We report the probable detection of energetic neutral hydrogen atoms (ENAs) at >0.8 MeV in several large solar energetic particle events observed between 1997 and 2004. The low Earth orbiting SAMPEX satellite detected transient increases of quasi-trapped equatorial protons beginning typically ∼3 hr after the X-ray flare and lasting for up to several hours. Since the magnetic cutoff rigidity is >10 GV at the magnetic latitude where the particles were observed, we interpret the signal as due to ENAs that penetrate Earth’s magnetic field and charge exchange in the upper atmosphere, whereupon the charged particles may become trapped. One event outside our survey period (2006 December 5) had previously reported solar flare ENAs, the only example of this phenomenon of which we are aware. Although the statistics are limited, the events we report suggest that the ENAs are produced as the flare-associated coronal mass wjection moves through the corona, as concluded previously for the 2006 December 5 event. The finding of ENAs emitted in conjunction with large solar flares opens a new avenue to understanding these events.

Close to the peak of a magnetic storm, the ratio of the O+ to the H+ contribution to the ring current energy, U(O+)/U(H+), depends on both solar EUV flux and storm size. This ratio provides a lower limit on the importance of the ionospheric plasma source compared to the solar wind source since only the ionosphere supplies O+ while both the solar wind and the ionosphere supply H+. We have used measurements from the CHEM ion spectrometer on the Active Magnetospheric Particle Tracer Explorers (AMPTE) CCE spacecraft to assess U(O+)/U(H+) near storm maximum. CHEM measured equatorial ion composition over the energy per charge range 1.5–300 keV/e for more than 4 years, from solar minimum through much of the rising phase of Solar Cycle 22. Thus our data set is large, including 67 magnetic storms with minimum Dst values less than −50 nT. To estimate the O+ and H+ contributions to the ring current energy, we have multiplied local measurements of each ion's energy density along the CCE trajectory by the appropriate dipole L shell volume and summed over L values from two to seven. We have used a bilinear regression to assess the dependence of U(O+)/U(H+) on solar EUV flux, parameterized by F10.7, and storm size, parameterized by minimum Dst value Dstmin. We have found that both F10.7 and Dstmin are important (and nearly independent) predictors of U(O+)/U(H+). Only four storms of 67 had U(O+)/U(H+) > 1. Two of the four were fairly small, with |Dstmin| < 100 nT but with high F10.7 values. This shows that O+ can sometimes contribute the majority of the ring current energy during small storms when the solar EUV flux is high (i.e., near solar maximum) as well as during very large storms throughout the solar cycle.

We report the results of a study of the contributions of the largescale magnetospheric currents to the observed Dst variation. Ground‐based magnetometer data during four magnetic storms (January 27–30, 1985; November 23–27, 1986; January 14–16, 1988; and May 6–8, 1988) were used to calculate Dst, and the paraboloid model of the magnetospheric magnetic field [Alexeev et al., 1996] was used to determine the contribution of each magnetospheric current system. Input data for our model were the solar wind plasma parameters, the interplanetary magnetic field (IMF) Bz, DMSP F6, F7, F8, and F9 satellite observations of precipitating auroral particles, and Active Magnetospheric Particle Tracer Explorers (AMPTE)/CCE satellite measurements of the total energy of the ring current ions with energy per charge between 1.5 and 300 keVq−1. We found good agreement between observed and modeled magnetic fields during the main phase of the magnetic storms. Using the paraboloid model, we have determined the contributions to Dst of different magnetospheric current systems including the magnetopause current BCF, the symmetric ring current BR, and the geotail current BT‐Such separation shows that values of BT and BCF are comparable with the value of BR during the main phase of the storms. During the recovery phase the effect of BR predominates.

The Dessler-Parker-Sckopke relation (DPS) predicts a linear dependence of the perturbation magnetic field at the surface of the Earth on the total ring current kinetic energy. In this paper, we test DPS by using measurements of the major ring current ion species made by the charge-energy-mass spectrometer on the Active Magnetospheric Particle Tracer Explorers CCE spacecraft. We use spectra from passes through the equatorial storm time ring current near the maximum phase of 80 magnetic storms between 1984 and 1989 to estimate the global ring current energy content E-RC and compare it with the average value for Dst during each pass. Our work shows that DPS holds well on average. In particular, there is a strong linear correlation between ring current energy estimated from nightside ion measurements and the Dst index, and the slope of the least squares fit line giving Dst as a function of nightside E-RC is in good agreement with the prediction of DPS. In contrast, dayside measurements of E-RC do not yield a robust correlation with Dst. Although we cannot rule out the possibility that currents other than the ring current (for example, tail currents and the magnetopause current) may cause large magnetic perturbations, we conclude that these perturbations, if they exist, must be largely compensating. By examining how the ratio of Dst to E-RC varies with the local time sector of the in situ ion measurements, we obtain statistical information on the anisotropy of the storm time ring current. We find that the largest values of E-RC/Dst result from nightside measurements and the smallest values result from measurements in the 0600 to 1200 LT region, as would be expected for an ion population injected on the nightside that must drift westward around the Earth, undergoing losses, to reach the dayside morning sector.

We have used the Low‐Energy Ion Composition Analyzer (LICA) instrument on the low altitude, polar orbiting SAMPEX spacecraft to survey energetic ions near the magnetic equator from late 1992 through 1998; that is, through the declining phase of Solar Cycle 22, solar minimum, and into the rise of Solar Cycle 23. This survey gives us a unique opportunity to examine both the long‐term variation in the low‐altitude equatorial ion population and short‐term enhancements that occur during magnetic storms. During the survey period, 40 storms with minimum Dst ≤ 100 nT occurred: the majority were accompanied by increases in the equatorial ion flux. Although LICA detects ions with energies far above the bulk of the ring current ion population, the times of the maximum equatorial fluxes clustered around the time of minimum Dst, i.e., the time of maximum ring current energy content. The storm associated flux maxima were unevenly distributed in geographic longitude, with the maximum flux enhancements occurring at longitudes just west of the South Atlantic Anomaly. Except for an increase in 1994, the quiet time monthly average equatorial flux declined steadily from 1992 until early 1998; then it began to rise again. The monthly average equatorial ion fluxes had a very significant correlation with the Ap index during this period (R = 0.54), indicating that geomagnetic activity dominated the long‐term variation. During the survey, we also discovered enhancements in the equatorial ion flux that occurred shortly after the onsets of three recent, large solar energetic particle events. These enhancements began well before the commencements of the associated geomagnetic storms. The major ion species present were H, He, C, and O, therefore ruling out an ionospheric source. These ions could not have penetrated directly from interplanetary space to the magnetic equator, and we do not understand the mechanism that produces the SEP‐associated enhancements.

Using instrumentation on board the SAMPEX satellite at ∼600 km we present new measurements of the composition of a population of >0.49 MeV/nucleon ions that are trapped near the geomagnetic equator. These ions gain access to the SAMPEX orbit as neutral particles via charge exchange collisions that continually occur within the higher altitude radiation belt and contribute to the decay of storm‐time enhanced ring current. At low altitudes a second charge exchange collision produces a trapped ion belt near the magnetic equator. This belt is ephemeral, as it is necessarily observed on drift paths that reach deep into the atmosphere near the South Atlantic Anomaly. We find the relative abundances of H:He:C:O:Ne‐Fe are 7000 : 7.5 : 0.14 : 1.0 : 0.03. The low altitude particles are underabundant in Z>8 ions compared to the radiation belt, possibly because the radiation belt Z>8 ions may not be stably trapped once they charge exchange to higher rigidities. We cannot rule out the possible contribution of trapped anomalous cosmic rays to the heavy ions observed above 0.49 MeV/nucleon in this low altitude population.

The SAMPEX satellite, an international collaboration with Germany, is the first in a series of small explorer missions that NASA began in 1989 to perform astrophysics and space physics investigations with small spacecraft launched on expendable launch vehicles. SAMPEX was launched from VAFB on a Scout rocket in July 1992, just 39 months after selection by NASA. Operating in an 82/spl deg/ inclination orbit with altitudes between 520 and 670 km, the 350 lb spacecraft has performed flawlessly since launch. The spacecraft bus was developed by the Small Explorer project at Goddard Space Flight Center. SAMPEX carries a payload of four scientific instruments that study particles originating at the Sun, in the so-called anomalous cosmic rays, and in the magnetosphere. The SAMPEX instruments have sensitivities >100 times larger than previous low Earth orbit spacecraft, that have led to new discoveries such as a new radiation belt of interstellar material and rare hydrogen and helium isotopes trapped in the radiation belts. SAMPEX provides routine global maps of the magnetosphere, and has given new insights into the processes by which radiation levels through the entire magnetosphere can become greatly enhanced, leading to operating spacecraft anomalies. The authors give an overview of the SAMPEX scientific goals, instrumentation, and mission development approach.

The development of the ring current ions in the inner magnetosphere during the main phase of a magnetic storm is studied. The temporal and spatial evolution of the ion phase space densities in a dipole field are calculated using a three dimensional ring current model, considering charge exchange and Coulomb losses along drift paths. The simulation starts with a quiet time distribution. The model is tested by comparing calculated ion fluxes with Active Magnetospheric Particle Tracer Explorers/CCE measurement during the storm main phase on May 2, 1986. Most of the calculated omnidirectional fluxes are in good agreement with the data except on the dayside inner edge (L < 2.5) of the ring current, where the ion fluxes are underestimated. The model also reproduces the measured pitch angle distributions of ions with energies below 10 keV. At higher energy, an additional diffusion in pitch angle is necessary in order to fit the data. The role of the induced electric field on the ring current dynamics is also examined by simulating a series of substorm activities represented by stretching and collapsing the magnetic field lines. In response to the impulsively changing fields, the calculated ion energy content fluctuates about a mean value that grows steadily with the enhanced quiescent field.

The retarding potential analyzer on the DMSP F8 satellite measured ion density, composition, temperature, and ram flow velocity at 840‐km altitude near the dawn and dusk meridians close to solar minimum. Nine days of data were selected for study to represent the summer and winter solstices and the autumnal equinox under quiet, moderately active, and disturbed geomagnetic conditions. The observations revealed extensive regions of light‐ion dominance along both the dawn and dusk legs of the DMSP F8 orbit. These regions showed seasonal, longitudinal, and geomagnetic control, with light ions commonly predominating in places where the subsatellite ionosphere was relatively cold. Field‐aligned plasma flows also were detected. In the morning, ions flowed toward the equator from both sides. In the evening, DMSP F8 detected flows that either diverged away from the equator or were directed toward the northern hemisphere. The effects of diurnal variations in plasma pressure gradients in the ionosphere and plasmasphere, momentum coupling between neutral winds and ions at the feet of field lines, and E × B drifts qualitatively explain most features of these composition and velocity measurements.

We study the ionospheric convection that prevails during the passage of a magnetic cloud past the Earth. For the cloud studied here, January 13‐15, 1988, the ionospheric convection was measured almost continually in cross sections of the polar cap and auroral zone by the Defense Meteorological Satellite Program (DMSP) polar‐orbiting satellite. In turn, the conditions in the magnetic cloud are like those of a controlled laboratory experiment: First, the magnetic field changes smoothly and slowly on time scales much longer than the expected ionospheric response time, ensuring that the external interplanetary conditions giving rise to any ionospheric flow pattern are known to unprecedented accuracy. Second, over the longer time scale of the magnetic cloud passage, the magnetic field vector rotates by over 180° such that the magnetic cloud divides into two intervals of northward (Bz > 0) and southward (Bz < 0) pointing interplanetary magnetic field (IMF) of 11 and 18 hours duration, respectively. During the former interval our observations show that (1) for strongly northward IMF (Bz > 18 nT) the convection in one (the southern) hemisphere is characterized by a two‐cell convection pattern confined to high latitudes (≥ 75°) with sunward flow over the pole (“reverse” two‐cell convection). (2) The strength of the flows (∼1 km/s) is comparable to that seen under southward IMF later. (3) Superimposed on this convection pattern there are clear dawn‐dusk asymmetries associated with a one‐cell convection component whose sense depends on the polarity of the magnetic cloud's large east‐west magnetic field component. (4) Whilst the flows in the southern hemisphere are ordered into a well‐defined convection pattern, the flows in the northern hemisphere are very irregular, varying on short spatial scales. When the cloud's magnetic field turns southward the following observations were made: (1) The convection is characterized by a two‐cell pattern extending to lower latitudes (∼50°) with antisunward flow over the pole (“standard” two‐cell convection). (2) There is no evident interhemisphere difference in the structure and strength of the convection. (3) Superimposed dawn‐dusk asymmetries in the flow pattern are observed which are only in part attributable to the east‐west component of the magnetic field. (4) A dawn‐dusk asymmetry in the latitude of the convection reversal boundary also exists, which is most pronounced in the northern hemisphere (up to 10° difference). We study the transition from a reverse to standard two‐cell convection pattern and find that because of the large By component of the magnetic field inside the magnetic cloud, this transition actually takes place before the external field turns southward. It occurs when the magnetic shear angle between the subsolar magnetospheric field and the magnetic field of the magnetic cloud is ∼70°. We consider the effect of the Bz component of the cloud magnetic field on the size of the open field line region or polar cap, inferred from the convection pattern. We argue that the long‐term variation in the polar cap size is determined by the temporal gradient of the Bz component. We investigate the strength of the ionospheric convection, parameterized by the maximum potential difference across the convection pattern, as a function of the Bz component of the cloud magnetic field, extending the range of similar investigations. For strongly northward IMF a potential difference of 60‐80 kV is observed across the reverse two‐cell convection pattern. For southward IMF the potential difference across the standard two‐cell convection pattern changes almost linearly with Bz, attaining its maximum value of ∼180 kV at the extreme value of Bz = −19 nT reached in this cloud.

Transient or patchy magnetic field line merging on the dayside magnetopause, giving rise to flux transfer events (FTEs), is thought to play a significant role in energizing high-latitude ionospheric convection during periods of southward interplanetary magnetic field. Several transient velocity patterns in the cusp ionosphere have been presented as candidate FTE signatures. Instrument limitations, combined with uncertainties about the magnetopause processes causing individual velocity transients, mean that definitive observations of the ionospheric signature of FTEs have yet to be presented. This paper describes combined observations by the PACE HF backscatter radar and the DMSP F9 polar-orbiting satellite of a transient velocity signature in the southern hemisphere ionospheric cusp. The prevailing solar wind conditions suggest that it is the result of enhanced magnetic merging at the magnetopause. The satellite particle precipitation data associated with the transient are typically cusplike in nature. The presence of spatially discrete patches of accelerated ions at the equatorward edge of the cusp is consistent with the ion acceleration that could occur with merging. The combined radar line-of-sight velocity data and the satellite transverse plasma drift data are consistent with a channel of enhanced convection superposed on the ambient cusp plasma flow. This channel is at least 900 km in longitudinal extent but only 100 km wide. It is zonally aligned for most of its extent, except at the western limit where it rotates sharply poleward. Weak return flow is observed outside the channel. These observations are compared with and contrasted to similar events seen by the EISCAT radar and by optical instruments.

We have modeled plasma transport in the low‐latitude and equatorial ionosphere during the great magnetic storm of March 1989. Our goal was to provide a consistent explanation for the DMSP (Defense Meteorological Satellite Program) observations of dramatic decreases in ion density and rapid ion drifts in the low latitude ionosphere over South America during the storm. The modeling effort supports the hypothesis that abnormally large upward drifts lifted F region plasma above the satellite's altitude and created the density depletions observed by DMSP. Modeled O+ densities at the satellite's altitude have a strong qualitative resemblance to DMSP observations. Both the model and the observations indicate a deep density trough with extremely sharp boundaries surrounding the equator. The widths of both the modeled and the observed equatorial troughs increase with time. Vertical ion drifts predicted by the model also have been compared with DMSP measurements. Like the observed vertical drifts, the modeled drifts reversed sign near the trough boundaries. The modeled vertical drifts are of the same order and direction as the vertical component of E × B convection near the equator, but of opposite direction (downward) near the trough boundaries and outside of the trough.

The direct injection of magnetosheath plasma into the cusp produces at low altitude a precipitation regime with an energy‐latitude dispersion—the more poleward portion of which we herein term the “cusp plume.” An extensive survey of the Defense Meteorological Satellite Program (DMSP) F7 and F9 32 eV to 30 keV precipitating particle data shows that similar dispersive signatures exist over much of the dayside, just poleward of the auroral oval. Away from noon (or more precisely, anywhere not immediately poleward of the cusp) the fluxes are reduced by a factor of about 10 as compared to the cusp plume, but other characteristics are quite similar. For example, the inferred temperatures and flow velocities, and the characteristic decline of energy and number flux with increasing latitude is essentially the same in a longitudinally broad ring of precipitation a few degrees thick in latitude over much of the dayside. We conclude that the field lines on which such precipitation occurs thread the magnetospheric plasma mantle over the entire longitudinally extended ring. Besides the location of occurrence (i.e., immediately poleward of the dayside oval), the identification is based especially on the associated very soft ion spectra, which have densities from a few times 10−2 to a few times 10−1/cm³; on the temperature range, which is from a few tens of eV up to about 200 eV; and on the characteristic gradients with latitude. Further corroborating evidence that the precipitation is associated with field lines which thread the plasma mantle includes drift meter observations which show that regions so identified based on the particle data consistently lie on antisunward convecting field lines. Our observations indicate that some dayside high‐latitude auroral features just poleward of the auroral oval are embedded in the plasma mantle.

The dayside zone of soft precipitation can be divided into four distinct types of plasma regimes, each corresponding to the respective magnetospheric source region: the cusp, the mantle, the low‐latitude boundary layer (LLBL), and the dayside extension of the BPS. Based on a detailed spectral study, including comparisons with nonsimultaneous ISEE 1 satellite LLBL data, we identify regions of LLBL‐type plasma in the DMSP data set and compare these plasma boundaries with convection reversal boundaries (CRBs) as determined by either Sondrestrom or the drift meter instrument on board the DMSP F9 spacecraft. The nine cases considered are all in the prenoon local time sector. We find that in eight of the nine cases the CRB occurs within the LLBL as expected, generally near to, but not coincident with, the equatorward edge of the LLBL‐type plasma. In our sample set, chosen for cases with latitudinally wide, easily identifiable LLBL signatures, the average latitudinal width was 1.85° magnetic latitude. The CRB, defined as the onset of steady antisunward convection, occurred about 30% of this width beyond the equatorward onset of LLBL‐type particles. The most equatorward portion of the region with LLBL‐type plasma usually had near‐zero or erratic convection and may correspond to the “stagnation region” reported from ISEE observations. The potential drop observed across the low‐altitude LLBL is roughly estimated to be typically ∼5 keV. A summary is given on how the various high‐altitude sources can be identified when plasma regions are observed at low altitude in the dayside auroral oval.

Early on March 14, 1989, a thermal plasma probe on the Defense Meteorological Satellite Program (DMSP) F9 spacecraft detected extensive and dramatic decreases in the ion density at 840 km, near 2130 LT, during two consecutive transequatorial passes over South America. The order of magnitude decreases in the ion density extended more than 4000 km along the satellite track. The depletions were accompanied by upward and westward plasma drifts, both in excess of 100 m/s. Their onsets and terminations were marked by extremely sharp density gradients. DMSP F9 observed no similar depletions over the Atlantic during preceding orbits. A partial depletion was detected over the eastern Pacific during the following orbit. The DMSP F9 ground track passed slightly west of a Brazilian total electron content (TEC) station and two Brazilian ionosondes during the first depletion encounter. The TEC fell far below normal during the night of March 13–14. The ionosonde measurements indicate that, in the hour after sunset, before DMSP passed through the depletions, the F2 layer rose rapidly and disappeared, but at the time of the first depletion encounter, hmF2 was decreasing over one of the stations. The DMSP F8 satellite, which orbits in the dawn‐dusk meridian, made related measurements on March 13 and 14. Crossing the equator at dust on March 13, at the same longitude where DMSP F9 encountered the first depletion, DMSP F8 detected upward and westward drifts, but it measured extremely large rather then depleted ion densities. During two dawn passes over the eastern Pacific on March 14, DMSP F8 observed depletions somewhat similar to those detected by DMSP F9. Large westward drifts accompanied the depletions detected by DMSP F8. It is quite probable that the morningside depletions detected on March 14 are remnants of those detected earlier by DMSP F9 in the evening sector. We develop a phenomenological model reconciling DMSP F8, F9, and ground‐based measurements. Our calculations show that rapid upward drifts sustained for several hours can produce depletions in the equatorial ion density with sharp gradients at their high‐latitude boundaries, consistent with the data. We discuss possible contributing mechanisms for generating these upward drifts. These include direct penetration of the magnetospheric electric field to low latitudes, the electric fields generated by the disturbance dynamo, and the effects of conductivity gradients near the dusk terminator and the South Atlantic anomaly.

The Geospace Environment Modeling (GEM) Program is directed toward modeling the coupled solar wind/magnetosphere/ionosphere system. The inter‐calibration of ground‐based observations of the ionosphere and satellite observations has been identified as an essential step in tying together the data to produce a global picture of geospace.On October 10,1988 the DMSP‐F9 satellite passed through the southern hemisphere cusp while a coherent scatter HF‐radar was observing 10‐m scale irregularities present in the ionosphere. The combined data indicate that these irregularities were being generated in the cusp, and that the cusp was a region of greater than normal electric field turbulence. The radar data indicate that the cusp was co‐located with the region where the ionospheric convection rotated from sunward to anti‐sunward with increasing latitude. These observations provide an unambiguous case where simultaneous satellite and ground‐based observations of the cusp can be compared.