Inner magnetospheric electric fields are often complicated by a reduction of the total electric field at low-latitudes called shielding and an intensification of the electric field equatorward of the auroral oval called a subauroral polarization stream (SAPS). To better understand the stormtime behavior of these electric fields, computer experiments have been conducted using the Rice Convection Model (RCM). The RCM computes the Region-2 Birkeland currents and the inner magnetospheric electric field and particle distribution. It uses as input a magnetic field and the plasma sheet pressure and energy distribution and assumes that particle loss is negligible. The base or control run used realistic model inputs, while the experiments varied fundamental model inputs to test the electric field's sensitivity to magnetospheric conditions. The first experiment is a high-pressure or dense plasma sheet run, where the boundary density and pressure are a factor of three higher than in the base run. The high-pressure run produced better shielding and a stronger SAPS. The second experiment is a cold plasma sheet run in which the plasma sheet temperature was reduced by a factor of thirty, while the pressure remained constant. In this run, the gradient and curvature drift is small compared with the × drift. With particle transport similar to ideal MHD, the cold plasma sheet run produced excellent shielding and strong Region-2 currents. This is in contrast to the weak shielding and Region-2 currents produced by global MHD codes. In addition, the SAPS feature is significantly strengthened. The final experiment held the magnetic field constant. This run produced marginally better shielding. However, the magnetospheric signature of SAPS is very weak, suggesting that magnetic field variations are important in SAPS formation.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 In addition to shielding, subauroral polarization streams (SAPS) [Foster and Burke, 2002] are significant deviations from a simple convection electric field. SAPS are strong poleward directed electric fields equatorward of the electron aurora in the evening and night sectors. Foster and Vo  has shown that SAPS are common ionospheric features during active conditions, and Rowland and Wygant  have found similar occurrences in the magnetospheric signatures of SAPS. Like the shielding phenomenon, SAPS are also associated with the Region-2 current system. During periods of intense magnetospheric convection, the inner edge of the plasma sheet ions can be pushed Earthward of the plasma sheet electrons. These electrons precipitate into the ionosphere to generate the high-conductivity, diffuse auroral oval. Since the Region-2 currents are equatorward of the high-conductivity auroral oval, the electric field generated by the horizontal closure current can become significant. It adds to the convection field poleward of Region-2 currents and reduces the convection field equatorward of Region-2 currents.
 Since both shielding and SAPS are strongly driven by the Region-2 currents, the subauroral electric field is affected by magnetospheric conditions. Vasyliunas  showed that the Region-2 Birkeland currents j∥ flowing into symmetric ionospheres are given by
where B is the magnetic field strength, the subscripts i and eq denote the value in the ionosphere and the magnetic equatorial plane, respectively, is the unit vector of the magnetic field direction, V is the volume of a magnetic flux tube containing 1 Wb of magnetic flux, and p is the thermal pressure of the magnetospheric plasma. In the inner magnetosphere, the pressure gradient is dominated by the inner edge of the plasma sheet. Thus thermal properties of the plasma sheet affect the strength and distribution of the Birkeland currents and through them the structure of the subauroral electric fields. Similarly, equation (1) indicates that changes in the magnetic field topography will change the Birkeland currents and thus the electric field distribution. By studying the effects of plasma sheet and magnetic field variations, the behavior of the subauroral electric field can be better understood.
 This paper presents simulations of the inner magnetospheric electric field by the Rice Convection Model (RCM). These simulations examine a major magnetic storm, the 4–6 June 1991 storm in which the Dst dropped to −223 nT, and use a realistic (but not self-consistent) magnetic field model that varies in time to remain consistent with certain geophysical parameters. A base run was conducted, which shows the electric field behavior for realistic plasma sheet and magnetic conditions. By varying the density (termed the dense or high-pressure run) and the temperature (cold run) of the plasma sheet and by using a constant magnetic field (constant magnetic field run), we are able to characterize the sensitivity of the subauroral electric fields to plasma sheet conditions and magnetic field variability.
2. Rice Convection Model
 The Rice Convection Model (RCM) self-consistently calculates the inner magnetospheric particle distribution, the Region-2 Birkeland currents, and resulting electric field patterns. At present the RCM is not fully self-consistent, since the magnetic field is assumed and magnetic field generated by the Birkeland currents are not calculated. The particle distribution is represented by a series of particle populations of a given adiabatic energy invariant, λ, and a given charge q. The adiabatic energy invariant, a form of the adiabatic pressure invariant pV5/3, is given by
where E is the particle energy and V is the magnetic flux tube volume per unit magnetic flux ∫ ds/B. The adiabatic energy invariant is conserved assuming slow flow (v ≪ vAlfven and v ≪ vthermal) and isotropic particle distributions (maintained by strong pitch-angle scattering) [Wolf, 1983; Heinemann and Wolf, 2001]. Isotropy is a valid assumption in the middle plasma sheet [Sergeev et al., 1993]. The population of particles with adiabatic energy invariants within a given range Δλ around a given λ are considered separate particle species by the RCM. Thus an adiabatic density invariant ηs can be defined as the number of particles of a given energy invariant and charge within a flux tube of unit magnetic flux and is related to the density by
where n is the number density of particles of a given energy invariant and charge. The density invariant is conserved along a drift path if chemical processes (e.g., charge exchange) and precipitation are negligible. In this study, both chemical processes and particle precipitation are assumed to be unimportant.
 With this particle formulation, the RCM solves the fundamental equation of magnetospheric-ionosphere coupling [Vasyliunas, 1970; Wolf, 1983]
and the bounced-averaged guiding center drift drift,s [Wolf, 1983]
where ∇H is the horizontal gradient in the ionosphere, is a tensor of the field-line-integrated ionospheric conductance, Φ is the electric potential, I is the magnetic dip angle, B is the magnetic field strength, the subscripts i and eq denote the value in the ionosphere and the magnetic equatorial plane, respectively, is the unit vector of the magnetic field direction, and s designates a particle species of given λs, qs, and ηs. The only species dependent quantity in equation (5) is λ/q, meaning the drift is essentially a function of energy per charge. Thus the RCM treats protons of given λ exactly like He+ and O+ ions of the same energy invariant and does not consider the ratio of protons to heavier ions.
 The basic logic of the RCM (shown in Figure 1) is a modification of the Vasyliunas logic loop [Vasyliunas, 1970]. In the RCM logic diagram, rounded boxes designate RCM inputs, while square boxes indicate RCM calculations. In order to reduce numerical diffusion, the RCM uses a series of test particles to simulate the boundary of a particle population (usually the inner edge of the plasma sheet). Since η is constant along a drift path, these test particles provide contour lines of constant density invariant and allow the RCM to maintain sharp boundaries. The Rice Convection Model has been described in greater detail by Harel et al. [1981b], Wolf , and Toffoletto et al. .
3. Model Setup and the Base Run
Figure 2 shows the geophysical conditions of the storm on 4–5 June 1991. These runs use the Hilmer and Voigt  magnetic field model as a model input. The Hilmer-Voigt field model uses the magnetospheric standoff distance (computed from the solar wind ram pressure), Dst, and the low-latitude boundary of the auroral oval at midnight [Gussenhoven et al., 1983], hereafter referred to as ABI, to determine the magnetic field. The flux tube volumes and mapping information are computed for a series of mark times through the event by interpolating from previously computed models of discrete input values. At each time step, the magnetic field values are computed by linear interpolation of the magnetic field values at the mark times. This procedure, which has been well tested, saves a substantial amount of computer time.
 The electrical potential on the RCM's polar boundary is a sine function in magnetic local time with the maximum potential set equal to half of the polar cap potential drop (PCP) at 0600 MLT and -PCP/2 at 1800 MLT. The PCP determined from Defense Meteorological Satellite Program (DMSP) Ion Drift Meter observations of the ion drift is shown in Figure 3.
 The plasma sheet is represented by a series of test particles that represent the inner edge of the plasma sheet. These inner edges are placed along contours of constant flux tube volume starting just inside the noontime magnetopause. The electron temperature in the plasma sheet is a factor of 7.2 lower than the ion temperature, on average [Baumjohann et al., 1989]. Hence the pressure of the electron plasma sheet is usually not important, and the electron plasma sheet can often be ignored. However, in these experimental runs a single electron plasma sheet edge is used as a tracer. In addition, the preexisting ring current population is represented by a series of 12 edges placed at different locations in the inner magnetosphere. All of these edges are initially aligned with contours of the flux tube volume so that the initial Birkeland currents are zero and the currents can evolve naturally with time. The plasma population is calculated assuming a kappa distribution [Vasyliunas, 1968; Christon et al., 1989] from an empirical model used in the Magnetospheric Specification Model (MSM) [Lambour, 1994]. This empirical model uses plasma fluxes from a variety of observations and statistical studies [Huang and Frank, 1986; Garrett and Faudet, 1989; Baker et al., 1978, 1984, 1982; Spjeldvik and Rothwell, 1985; Vette, 1991; Lyons and Williams, 1975a, 1975b; Christon et al., 1991; Gloeckler and Hamilton, 1987; Hamilton et al., 1988], and fits the data to either a kappa or bi-kappa distribution at four mark radii (3, 4, 6.6, and 13 RE). The particle fluxes are binned by log10(energy) and Kp. The preexisting ring current population is calculated for Kp = 0 conditions. The value of the plasma sheet distribution function at the RCM's high-L boundary is set for Kp = 3 conditions. The Hardy et al.  empirical model of the auroral precipitation is used to calculate the auroral conductivity. The background (solar-produced) conductivity is calculated from the International Reference Ionosphere (IRI-95) [Bilitza, 1997] and the Mass Spectrometer Incoherent Scatter (MSIS) model [Hedin, 1991]. Although strong electric fields will reduce the ionospheric conductivity [Schunk et al., 1976], no attempt has been made to include this effect in any of these runs.
 The model run using these inputs is designated the base run. It best represents the actual plasma sheet conditions during this storm period. There are uncertainties within the RCM's inputs. This study is attempts to understand the effects of these uncertainties upon the model output. The model runs deviating from the base run are designated by the form of the deviation: a dense or high-pressure plasma sheet, a cold plasma sheet, and a constant magnetic field. Some base run results are presented in Figure 4, which shows the potential distribution of the electrostatic field (in the corotating frame) at eight times during the magnetic storm. These times represent the conditions near the beginning of the RCM simulation (0600 UT), at the beginning of the storm (1500 UT), during the first plasma sheet injection (2000 UT), during the recovery from the first injection (2500 UT), at the second injection (3200 UT), during the recovery afterwards (3400 UT), near the third injection (4600 UT), and at the end of the model simulation (4800 UT).
 The base run is characterized by weak shielding and strong SAPS. Throughout this run, a strong electric field exists at low L-shells. While this field is rotated with respect to the electric field in the tail, most of the imposed convection field reaches low L-shells through the storm. In addition, a strong electric field has formed at moderate L-shells (L ∼ 4 or 5) on the duskside by 2000 UT. This field enhancement is the magnetospheric signature of a subauroral polarization stream and agrees with the electric fields statistics found by Rowland and Wygant . The SAPS feature is most intense after the second injection but is a persistent feature throughout the storm.
4. Dense Plasma Sheet Experiment
 The first of the numerical experiments increased the density in the plasma sheet by a factor of three, with the temperature and the shape of the distribution function held constant. Figure 5 shows the equipotential distribution (where the corotation field has been subtracted off) at the eight mark times during the storm (these are the same times shown in Figure 4). Figure 6 provides a more detailed comparison of the electric field in the base and dense runs. The solid line shows the ratio of the radial electric field −δΦ/δr to the cross-tail electric field PCP/W (where W is the width of the magnetotail) along the the 1800 MLT line. The dotted line shows the same ratio for the high-pressure run. For perfect shielding, the ratio is zero; no shielding is ratio of one. The SAPS feature is indicated as a ratio over one in the L ∼ 4–5.5. The eight panels in this figure correspond to the same mark times as Figures 4 and 5.
 The first plot (starting at the upper left hand corner and going down in Figures 5 and 6) shows the potential distribution shortly after the model run begins but before the storm starts. Notice the similarity to the initial potential distribution and the electric field strength in the base run. Near the beginning of the storm at 1500 UT (second plot on the left hand side), the potential distribution is still very similar to that in the base run. The first injection occurred near 2000 UT (third plot, left side). The denser plasma sheet produces a stronger electric field in the afternoon sector outside of 5 RE (as seen by the higher concentration of equipotentials in Figure 5 and the stronger field strength in Figure 6) and a considerably weaker electric field inside (fewer equipotentials). In the morning sector, the electric field near the Earth is weaker than in the base run but not as dramatically as in the afternoon sector. The plot for 2500 UT shows the potential distribution near the recovery from the first injection. This plot shows a weaker low L-shell electric field in the dense plasma sheet run than in the base run. Outside of 4 RE, the electric field is stronger in both runs, but it is more intense in the dense plasma sheet run. The second injection occurred near 3200 UT. At this time, the dawnside shielding is roughly as strong as the duskside shielding. As the magnetic storm progresses, some plasma sheet particles become trapped which produces comparable amounts of shielding in both the dawn and dusk sectors. Figure 7 shows the location of the inner edge of average energy plasma sheet ions at UT 2500 and UT 3200. Notice that the dawn-dusk asymmetry is smaller at UT 3200. Additionally, the ionospheric conductances also have a dawn-dusk asymmetry that produces greater shielding on the duskside. Both the shielding and the SAPS structure remain strong in the dense plasma sheet run during the recovery from the second injection (3400 UT), the third injection (4600 UT), and the end of the model run (4800 UT).
 It is not surprising that increasing the plasma-sheet density without changing the temperature produces stronger shielding and more intense SAPS. Simple shielding theory [Jaggi and Wolf, 1973] indicates for a given energy distribution of plasma-sheet particles, the strength of the electric field Earthward of the shielding layer should be inversely proportional to the total number of particles per unit magnetic flux (i.e., the total η). A higher-pressure plasma sheet provides more particles for the Region-2 currents; a stronger current generates a more powerful closure current electric field.
5. Cold Plasma Sheet Experiment
 In order to change the plasma sheet temperature while maintaining the base run pressure, it is necessary to change the shape of the distribution function. Instead of the kappa distribution used in the base run, the plasma sheet is represented by a monoenergetic particle distribution of λ = 177.14eV RE−2/3/nT which corresponds to an energy of roughly 300 eV at 12 RE, about 1/30 of the base run plasma sheet temperature. Since the pressure invariant, pV5/3 ∼ ηsλs, is most directly related to the Birkeland currents, ηsλs (where sum is over the species s) was kept the same as in the base run and the invariant density in the single plasma sheet channel is increased to maintain the plasma sheet pressure. Figure 8 shows the potential distribution (without the corotation field) from the cold plasma sheet run at the same mark times shown in Figure 4. Figure 9 provides a more detailed comparison of the electric field in the base and cold runs. The format is identical to Figure 6. The solid line shows the base run ratio, while the dotted line shows the same ratio for the cold plasma sheet run.
 The cold plasma sheet results do not differ significantly from the base run results at 0600 UT. However, the cold plasma sheet quickly begins to shield the inner magnetosphere from the convection electric field. This shielding, which only weakly develops in the base run late in the storm, now develops before the storm really starts. By 1500 UT, stronger shielding has already developed with the dusk electric field inside of 4 RE being roughly 40% of the cross-tail field in the cold plasma sheet run compared with 60% in the base run. The shielding produced in the cold plasma sheet run continues to strengthen during the storm. It becomes very strong by 2000 UT where the electric field inside of the 4 RE is less than 0.5 kV/RE. This increased shielding agrees with previous studies [e.g., Spiro et al., 1981]. Outside of the plasma sheet inner edge, the total electric field is dramatically stronger (three to four times the cross-tail electric field strength). This is particularly true in the morning and afternoon sectors. These regions of intense field strength have a very different structure from similar regions in the base run. They are more symmetric about noon. In the base run, the strong electric field regions are seen in the afternoon and dusk sectors (with a small region around midnight) but not in the morning sector. However, in the cold plasma sheet run, the prenoon electric field is quite strong, indicating the penetration of the plasma sheet ions inside of the electron plasma sheet in the dawn sector.
 In addition to producing better shielding, the cold plasma sheet generated a different potential pattern. The shielding in the base and dense plasma sheet runs first develops in the afternoon sector and remains stronger on the duskside. However, in the cold plasma sheet run, the shielding is more uniform, with comparable shielding on both the dawnside and duskside. This phenomenon is associated with the thickness of the shielding layer, which is different between the base and the cold plasma sheet runs. In the cold plasma sheet run, the shielding layer is very thin. In the base run, it is more spread out. While both of these effects are produced in part by the use of only one plasma sheet edge in the cold plasma sheet run, both the increased thickness of the shielding region and the stronger dawn-dusk asymmetry in the shielding are related to the temperature of the plasma sheet. In a warm plasma sheet like the one used in the base run, the shielding layer spreads out as the colder parts of the plasma sheet penetrate closer to the Earth. The separation of plasma sheet edges is greater on the dawnside of the magnetosphere because gradient and curvature drifts of the particles oppose the drift caused by the convection field. However, for a sufficiently cold plasma sheet, there is little spread in the shielding layer because the main pressure-bearing plasma sheet ions are not significantly affected by gradient-curvature drifts and penetrate to the same radius.
 The cold plasma sheet produces greater shielding because the colder plasma sheet penetrates to lower L-shells. Figure 10 compares the location of the inner edge of the plasma sheet for the mean energy of the plasma sheet particles in both the base run and the plasma sheet in the cold run at 2500 UT. The cold plasma sheet penetrates more symmetrically in the dawn-dusk plane and penetrates closer to the Earth. While the plasma sheet in the base run penetrates to the same location as the cold plasma sheet at dusk (roughly L ∼ 4), the warmer base run plasma sheet is significantly farther from the Earth at dawn (roughly L ∼ 6) where the gradient-curvature drift opposes the convection.
 The deeper plasma sheet penetration increases the shielding. Although an initial examination of the magnetosphere-ionosphere coupling equation (4) suggests that the temperature of the plasma sheet should not substantially affect the shielding electric field if the pressure invariant, pV5/3 ≈ Σηλ, is held constant, this experiment shows that the plasma sheet temperature does affect the development of the electric field. Equation (4) can be simplified to
 1. The plasma sheet can be treated as a single fluid of energy invariant λ and density invariant η.
 2. The only gradient in λη is at the inner edge of the plasma sheet, and the inner edge of the plasma sheet is a sharp boundary. Mathematically, this statement is ∇(λη) ≈ Δ(λη)δ(L − Lps)/RE, where Lps is the L-shell of the plasma sheet inner edge.
 3. The conductance is a scalar constant Σ.
 4. The ionospheric electric field varies only in colatitude giving ∇ · ≈ (Ri sin(θ))−1d(E sin(θ))/dθ.
 5. The magnetic field is a dipole, where Beq = −MRE3/L3, L = (sin θ)−2, Bi is a constant, sin(I) = ∼ 1, and
where M is the Earth's magnetic moment.
 6. Ri ≈ RE.
 Thus the change in the electric field across the inner edge will be
From Figure 10, the plasma sheet inner edge location at midnight is Lbase ≈ 5 and Lcold ≈ 3.75. Using these values, the relative magnitudes of the change in the electric field across the plasma sheet inner edge are
At 2500 UT the average electric fields inside of L = 4 at midnight for the base run is ∼1.05 mV/m and ∼0.05 mV/m for the cold plasma sheet run. The electric field at midnight around L = 4 is ∼1.3 mV/m for the base run and ∼0.82 mV/m for the cold run. Therefore the change in the electric field at 4 RE is roughly 1.30 mV/m − 1.05 mV/m = 0.25 mV/m in the base run and 0.77 mV/m in the cold run, a factor of 3.1. This is good agreement, considering the number of simplifying assumptions made in this discussion.
 The cold plasma sheet run is very similar to the plasma sheet conditions in global MHD models. These models frequently exhibit Region-2 currents that are unrealistically weak and often poorly defined. It is tempting to attribute this persistent problem to the fact that MHD neglects particle transport by gradient/curvature drift. That is, the ideal-MHD assumption that + × = 0 implies that v⊥ = × /B2, which corresponds to neglect of the gradient/curvature drift term in equation (5). This assumption is unrealistic in the inner plasma sheet and ring current. The much smaller λ (a factor of 30 less than the normal plasma sheet value) particles gradient/curvature drift very slowly and are thus close to conforming with ideal MHD. The cold plasma sheet run (Figure 9) showed much better shielding than the base run, which implies stronger Region-2 Birkeland currents. Figure 11 compares the total Region-2 Birkeland current per unit local time for the base and cold plasma sheet runs. For a given plasma sheet pressure, a hotter plasma sheet reduces the shielding and produces weaker Region-2 currents. As discussed above, strong gradient/curvature drift inhibits plasma sheet particles from getting close to the Earth, reducing their ability to form the asymmetric ring current that connects to the Region-2 Birkeland currents. It should also be noted that Southwood  demonstrated Region-2 currents and the shielding effect can be derived from ideal MHD. While this work does not suggest a cause for the weak MHD Region-2 currents, it does suggest that the cooler temperatures and the neglect of gradient/curvature drift, common in MHD models, should produce stronger, not weaker, currents.
6. Constant Magnetic Field Experiment
 In the base run the magnetic field changes at every time step to remain consistent with the observed geomagnetic conditions. For comparison, the RCM has been run with a constant magnetic field that is set according to the geomagnetic conditions for 1330 UT on 4 June. The results of this run are shown in Figure 12 which gives the potential distribution (without the corotation potentials) for each of the eight times shown in Figure 4. Figure 13 provides a more detailed comparison of the electric field in the base and constant magnetic field runs. The format is same as Figure 6. The solid line shows the base run ratio, while the dotted line shows the same ratio for the constant magnetic field run. The potential distribution for the constant magnetic field run looks similar to the distribution produced by the base run at 600 and 1500 UT because the magnetic field changes very slowly through this period in the base run. At the first injection the potential distributions differ. The constant magnetic field run produces slightly better shielding and weaker SAPS structures. The shielding grows stronger through the second injection at 3400 UT and is strongest in the night sector. In the base run with the changing magnetic field, the electric field is very strong in the afternoon sector, indicating a SAPS. However, a similar structure only weakly develops in the constant magnetic field structure. The most noticeable difference between the potential distributions created by the base and constant magnetic field runs is the rotation of the low L-shell pattern towards midnight. In the constant magnetic field run the strongest electric field is generated around midnight instead of dusk as in the base run. The exception is the strong afternoon field at 3400 UT.
 Both the shielding and the SAPS have been substantially altered by allowing realistic magnetic field variations. There are three principle reasons for this. The first reason is the development of the induction electric field. In the RCM, magnetic fields are treated as equipotentials with stationary ionospheric footpoints. In the base run the magnetic field changes with solar wind and geomagnetic conditions. When the magnetic field is reconfigured, the ionospheric electric potential maps to a different location in the magnetic equatorial plane. An alternate view of this phenomenon [Fejer et al., 1990] is that the equatorial crossing point of a field line moves with a velocity vfield and an induction electric field = −field × is inferred. As the magnetic field compresses in response to high solar wind ram pressure, most of the equatorial field line crossings move Earthward. This generates an azimuthal induction electric field. In the alternative view the equipotentials map to a lower L-shell, which strengthens the azimuthal electric field (through r−1dΦ/dϕ). A similar argument holds a weaker azimuthal electric field in an expanding magnetosphere. As the auroral zone expands in the storm's main phase and the magnetic field is stretched tailward in the midnight sector, an inductive electric field is created that is directed from dusk to dawn at midnight in opposition to the convection field. In a period when the nightside magnetic field becomes more dipolar, the inductive electric field is directed from dawn to dusk at midnight adding to the convection electric field. Since the magnetic field lines at midnight move the most, the inductive electric field will have the greatest impact around midnight. This helps explain the differences in the midnight potential pattern.
 Another reason is the compression and the expansion of the magnetosphere. The potential drop across the magnetosphere is the same in both the base run and the constant magnetic field run. However, the magnetosphere is smaller for most of the base run strengthening the electric field since the same potential is applied across a smaller magnetosphere. Figure 13 shows the ratio of the radial electric field to the cross-tail field strength and thus accounts for the compression/expansion effect. The differences in Figure 13 indicate that the stronger cross-tail electric field is not the dominate cause of the differences in the potential distributions in Figures 4 and 12.
 Finally, Fejer et al.  argued that a changing magnetic field will change the structure of the shielding layer. When a magnetic field line is stretched, it pulls the particles on the field line with it in order to maintain the “frozen-in-flux” condition. This motion changes the shielding by moving the plasma sheet inner edge tailward and altering the cross product in equation (4) (generally reducing the angle between the gradients). At midnight, this motion helps to reduce the potential electric field by putting the plasma sheet inner edge farther from the Earth (the L−8/3(L − 1)−1/2 dependence in equation (7)) and increasing the volume of the flux tube (the V−5/3 term in equation (4)). Both of these factors reduce the shielding. The effects of a changing magnetic field are most noticeable in the midnight sector because that is where magnetic field lines change the most during a magnetic storm.
 Although the inner edge of the plasma sheet generally shields the inner magnetosphere from the full force of the convection electric field, our base run simulation of the main phase of the major storm of 4–5 June 1991 shows that some of the strongest electric fields in the entire equatorial magnetosphere occurred near the Earth. The convective × drift sometimes seem focused into the inner magnetosphere, rather than diverting around it. These strong electric fields are the magnetospheric signature of subauroral polarization streams and agree with the statistical electric fields seen by Rowland and Wygant . To better understand how plasma sheet conditions and magnetic field variations affect the electric field, numerical experiments with the Rice Convection Model have been conducted. These experiments show that
 1. The strength of shielding increases with the value of pV5/3 assumed in the plasma sheet. Similarly, the strength of the SAPS increases with pV5/3.
 2. Shielding and the Region-2 currents tend to be strong if the plasma sheet is cold and dense and weak when it is hot and rarefied. The use of the ideal-MHD approximation v⟂ = × /B2 should lead to an overestimate of shielding and an overestimate of Region-2 currents. In contrast, MHD models tend to produce weak Region-2 currents and insufficient shielding. This is a meaningful inconsistency which should be further investigated by MHD modelers.
 3. A time-varying magnetic field alters the distribution of electrical potential. In particular, a time-varying magnetic field moves the strong mid-L-shell electric field (a SAPS) towards dusk and intensifies this field. It also reduces the shielding through the combined effects of an induction electric field, stronger fields due to compression, and the tailward motion of the plasma obeying “frozen-in-flux” as the field is stretched tailward.
 The author would like to thank Richard Wolf and Robert Spiro for their input into this study and for providing important insight into the model results. In addition, the author would like to thank Stanislav Sazykin, Bela Fejer, and Bill Burke for enlightening conversations and insight. The DMSP data were provided courtesy of William Denig of the Air Force Research Laboratory. This work was supported by the NASA Sun-Earth Connection Theory Program under grant NAG5-11881 and by the Applied Research Laboratories, The University of Texas at Austin under an ARL:UT postdoctoral fellowship.
 Arthur Richmond thanks John C. Foster and Michael A. Heinemann for their assistance in evaluating this paper.