects of radial diffusion induced by ULF variations. Selesnick et al. [1997] solved a lossy radial diffusion equation during an extended quiet period occurring in 1996, fit to data to empirically determine the relevant rates of radial diffusion and loss. Several works have compared the results of radial diffusion simulations, using presumed diffusion coefficients, to data, in an effort to determine contrasting effects of radial diffusion vs. local heating [e.g., Summers et al., 1998]. Brautigam and Albert [2000] solved the radial diffusion equation in an analysis of a storm occurring in October of 1990. They considered the effects of both electromagnetic and potential electric fields, using symmetric forms for DLL that varied with Kp to capture the effects of increased ULF activity during geomagnetically-active times. Comparing the results to CRRES observations, they concluded that radial diffusion was the dominant process among electrons at keV energies, 11 but that radial diffusion alone was not enough to account for the dynamics of >MeV electrons. Shprits and Thorne [2004] used the diffusion coefficients corresponding to the electroB , Sym magnetic DLL coefficients of Brautigam and Albert, in a study of the competing effects of diffusion and loss in the outer zone, and similarly concluded that radial diffusion was not the only process at work in the radiation belts. The Salammbô diffusive radiation belt code [Beutier and Boscher, 1995; Bourdarie et al., 1996] attempts to simulate not only the effects of radial diffusion, but also include the effects of high-frequency interactions that lead to pitch angle scattering and local heating. Rates of radial diffusion are handled in an activity-dependent manner, while pitch angle diffusion is modeled considering plasmaspheric hiss at the plasmapause. Albert and Young [2004] noted that simulating the effects of processes with such disparate characteristic time scales (particle drift vs gyromotion) can be problematic, leading to large and rapidly varying cross terms in the full diffusion tensor. They suggest a particular coordinate transformation as a means of avoiding this problem and allowing simulation including the range of high- and low-frequency effects. Insight can be gained through a combination of both test particle simulations and solution of the radial diffusion equation. Riley and Wolf [1992] used a time-dependent, observation-based model electric field for the storm event of August 1990, and used it to compare results of a drift calculation to results of a radial diffusion simulation, the latter using the power spectrum of the fluctuations in the electric field model to fix the symmetric electric diffusion coefficient considered in the study. They concluded the diffusion formalism used gave only roughly right answers for a single real storm but did much better for an average over a statistical ensemble of storms. Ukhorskiy et al. [2006] used a model of the electric fields induced by solar wind pressure variations during a period with near-zero Dst to drive a kinetic simulation of the radiation belts. They found that the induced fields resulted in effective radial transport of energetic electrons, but similarly found that the resulting particle distribution was not well described in terms of simple radial diffusion. Fei et al. [2006] conducted a study comparing a radial diffusion simulation with the results of an MHD/particle simulation of the September 1998 storm. This study conducted a global analysis of the ULF wave power occurring in the MHD model, including azimuthal mode structure using the methods of Holzworth and Mozer [1979]. They applied power appropriately to the range of symmetric and asymmetric magnetic and electric diffusion coefficients, equations (5) through (8), and used boundary conditions based on the results of a test particle simulation using the same fields. Plate 1 shows an example of the comparison between the radial diffusion and test particle results. Fei et al [2006] were able to find much better GM01069_Ch12.qxd 12 11/9/06 8:52 PM Page 12 A REVIEW OF ULF/RADIATION BELT INTERACTIONS Plate 1. Radial diffusion solution (left) compared with an MHD/particle solution (right), for the September 1998 geomagnetic storm [Fei et al., 2006]. They find effective solution of the radial diffusion equation requires detailed knowledge of the driving diffusion coefficients, and hence, ULF activity. Plate 2. keV electrons in the plasmasheet may be convectively injected into the inner magnetosphere, gaining energy through conservation of the first invariant in the process [Elkington et al., 2004]. The plasmasheet population acts as a boundary condition on the trapped particles undergoing diffusion in the inner magnetosphere. GM01069_Ch12.qxd 11/9/06 8:52 PM Page 13 ELKINGTON agreement between the two simulation methods when using all the appropriate symmetric and asymmetric diffusion coefficients, than using any single or simplified diffusion coefficient alone. 7. DOES RADIAL DIFFUSION ACT AS A SOURCE OR LOSS? As with any stochastic process acting on a distribution of particles, radial diffusion will generally act to smooth out gradients in the space describing that distribution. Therefore, for radial diffusion to act alone in populating the inner regions of the radiation belts, a source of particles at high L must be present. That is, the phase space density of the particle must generally increase out to the particle trapping boundary. Lacking this source at high L, the effect of diffusion driven by ULF variations will be to deplete the belts as particles are transported from regions of higher phase space density through the outer trapping boundary. Figure 8 illustrates three scenarios by which the distribution function of the radiation belts may evolve [Green and Kivelson, 2004]. In the leftmost panel, radial diffusion acts alone on a distribution with a monotonically increasing density. By Fick’s Law [Walt, 1994], this outward gradient will result in an inward flux of particles from the high-L boundary. In the center panel, a local heating mechanism acts to increase the phase space density at a limited range of L values, producing a local maximum in the phase space density relatively low in the magnetosphere. In this case, radial diffusion will act to move particles outward in the high-L regions where the gradient is negative, and inward in the inner regions where the gradient is positive. A similar situation will result in case illustrated in the rightmost panel, where a sudden drop in phase space density at the boundary again results in a peak at relatively low L. In either of the two 13 latter cases, those particles with a net outward drift will presumably be lost to the magnetopause or outer stable trapping boundary. Significant observational effort has been made to determine the evolving nature of the phase space density profiles, similar to the illustrations in Figure 8, as a means of determining those times when radial diffusion associated with ULF activity may be dominant, vs. those times when local acceleration processes may be active. Typically, these studies are made difficult by uncertainties in the global magnetic field configuration that dictates the particle adiabatic invariants. Recent studies include those by Green and Kivelson [2004], who used observations from the polar-orbiting spacecraft Polar to suggest that local heating processes were frequently active during the recovery phase of geomagnetic storms, forming peaks in the phase space density in the inner magnetosphere over the course of a few days following storm onset. During these times, radial diffusion driven by ULF variations would act as a loss as particles were driven out through the outer boundary. Onsager et al. [2004] used two geosynchronous spacecraft at different magnetic latitudes to calculate the phase space density gradient at L ∼ 6.6, and found at various times both positive and negative gradients with respect to L. Taylor et al. [2004] used equatorial measurements from the Cluster constellation of spacecraft, in conjunction with geosynchronous and other observations of the inner magnetosphere, and showed that there was often sufficient phase space density to allow for inward radial transport from the plasmasheet to populate the belts. Simulations by Elkington et al. [2004] for the March 2001 storm, depicted in Plate 2, provide insight into the possibility of keV plasmasheet electrons being transported inward and trapped in the radiation belts. They find that during times of strong global convection associated with geomagnetic storms, a sufficient number of Figure 8. Radial phase space density profiles as related to various processes acting on the radiation belts [Green and kivelson, 2004]. In the left panel, radial diffusion acts alone on a distribution with a monotonically-increasing distribution function. Following panels indicate phase space density bumps that result from internal acceleration (center) and a change in the outer boundary condition (right). In either of the second two cases, radial diffusion at high L will act as a loss process. GM01069_Ch12.qxd 14 11/9/06 8:52 PM Page 14 A REVIEW OF ULF/RADIATION BELT INTERACTIONS energetic electrons may be trapped in the inner magnetosphere to contribute to the overall distribution of radiation belt electrons at MeV energies. However, during times of northward IMF (last frame of Plate 2), higher energy plasmasheet electrons do not always have access to the inner magnetosphere. 8. SUMMARY AND OPEN QUESTIONS Understanding (to the point of being able to quantify) the relevant effects of ULF waves on energetic particle dynamics in the radiation belts requires knowledge of a number of specific properties of the driving perturbations and global fields. Therefore, we conclude this paper with a few of the outstanding questions related to ULF-radiation belt interactions, and a review of ULF wave properties necessary to establish and quantify their effect on the radiation belts. A basic question related to the transport of radiation belt electrons is the extent to which the dynamics may be considered in a diffusive framework, versus the extent to which kinetic effects must be taken into account. The modes of interaction described by Liu et al. [1999] and Summers and Ma [2000] explicitly involve violation of the first adiabatic invariant, and therefore cannot be considered purely in the framework of radial diffusion. Further, Riley and Wolf [1992], Elkington et al. [2003], and Ukhorskiy et al. [2005, 2006] all noted that radial diffusion only provided an adequate description of particle transport when averaged over a statistical ensemble of driving conditions. In consideration of single events, diffusion gave only a rough description of the electron dynamics, with considerable variation from the diffusive formalism that may have been attributable to the details of the driving fields and only examined from a kinetic point of view. A closely-related question is the relative importance of, and differences between, intrinsic vs. extrinsic stochasticity; that is, diffusion resulting from randomness inherent in the external driving perturbations (e.g., the solar wind), as opposed to diffusion resulting from the intrinsic spectral [Elkington et al., 2003] or spatial [Degeling et al., 2006] properties of the waves themselves. In addition to the physical properties of the ULF variations affecting the transport of radiation belt electrons in the inner magnetosphere, the global configuration of the background magnetic field may prove to have a significant effect on the interaction of ULF power with energetic particles. Elkington et al. [2003] demonstrated that global distortion resulting from Chapman-Ferraro currents at the magnetopause can lead to additional modes of interaction between particles and waves. The influence of global, steady-state convection can similarly alter basic particle drift paths and change the energy gains and losses associated with interaction with ULF waves, as can the storm-time evolution of the ring current and partial ring current. The effect of these latter interactions have not yet been examined in detail. The efficiency of the interaction between ULF power and energetic electrons depends on the power spectrum at frequencies including ω = mωd, where ωd is the (energy-dependent) particle drift frequency and m is the azimuthal mode structure of the interacting waves [Fälthammar, 1965; Schulz and Lanzerotti, 1974; Hudson et al., 2000; Elkington et al., 1999, 2003; Perry et al., 2005; Summers and Ma, 2000] Resonances higher-order in m may be induced by global asymmetries in the guiding background magnetic field [Elkington et al., 2003]. The radial extent and azimuthal extent over which the waves globally occur determines the rate at which particles will be transported/energized through interaction with ULF waves in different regions of the magnetosphere [Mathie and Mann, 2000b; Elkington et al., 2003; Tan et al., 2004; Ukhorskiy et al., 2005; Perry et al., 2005]. For example, if waves from a particular source occur over only some azimuthal fraction of a drift shell, the total effect of the ULF waves on the particles on that shell will be reduced in proportion to the fractional azimuthal occurrence of the waves. Degeling et al. [2006] shows that local time asymmetries in the power itself is enough to induce diffusion. In accordance with the resonance condition suggested above, the direction of propagation of the ULF disturbances determines whether the particles may efficiently interact with the waves. In particular, only waves with positive mode numbers will effectively transport counterclockwise-drifting electrons [Ukhorskiy et al., 2005]; variations in a global convection electric field, which can be described in terms of “standing” waves consisting of counterpropagating waves of equal magnitude and phase, will interact with trapped electrons at a level dictated by the magnitude of only the positive-m component, E0/2. Further, simulations and theoretical considerations have suggested that electromagnetic vs. potential ULF variations have explicitly different effects in L and energy among trapped particle populations in the magnetosphere [Cornwall, 1968; Fälthammar, 1968; Schulz and Lanzerotti, 1974; Perry et al., 2005]. The difference between equations (5) and (6) illustrates this point. This has particular significance, for example, when ULF variations in a given region are dominated by convective rather than compressional waves. The depth to which particles may be effectively transported in the radiation belts via interaction with ULF waves therefore depends not only on the radial penetration of the ULF waves, but explicitly on the characteristic nature of the waves as well. The need for more detailed understanding of the global nature and occurrence of the waves driving energetic particle dynamics has been underscored by recent research works by GM01069_Ch12.qxd 11/9/06 8:52 PM Page 15 ELKINGTON Perry et al. [2005], Brautigam et al. [2005], Ukhorskiy et al. [2005], and Fei et al. [2006]. For example, the observational and modeling efforts undertaken in Brautigam et al. [2005] and Ukhorskiy et al. [2005] were forced to make an m = 1 assumption for all wave power considered when estimating the effect of ULF waves on energetic particles. Perry et al. [2005] attempted to apply realistic radial and spectral envelopes to the particle simulations conducted in that study, which had to be constructed based on a limited combination of measurements [Bloom and Singer, 1995] and global MHD results [Elkington et al., 2004]. Fei et al. [2006] used a model-model comparison between a radial diffusion formulation of the radiation belts and the evolution suggested by global MHD/particle simulations of the September 1998 storm, and found that the best agreement was only obtainable when the diffusion simulation incorporated global properties such as mode structure and radial and spectral profiles of ULF activity into the appropriate transport coefficients. Indeed, ultimately disentangling the relative effectiveness of various processes acting in the radiation belts from among the zoo of possible high- (VLF) [e.g., Summers et al., 1998; Horne and Thorne, 1998] and low-frequency interactions will require an extensive knowledge of the global characteristics of the waves involved in each of these processes. Acknowledgments. Work has been supported by NSF grant ATM-0303209 and NASA grants NNG05-GG27G and NNG05GK04G at the University of Colorado. REFERENCES Albert, J.M., and S.L. Young, Multidimensional quasi-linear diffusion of radiation belt electrons, Geophys. Res. Lett., 32, L14,110, doi:10.1029/2005GL023191, 2004. Baker, D.N., S. Kanekal, J.B. Blake, B. Klecker, and G. Rostoker, Satellite anomalies linked to electron increase in the magnetosphere, Eos Trans. AGU, 75, 401, 1994. Baker, D.N., J.H. Allen, S.G. Kanekal, and G.D. Reeves, Disturbed space environment may have been related to pager satellite failure, Eos Trans. AGU, 79, 477, 1998a. Baker, D.N., et al., Recurrent geomagnetic storms and relativistic electron enhancements in the outer magnetosphere: ISTP coordinated measurements, J. Geophys. Res., 102 (A7), 14,141, 1997. Baker, D.N., et al., A strong CME-related magnetic cloud interaction with the Earth’s magnetosphere: ISTP observations of rapid relativistic electron acceleration on May 15, 1997, Geophys. Res. Lett., 25, 2975, 1998b. Baker, D.N., et al., Coronal mass ejections, magnetic clouds, and relativistic magnetospheric electron events: ISTP, J. Geophys. Res., 103, 17,279, 1998c. Beutier, T., and D. Boscher, A three-dimensional analysis of the electron radiation belt by the Salammbô code, J. Geophys. Res., 100, 14,853, 1995. Blake, J.B., M.S. Gussenhoven, E.G. Mullen, and R.W. Fillius, Identification of an unexpected space radiation hazard, IEEE Trans. Nuc. Sci., 39, 1761, 1992a. Blake, J.B., W.A. Kolasinski, R.W. Fillius, and E.G. Mullen, Injection of electrons and protons with energies of tens of MeV into L < 3 on 24 March 1991, Geophys. Res. Lett., 19, 821, 1992b. Bloom, R.M., and H.J. Singer, Diurnal trends in geomagnetic noise power in the Pc2 through Pc5 bands at low geomagnetic latitude, J. Geophys. Res., 100, 14,943, 1995. Bourdarie, S., D. Boscher, T. Beutier, J.-A. Sauvaud, and M. Blanc, Magnetic storm modeling in the Earth’s electron belt by the Salammbô code, J. Geophys. Res., 101 (A12), 27,171, 1996. Brautigam, D.H., and J.M. Albert, Radial diffusion analysis of outer radiation belt electrons during the October 9, 1990, magnetic storm, J. Geophys. Res., 105 (A1), 291, 2000. Brautigam, D.H., G.P. Ginet, J.M. Albert, J.R. Wygant, D.E. Rowland, A. Ling, and J. Bass, CRRES electric field power spectra and radial diffusion coefficients, J. Geophys. Res., 110 (A2), A02,214, doi:10.1029/2004JA010612, 2005. 15 Brizard, A.J., and A.A. Chan, Relativistic bounce-averaged quasilinear diffusion equation for low-frequency electromagnetic fluctuations, Phys. Plasmas, 8 (11), 4762, 2001. Cahill, L.J., Jr., and J.R. Winckler, Periodic magnetopause oscillations observed with the GOES satellites on March 24, 1991, J. Geophys. Res., 97 (A6), 8239, 1992. Chiu, Y.T., R.W. Nightingale, and M.A. Rinaldi, Simultaneous radial and pitch angle diffusion in the outer electron radiation belt, J. Geophys. Res., 93 (a4), 2619, 1988. Cornwall, J.M., Diffusion processes influenced by conjugate-point wave phenomena, Radio Sci., 3 (7), 740, 1968. Degeling, A.W., R. Rankin, K. Kabin, R. Marchand, and I.R. Mann, The effect of ULF compressional modes and field line resonances on relativistic electron dynamics, Adv. Space Res., in press, 2006. Dungey, J.W., Hydromagnetic waves, in Physics of Geomagnetic Phenomena, vol. 2, edited by S. Matsushita, and W.H. Campbell, chap. V-1, p. 913, Academic Press, New York, 1967. Elkington, S.R., M.K. Hudson, and A.A. Chan, Acceleration of relativistic electrons via drift-resonant interaction with toroidal-mode Pc-5 ULF oscillations, Geophys. Res. Lett., 26 (21), 3273, 1999. Elkington, S.R., M.K. Hudson, M.J. Wiltberger, and J.G. Lyon, MHD/Particle simulations of radiation belt dynamics, J. Atmos. Solar Terr. Phys., 64, 607, 2002. Elkington, S.R., M.K. Hudson, and A.A. Chan, Resonant acceleration and diffusion of outer zone electrons in an asymmetric geomagnetic field, J. Geophys. Res., 108 (A3), 1116, doi:10.1029/2001JA009202, 2003. Elkington, S.R., M. Wiltberger, A.A. Chan, and D.N. Baker, Physical models of the geospace radiation environment, J. Atmos. Solar Terr. Phys., 66, 1371, doi:10.1016/ j.jastp.2004.03.023, 2004. Fälthammar, C.-G., Effects of time-dependent electric fields on geomagnetically trapped radiation, J. Geophys. Res., 70 (11), 2503, 1965. Fälthammar, C.-G., Radial diffusion by violation of the third adiabatic invariant, in Earth’s Particles and Fields, edited by B.M. McCormac, p. 157, NATO Advanced Study Institute, Reinhold Book Corp., New York, 1968. Fei, Y., A.A. Chan, S.R. Elkington, and M.J. Wiltberger, Radial diffusion simulation of relativistic electron transport by ULF waves in the September 1998 storm, J. Geophys. Res., (in review), 2006. Friedel, R.H.W., G.D. Reeves, and T. Obara, Relativistic electron dynamics in the inner magnetosphere – a review, J. Atmos. Solar Terr. Phys., 64 (2), 265, 2002. Green, J.C., and M.G. Kivelson, A tale of two theories: How the adiabatic response and ulf waves affect relativistic electrons, J. Geophys. Res., 106, 25,777, 2001. Green, J.C., and M.G. Kivelson, Relativistic electrons in the outer radiation belt: Differentiating between acceleration mechanisms, J. Geophys. Res., 109 (A3), A03213, doi:10.1029/2003JA010153, 2004. Greenstadt, E.W., R.L. McPherron, and K. Takahashi, Solar wind control of daytime, midperoid geomangetic pulsations, in ULF Pulsations in the Magnetosphere, vol. 89, edited by D.J. Southwood, Center for Academic Publications Japan, 1981. Hasegawa, Drift mirror instability in the magnetosphere, Phys. Fluids, 12, 2642, 1969. Holzworth, R.H., and F.S. Mozer, Direct evaluation of the radial diffusion coefficient near L = 6 due to electric field fluctuations, J. Geophys. Res., 84 (A6), 2559, 1979. Horne, R.B., and R.M. Thorne, Potential waves for relativistic electron scattering and stochastic acceleration during magnetic storms, Geophys. Res. Lett., 25 (15), 1231, doi:10.1029/98GL01002, 1998. Hudson, M.K., S.R. Elkington, J.G. Lyon, C.C. Goodrich, and T.J. Rosenberg, Simulation of radiation belt dynamics driven by solar wind variations, in Sun-Earth Plasma Connections, vol. 109, edited by J.L. Burch, R.L. Carovillano, and S.K. Antiochos, p. 171, AGU, Washington, D.C., 1999. Hudson, M.K., S.R. Elkington, J.G. Lyon, and C.C. Goodrich, Increase in relativistic electron flux in the inner magnetosphere: ULF wave mode structure, Adv. Space Res., 25 (12), 2327, 2000. Hudson, M.K., S.R. Elkington, J.G. Lyon, M.J. Wiltberger, and M. Lessard, Radiation belt electron acceleration by ULF wave drift resonance: Simulation of 1997 and 1998 storms, in Space Weather, vol. 125, edited by P. Song, H.J. Singer, and G.L. Siscoe, p. 289, AGU, Washington, D.C., 2001. Hughes, W.J., Magnetospheric ULF waves: A tutorial with a historical prespective, in Solar wind sources of magnetospheric Ultra-Low-Frequency waves, Geophysical Monographs, vol. 81, edited by M. Engebretson, K. Takahashi, and M. Scholer, pp. 1-11, AGU, Washington, D.C., 1994. Jacobs, J.A., Y. Kato, S. Matsushita, and V.A. Troitskaya, Classification of geomagnetic micropulsations, J. Geophys. Res., 69, 180, 1964. Kepko, L., and H.E. Spence, Observations of discrete, global magnetospheric oscillations directly driven by solar wind density variations, J. Geophys. Res., 108 (A6), 211, doi:10.1029/2002JA009676, 2003. Kepko, L., H.E. Spence, and H.J. Singer, ULF waves in the solar wind as direct drivers of magnetospheric pulsations, Geophys. Res. Lett., 29 (8), 39-1, doi:10.1029/ 2001GL014405, 2002. Kim, H.-J., and A.A. Chan, Fully adiabatic changes in storm-time relativistic electron fluxes, J. Geophys. Res., 102, 22,107, 1997. Kivelson, M.G., and D.J. Southwood, Hydromagnetic waves in the ionosphere, Geophys. Res. Lett., 15, 1271, 1988. GM01069_Ch12.qxd 16 11/9/06 8:52 PM Page 16 A REVIEW OF ULF/RADIATION BELT INTERACTIONS Lanzerotti, L.J., and C.G. Morgan, ULF geomagnetic power near L = 4 − 2. Temporal variation of the radial diffusion coefficient for relativistic electrons, J. Geophys. Res., 78 (22), 4600, 1973. Lanzerotti, L.J., C.G. MacLennan, and M. Schulz, Radial diffusion of outer-zone electrons: An empirical approach to third-invariant violation, J. Geophys. Res., 75 (28), 5351, 1970. Li, X., D.N. Baker, M. Temerin, D. Larson, R.P. Lin, G.D. Reeves, M. Looper, S.G. Kanekal, and R.A. Mewaldt, Are energetic electrons in the solar wind the source of the outer radiation belt?, Geophys. Res. Lett., 24, 923, 1997. Li, X., et al., Energetic electron injections into the inner magnetosphere during the January 10-11, 1997 magnetic cloud event, Geophys. Res. Lett., 25, 2561, doi:10.1029/98GL00036, 1998. Liu, W.W., G. Rostoker, and D.N. Baker, Internal acceleration of relativistic electrons by large-amplitude ULF pulsations, J. Geophys. Res., 104 (A8), 17,391, 1999. Loto’aniu, T.M., I.R. Mann, L.G. Ozeke, A.A. Chan, Z.C. Dent, and D.K. Milling, Radial diffusion of relativistic electrons into the radiation belt slot region during the 2003 Halloween geomagnetic storms, J. Geophys. Res., in press, 2006. Lysak, R.L., and D. Lee, Response of the dipole magnetosphere to pressure pulses, Geophys. Res. Lett., 19, 937, 1992. Mann, I.R., A.N. Wright, and P.S. Cally, Coupling of magnetospheric cavity modes to field line resonances: A study of resonance widths, J. Geophys. Res., 100, 19,441, 1995. Mann, I.R., A.N. Wright, K.J. Mills, and V.M. Nakariakov, Excitation of magnetospheric waveguide modes by magnetosheath flows, J. Geophys. Res., 104 (A1), 333, 1999. Mann, I.R., T.P. O’Brien, and D.K. Milling, Correlations between ULF wave power, solar wind speed, and relativistic electron flux in the magnetosphere: solar cycle dependence, J. Atmos. Solar Terr. Phys., 66, 187, doi:10.1016/j.jastp.2003.10.002, 2004. Mathie, R.A., and I.R. Mann, A correlation between extended intervals of ULF wave power and storm-time geosynchronous relativistic electron flux enhancements, Geophys. Res. Lett., 27 (20), 3261, 2000a. Mathie, R.A., and I.R. Mann, Observations of Pc5 field line resonance azimuthal phase speeds: A diagnostic of their excitation mechanism, J. Geophys. Res., 105 (A5), 10,713, 2000b. Miura, A., Kelvin-helmholtz instability at the magnetospheric boundary: Dependence on the magnetosheath sonic mach number, J. Geophys. Res., 97, 10,655, 1992. Nakamura, R., J.B. Blake, S.R. Elkington, D.N. Baker, W. Baumjohann, and B. Klecker, Relationships between ULF waves and radiation belt electrons during the March 10, 1998 storm, Adv. Space Res., 30 (10), 2163, 2002. Newkirk, L.L., and M. Walt, Radial diffusion coefficient for electrons at low L values, J. Geophys. Res., 73 (3), 1013, 1968a. Newkirk, L.L., and M. Walt, Radial diffusion coefficient for electrons at 1.76 < L < 5, J. Geophys. Res., 73 (23), 7231, 1968b. Northrop, T.G., The Adiabatic Motion of Charged Particles, 109 pp., Interscience Publishers, New York, 1963. O’Brien, T.P., R.L. McPherron, D. Sornette, G.D. Reeves, R. Friedel, and H.J. Singer, Which magnetic storms produce relativistic electrons at geosynchronous orbit?, J. Geophys. Res., 106 (A8), 15,533, 2001. O’Brien, T.P., K.R. Lorentzen, I.R. Mann, N.P. Meredith, J.B. Blake, J.F. Fennel, M.D. Looper, D.K. Milling, and R.R. Anderson, Energization of relativistic electrons in the presence of ULF power and MeV microbursts: Evidence for dual ULF and VLF acceleration, J. Geophys. Res., 106 (A8), 15,533, 2003. Onsager, T.G., A.A. Chan, Y. Fei, S.R. Elkington, J.C. Green, and H.J. Singer, The radial gradient of relativistic electrons at geosynchronous orbit, J. Geophys. Res., 109 (A5), A05221, doi:10.1029/2003JA010368, 2004. Ozeke, L., and I.R. Mann, Modeling the properties of guided poloidal Alfvén waves with finite asymmetric ionospheric conductivities in a dipole field, J. Geophys. Res., 109, A5205, 2004. Paulikas, G.A., and J.B. Blake, Effects of the solar wind on magnetospheric dynamics: Energetic electrons at the synchronous orbit, in Quantitative Modeling of Magnetospheric Processes, vol. 21, edited by W.P. Olsen, p. 180, AGU, Washington, D.C., 1979. Perry, K.L., M.K. Hudson, and S.R. Elkington, Incorporating spectral characteristics of Pc5 waves into three-dimensional radiation belt modeling and the diffusion of relativistic electrons, J. Geophys. Res., 110 (A3), A03,215, doi:10.1029/2004JA010760, 2005. Reeves, G.D., K.L. McAdams, R.H.W. Friedel, and T.P. O’Brien, Acceleration and loss of relativistic electrons during geomagnetic storms, Geophys. Res. Lett., 30 (10), 36-1, doi:10.1029/2002GL016513, 2003. Ridley, A.J., G. Lu, C.R. Clauer, and V.O. Papitashvili, Ionospheric convection during nonsteady interplanetary magnetic field conditions, J. Geophys. Res., 102, 14,563, 1997. Ridley, A.J., G. Lu, C.R. Clauer, and V.O. Papitashvili, A statistical study of the ionospheric convection response to changing interplanetary magnetic field conditions using the assimilative mapping of ionospheric electrodynamics technique, J. Geophys. Res., 103, 4023, 1998. Riley, P., and R.A. Wolf, Comparison of diffusion and particle drift descriptions of radial transport in the Earth’s inner magnetosphere, J. Geophys. Res., 97 (A11), 16,865, 1992. Roederer, J.G., Dynamics of Geomagnetically Trapped Radiation, 166 pp., SpringerVerlag, New York, 1970. Rostoker, G., S. Skone, and D.N. Baker, On the origin of relativistic electrons in the magnetosphere associated with some geomagnetic storms, Geophys. Res. Lett., 25, 3701, 1998. Ruohoniemi, J.M., and R.A. Greenwald, The response of high-latitude convection to a sudden southward IMF turning, Geophys. Res. Lett., 25, 2913, 1998. Schulz, M., Canonical coordinates for radiation-belt modeling, in Radiation Belts: Models and Standards, vol. 97, edited by J.F. Lemaire, D. Heynderickx, and D.N. Baker, p. 153, AGU, Washington, D.C., 1996. Schulz, M., and A. Eviatar, Diffusion of equatorial particles in the outer radiation zone, J. Geophys. Res., 74 (9), 2182, 1969. Schulz, M., and L.J. Lanzerotti, Particle Diffusion in the Radiation Belts, Physics and Chemistry in Space, vol. 7, 215 pp., Springer-Verlag, New York, 1974. Selesnick, R.S., J.B. Blake, W.A. Kolasinski, and T.A. Fritz, A quiescent state of 3 to 8 MeV radiation belt electrons, Geophys. Res. Lett., 24 (11), 1343, 1997. Shprits, Y.Y., and R.M. Thorne, Time dependent radial diffusion modeling of relativistic electrons with realistic loss rates, Geophys. Res. Lett., 31, L08,805, doi:10.1029/ 2004GL019591, 2004. Sibeck, D.G., W. Baumjohann, R.C. Elphic, D.H. Fairfield, and J.F. Fennell, The magnetospheric response to 8-minute period strong-amplitude upstream pressure variations, J. Geophys. Res., 94, 2505, 1989a. Sibeck, D.G., R.E. Lopez, and W. Baumjohann, Solar wind dynamic pressure variations and transient magnetospheric signatures, Geophys. Res. Lett., 16, 13, 1989b. Southwood, D.J., J.W. Dungey, and R.J. Etherington, Bounce resonant interaction between pulsations and trapped particles, Planet. Space Sci., 17, 349, 1969. Summers, D., and C.-Y. Ma, Rapid acceleration of electrons in the magnetosphere by fast-mode MHD waves, J. Geophys. Res., 105 (A7), 15,887, 2000. Summers, D., R.M. Thorne, and F.L. Xiao, Relativistic theory of wave-particle resonant diffusion with application to electron acceleration in the magnetosphere, J. Geophys. Res., 103 (A9), 20,487, 1998. Takahashi, K., and R.L. McPherron, Standing hydromagnetic oscillations in the magnetosphere, Planet. Space Sci., 32, 1343, 1984. Takahashi, K., P.R. Higbie, and D.N. Baker, Azimuthal propagation and frequency characteristics of compressional Pc 5 waves, J. Geophys. Res., 90, 1473, 1985a. Takahashi, K., C.T. Russell, and R.R. Anderson, ISEE 1 and 2 observations of the spatial structure of a compressional Pc 5 wave, Geophys. Res. Lett., 12, 613, 1985b. Tan, L.C., S.F. Fung, and X. Shao, Observation of magnetospheric relativistic electrons accelerated by Pc-5 ULF waves, Geophys. Res. Lett., 31 (14), L14,802, 2004. Taylor, M.G. G.T., R.H.W. Friedel, G.D. Reeves, M.W. Dunlop, T.A. Fritz, P.W. Daly, and A. Balogh, Multisatellite measurements of electron phase space density gradients in the Earth’s inner and outer magnetosphere, J. Geophys. Res., 109, A05,220, doi:10.1029/2003JA010294, 2004. Tomassian, A.D., T.A. Farley, and A.L. Vampola, Inner-zone energetic electron repopulation by radial diffusion, J. Geophys. Res., 77 (19), 3441, 1972. Ukhorskiy, A.Y., K. Takahashi, B.J. Anderson, and H. Korth, The impact of inner magnetospheric ULF waves on the outer radiation belt electrons, J. Geophys. Res., 110, A10,202, doi:10.1029/2005JA011017, 2005. Ukhorskiy, A.Y., B.J. Anderson, and K. Takahashi, The impact of ULF oscillations in the solar wind dynamic pressure on the outer radiation belt electrons, Geophys. Res. Lett., 33, L06,111, doi:10.1029/2005GL024380, 2006. Walt, M., Introduction to Geomagnetically Trapped Radiation, Cambridge atmospheric and space science series, Cambridge University Press, Cambridge, 1994. Wrenn, G.L., Conclusive evidence for internal dielectric charging anomalies on geosynchronous communications spacecraft, J. Spacecr. Rockets, 32, 514, 1995. Wright, A.N., and G.J. Rickard, ULF pulsations driven by magnetopause motions: Azimuthal phase characteristics, J. Geophys. Res., 100, 23,703, 1995. S.R. Elkington, Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder, Colorado 80303, USA. ([email protected])