Margaret W. Chen

Simulations of phase space distributions of storm time proton ring current

We use results of guiding-center simulations of ion transport to map phase space densities of the stormtime proton ring current. We model a storm as a sequence of substorm-associated enhancements in the convection electric field. Our pre-storm phase space distribution is an analytical solution to a steady-state transport model in which quiet-time radial diffusion balances charge exchange. This pre-storm phase space spectra at L∼2-4 reproduce many of the features found in observed quiet-time spectra. Using results from simulations of ion transport during model storms having main phases of 3, 6, and 12 hr, we map phase space distributions from the pre-storm distribution in accordance with Liouvilles theorem. We find stormtime enhancements in the phase space densities at energies E∼30-160 keV for L∼2.5-4. These enhancements agree well with the observed stormtime ring current. For storms with shorter main phases (∼3 hr), the enhancements are caused mainly by the trapping of ions injected from open night side trajectories, and diffusive transport of higher-energy (≳ 160 keV) ions contributes little to the stormtime ring current. However, the stormtime ring current is augmented also by the diffusive transport of higher-energy ions (E≳ 160 keV) during storms having longer main phases (≳ 6 hr). In order to account for the increase in Dst associated with the formation of the stormtime ring current, we estimate the enhancement in particle-energy content that results from stormtime ion transport in the equatorial magnetosphere. We find that transport alone cannot account for the entire increase in |Dst| typical of a major storm. However, we can account for the entire increase in |Dst| by realistically increasing the stormtime outer boundary value of the phase space density relative to the quiet-time value. We compute the magnetic field produced by the ring current itself and find that radial profiles of the magnetic field depression resemble those obtained from observational data.

Stormtime transport of ring current and radiation belt ions

This is an investigation of stormtime particle transport that leads to formation of the ring current. Our method is to trace the guiding-center motion of representative ions (having selected first adiabatic invariants µ) in response to model substorm-associated impulses in the convection electric field. We compare our simulation results qualitatively with existing analytically tractable idealizations of particle transport (direct convective access and radial diffusion) in order to assess the limits of validity of these approximations. For µ ≲ 10 MeV/G (E ≲ 110 keV at L ≈ 3) the ion drift period on the final (ring-current) drift shell of interest (L ≈ 3) exceeds the duration of the main phase of our model storm, and we find that the transport of ions to this drift shell is appropriately idealized as direct convective access, typically from open drift paths. Ion transport to a final closed drift path from an open (plasma-sheet) drift trajectory is possible for those portions of that drift path that lie outside the mean stormtime separatrix between closed and open drift trajectories. For µ ∼ 10-25 MeV/G (110 keV ≲ E ≲ 280 keV at L ≈ 3) the drift period at L ≈ 3 is comparable to the postulated 3-hr duration of the storm, and the mode of transport is transitional between direct convective access and transport that resembles radial diffusion. (This particle population is transitional between the ring current and radiation belt). For µ ≳ 25 MeV/G (radiation-belt ions having E ≳ 280 keV at L ≈ 3) the ion drift period is considerably shorter than the main phase of a typical storm, and ions gain access to the ring-current region essentially via radial diffusion. By computing the mean and mean-square cumulative changes in 1/L among (in this case) 12 representative ions equally spaced in drift time around the steady-state drift shell of interest (L ≈ 3), we have estimated (from both our forward and our time-reversed simulations) the time-integrated radial-diffusion coefficients DLLsim for particles having selected values of µ ≳ 15 MeV/G. The results agree surprisingly well with the predictions (DLLql) of quasilinear radial-diffusion theory, despite the rather brief duration (≈ 3 hr) of our model storm and despite the extreme variability (with frequency) of the spectral-density function that characterizes the applied electric field during our model storm. As expected, the values of DLLsim deduced (respectively) from our forward and time-reversed simulations agree even better with each other and with DLLsim when the impulse amplitudes which characterize the individual substorms of our model storm are systematically reduced.

Simulations of diffuse aurora with plasma sheet electrons in pitch angle diffusion less than everywhere strong

To investigate the spatial and spectral structure of the diffuse aurora during a model geomagnetic storm characterized by random impulses in the cross-magnetospheric convection electric field, we simulate the bounce-averaged drift motion and precipitation of plasma sheet electrons. Bounce-averaged drift trajectories are computed from a Hamiltonian formulation in which we have treated the plasma sheet electrons as though they were undergoing strong pitch angle diffusion in Dungeys model magnetosphere (dipole field plus uniform southward Bz). Using the simulation results, we map phase space densities from near the nightside neutral line according to Liouvilles theorem, modified to account for particle losses consistent with the postulated pitch angle scattering. We consider three different idealized scattering rate models for the plasma sheet electrons: (1) strong diffusion everywhere, (2) an MLT-independent model for diffusion that is less than everywhere strong, and (3) an MLT-dependent scattering rate model with the same azimuthal average as the MLT-independent model. We evaluate the precipitating energy flux and mean energy at ionospheric altitude h = 127.4 km, as well as the ionospheric Hall and Pedersen conductances, for the three different scattering rate models considered, and we compare the results with each other and with available observations. The limit of everywhere strong pitch angle diffusion yields a simulated diffuse aurora that seems too intense at its maximum (storm time energy flux of ∼5–10 erg cm−2 s−1) and seems too concentrated near midnight in magnetic local time (MLT). The distribution of energy flux above half maximum extends from ∼2000–0700 MLT throughout the model storm. Corresponding maxima in simulated Hall (∼18–34 mhos) and Pedersen (∼9–11 mhos) conductances also seem too large. Our model for pitch angle diffusion that is less than everywhere strong yields a more realistic maximum energy flux (∼2 erg cm−2 s−1) and yields a realistically broader distribution in MLT, essentially because the plasma sheet electrons of interest live longer before precipitating. An even more realistic MLT distribution of precipitating energy flux is obtained by modulating the latter scattering rate model with an MLT dependence based on the observed statistical distribution of waves with respect to MLT. With such a model we can account for intensifications of diffuse auroral electron precipitation found near dawn and late in the morning quadrant in both statistical and storm event studies.

CRRES Observations of the Composition of the Ring-Current Ion Populations.

Abstract : The Magnetospheric Ion Composition Spectrometer onboard the CRRES spacecraft provided mass and charge state composition data for positive ions in the energy-per-charge range 10-425 keV/e. The CRRES data is compared to the AMPTE/CCE observations during both geomagnetically quiet and active periods. The CRRES average radial profiles of H+, He+, and He++ during quiet intervals are remarkably similar to those measured by CCE. The excess of ions measured by CRRES at L <4 compared to standard ion transport models tends to support the necessity of additional ion radial diffusion by ionospheric electric-field variations. A summary is also given of the measured storm-time variations of the major ion populations during the large storm of March 1991. The results are compared to previous observations by the AMPTE/CCE spacecraft during a large storm. The CRRES data confirm that the rapid initial recovery of the Dst magnetic index is due to a momentary change of the relative ion composition of the ring current to an oxygen-dominated state. A preliminary test of the Dessler-Parker-Sckopke relation between the ion energy and the global magnetic perturbation shows that the observed particle fluxes during the March 1991 storm could account for only 30-50% of the variation of the Dst magnetic index. (AN)

Modeling the quiet-time inner plasma sheet protons

In order to understand the characteristics of the quiet time inner plasma sheet protons, we use a modified version of the Magnetospheric Specification Model to simulate the bounce averaged electric and magnetic drift of isotropic plasma sheet protons in an approximately self-consistent magnetic field. Proton differential fluxes are assigned to the model boundary to mimic a mixed tail source consisting of hot plasma from the distant tail and cooler plasma from the low latitude boundary layer (LLBL). The source is local time dependent and is based on Geotail observations and the results of the finite tail width convection model. For the purpose of self-consistently simulating plasma motion and a magnetic field, the Tsyganenko 96 magnetic field model is incorporated with additional adjustable ring-current shaped current loops. We obtain equatorial proton flow and midnight and equatorial profiles of proton pressure, number density, and temperature. We find that our results agree well with observations. This indicates that the drift motion dominates the plasma transport in the quiet time inner plasma sheet. Our simulations show that cold plasma from the LLBL enhances the number density and the proton pressure in the inner plasma sheet and decreases the dawn-dusk asymmetry of the equatorial proton pressure. From our approximately force-balanced simulations the magnetic field responds to the increase of pressure gradient force in the inner plasma sheet by changing its configuration to give a stronger magnetic force. At the same time, the plasma dynamics is affected by the changing field configuration and its associated pressure gradient force becomes smaller. Our model predicts a quiet time magnetic field configuration with a local depression in the equatorial magnetic field strength at the inner edge of the plasma sheet and a cross-tail current separated from the ring current, results that are supported by observations. A scale analysis of our results shows that in the inner plasma sheet the magnitude of the Hall term in the generalized Ohms law is not small compared with the quiet time electric field. This suggests that the frozen-in condition E = −v×B is not valid in the inner plasma sheet and that the Hall term needs to be included to obtain an appropriate approximation of the generalized Ohms law in that region.

Simulations of storm time diffuse aurora with plasmasheet electrons in strong pitch angle diffusion

Using a guiding-center simulation of plasmasheet electrons postulated to be in strong pitch angle diffusion, we compute drift trajectories from a Hamiltonian formulation and map phase space densities from the nightside neutral line in Dungeys model magnetosphere (dipole field plus uniform southward Bz) according to Liouvilles theorem modified for exponential loss implicit in the strong diffusion hypothesis. From the resulting phase space distributions we compute the precipitating energy flux into the auroral ionosphere as functions of magnetic latitude and magnetic local time (MLT) so as to simulate numerically the spatial and spectral structure of diffuse auroral electron precipitation for comparison with observational data. Storm-associated impulses (by which we enhance the convection electric field) can typically transport plasmasheet electrons from the nightside neutral line to the near-midnight region of maximum precipitating energy flux (latitude ≈ 65°) in ∼20–30 min, which is roughly the strong diffusion lifetime of 4-keV electrons at the corresponding L value (≈5.7). The maximum precipitating electron energy flux in our simulation of the model storm is thus modulated by random variations in the mean cross-magnetospheric electric potential drop over the 20–30 min before the time of interest. Our results also show a consistent lack of precipitating electron energy flux in the afternoon quadrant, essentially because this is the last quadrant to be visited by plasmasheet electrons (and therefore features the most strongly attenuated phase space densities) as they drift through the magnetosphere on open trajectories. The result agrees qualitatively with the typically observed “darkness” of X-ray images of the diffuse aurora in that sector (1200–1800 MLT). While our simulation results locate the region of maximum energy flux slightly postmidnight during both prestorm and stormtime, in good agreement with previously published statistical compilations of auroral electron precipitation. Our simulations do not explain other large intensifications of electron precipitation (near dawn and in the morning sector) that are also found in both statistical and storm event studies. This suggests the need for a model in which pitch angle diffusion is less than everywhere strong throughout the plasma sheet.

Proton ring current pitch angle distributions: Comparison of simulations with CRRES observations

In this study we compare the proton pitch angle distributions (PADs) in the ring current region (L ∼ 3–4) obtained from Combined Release and Radiation Effects Satellite (CRRES) observations during the large magnetic storm (minimum Dst = −170 nT) on August 19, 1991, with results of phase-space mapping simulations in which we trace the bounce-averaged drift of protons during storm-associated enhancements in a model of the convection electric field. We map the phase-space density ƒ according to Liouvilles theorem except for attenuation by charge exchange, which we compute for both an empirical model [Rairden et al., 1986] and a theory-based model [Hodges, 1994] of the neutral H density distribution. We compare simulated pitch angle distributions at 48 keV, 81 keV, and 140 keV at L = 3 and L = 4 directly with the CRRES distributions at the same energies and L values before and during the storm. A steady-state application of our transport model, using the empirical neutral H density model of Rairden et al. [1986], reproduces the absolute intensities well except for E = 140 keV at L = 3 (M = 10 MeV/G) and E = 48 keV at L = 4 (M = 13 MeV/G). The anisotropies (A ∼ 0.2–0.8) of the CRRES and modeled pre-storm pitch angle distributions agree within factors ≲ 2. Time-dependent application of our transport model reproduces measured recovery phase anisotropies (t = 10–12 h after storm onset; A ∼ 0.4–1.2) similarly well at the selected energies and L values, but agreement between modeled and measured absolute intensities is energy-dependent and not consistently good. Our model underpredicts the proton intensities found by CRRES for E > 80 keV at L = 4 in early recovery phase (t = 10–12 h). Perhaps the impulsive stormtime convection electric field was stronger during the main phase than we have assumed here. Comparisons were more difficult in late recovery phase (t = 20 h) because CRRES was too far off the magnetic equator. Proton life-times inferred from the CRRES data during the recovery phase of this storm are considerably shorter than charge-exchange lifetimes for either model, but the empirical neutral H density model of Rairden et al. [1986] leads to smaller discrepancies with the CRRES data at all the selected energies and L values than the theory-based neutral H density model of Hodges [1994] for parameters that most closely represent the seasonal and solar maximum conditions of the August 19, 1991, storm. It appears that charge exchange alone is not enough to explain the observed rapid decay of the ring current proton intensities during the recovery phase of this storm.

Stormtime ring‐current formation: A comparison between single‐and double‐dip model storms with similar transport characteristics

Intense magnetic storms often develop in two stages such that a second ring current enhancement begins before the first ring current enhancement has recovered to the prestorm level. Since Dst traces of such storms exhibit two dips, we refer to these as double-dip storms. Here we compare double- and single-dip storms with similar convective and diffusive transport characteristics for effectiveness at forming the proton ring current. Our model storms consist of superposition of almost randomly occurring impulses in the convection electric field. We have synthesized a hypothetical double-dip storm consisting of a moderate 6-hour storm, followed by a 3-hour quiet interval and then by a more intense 15-hour storm, for a total duration of 24 hours. For comparison, we consider a single-dip model storm with an unmodulated 24-hour main phase during which the root-mean-square enhancement of the cross-tail potential drop is made equal to the timeweighted rms enhancement for the double-dip model storm. This leads to comparable timeaveraged diffusion coefficients for our single- and double-dip model storms. The mean enhancement of the cross-tail potential drop of the two storms are also comparable. When the stormtime proton plasma sheet distribution, the source of ring current protons, is left unchanged from its quiet (prestorm) level, we find little difference in proton energy content per unit R (which is geocentric distance normalized by RE) between our double- and single-dip model storms. The protonenergy content of the magnetosphere is roughly increased by a factor of 2.5 by either model storm under this scenario, in which the overall amount of stormtime transport, whether convective or diffusive, is nearly the same for the double- and single-dip model storms. As in our earlier work, we require here an enhanced stormtime plasma sheet population (in addition to enhanced particle transport) in order to achieve (for example) the 20-fold increase in |Dst| characteristic of a large storm. Only when we invoke a two-stage stormtime enhancement of the boundary (plasma sheet) phase space density in combination with the two-stage enhancement in particle transport, our double-dip model storm does show a much larger total energy content than our single-dip model storm with a one-stage enhancement of the boundary spectrum. This suggests that plasma sheet preconditioning may be important for the development of especially intense storms.

Observations of iron, silicon, and other heavy ions in the geostationary altitude region during late March 1991

Following the great sudden storm commencement of March 24, 1991, observations were made aboard CRRES of Fe, Mg, Si, and other heavy ions in the energy range of several tens to a few hundred keV/n. The charge states of these heavy ions were observed to change abruptly. For example, the dominant Fe charge state changed at one point from q = 9 + to q = 16 + . We suggest that ions from the solar corona, retaining the charge state frozen in near the Sun, were convected inward to the location of CRRES from the plasma sheet within minutes. The ions were energized in the convection process, and the ionic charge state was unaltered by charge exchange. The very large magnetospheric electric fields associated with the highly disturbed geomagnetic conditions were required to enable the ions to reach CRRES. Guiding center simulations were carried out which support this hypothesis. At least two ion populations were observed, corresponding to temperatures at two different coronal locations, and the observed changes in heavy-ion charge state were caused by a sequence of solar wind parcels sweeping over the Earths magnetosphere. The signature of a change from low to high temperature seems to fit well the suggestions of Tsurutani and Gonzales [1994], of a fast solar wind stream tamped by a slower one, as a generator of great storms.

Global storm time auroral X-ray morphology and timing and comparison with UV measurements

The Polar Ionospheric X-ray Imaging Experiment (PIXIE) on NASAs Polar spacecraft provides the first global images of the auroral oval in X-rays and allows very accurate measurements of the timing of geomagnetic disturbances to a degree of temporal resolution not available from previous imagers due to its photon counting characteristics. On October 19, 1998, a magnetic cloud associated with a CME encountered the Earths magnetopause near 0500 UT, generating a magnetic storm that reached a minimum value in Dst of -139 nT. The z component of the interplanetary magnetic field (IMF) (B z ) remained remarkably steady for the first 10 hours of the storm as did the solar wind particle pressure. The PIXIE and UVI instruments on the Polar spacecraft were both imaging the auroral oval from 0800 to 1800 UT; six distinct impulsive auroral enhancements were observed by the imagers during this time period. Global imaging combined with geosynchronous particle observations allowed classification of the geomagnetic disturbances associated with the events. Only two of the events were classified as substorms; one was classified as a poleward boundary intensification, one was a convection bay, and one was a pseudobreakup. A sixth event occurred after a dramatic northward turning of the IMF at the end of the 10-hour B z south period but was very weak and transient. The effects of the northward turning were counteracted by a simultaneous increase in the B y component of the IMF. The first sign of significant substorm activity occurred over 8 hours after the cloud encountered the Earth and was not associated with any change in the solar wind magnetic field or particle pressure. The cross polar cap potential remained large (> 100 kV), and most of the X-ray emissions observed were associated with enhanced earthward convection caused by large cross-tail electric fields; 50% were collected from the 0000 - 0600 magnetic local time (MLT) sector.