antisunward convection over the polar cap and
a return sunward flow at lower latitudes (Fig. 2).
Two features of the ionospheric convection pattern in Fig. 2 are noteworthy: (i) It is rotated 16.1°
clockwise (CW) about the magnetic pole, as determined by the angle between the line connecting
the potential minimum and maximum and the
dawn-dusk meridian; and (ii) more magnetic flux
circulates in the dusk cell than in the dawn cell,
which is manifest in the −6-kV offset in the dusk
cell relative to the dawn cell.
Simple models of the distribution of high-latitude Hall and Pedersen conductance and
field-aligned currents flowing between the magnetosphere and ionosphere show that the rotation in ionospheric convection is due to an auroral
band of enhanced Hall conductance (15, 16).
However, these idealized ionospheric models do
not yield asymmetry in magnetic flux circulation
between the dawn and dusk convection cells. In
specifying a given field-aligned current distribution as input to the well-known Poisson equation
for ionospheric electrodynamics (17), the models
do not self-consistently couple the magnetosphere
and ionosphere and cannot be used to analyze
linkages between asymmetries in the ionosphere
and magnetosphere. For a better understanding
of the relationship between the fast flows observed in the plasmasheet and convection in the
ionosphere, we ran global simulations of the solar
wind–magnetosphere–ionosphere (SW-M-I) interaction (18–20).
Each simulation was run for 8 hours (18).
The upwind boundary conditions during the
last 4 hours for the solar wind (SW) and interplanetary magnetic field in solar magnetic coordinates (8) were NSW = 5/cm3, TSW = 8.5 eV, Vx =
−300 km/s, Bz = −4 n T, and Vy,z = Bx,y = 0. We
then calculated 1-hour average states (Figs. 3 and 4)
during the last hour of the simulation.
Our first simulation (Fig. 3, left panels) has
uniform Pedersen (5 S) and Hall (10 S) conductance in the ionosphere. The resulting ionospheric
convection pattern and plasmasheet fast flows
are dawn-dusk symmetric. The second simulation
(Fig. 3, center panels) uses a causally regulated,
empirical conductance model (21), including contributions from extreme ultraviolet (EUV)–induced
ionization (day-to-night gradient in conductance)
and auroral electron precipitation, which is noticeable as an approximately circular band offset toward midnight from the magnetic pole
with greatest intensity near 69° magnetic latitude and 2300 magnetic local time (MLT). The
nightside ionospheric convection pattern is rotated CW 7° relative to the uniform conductance
run, and the magnitude of the dusk-cell potential
is 8.2 kV larger than that of the dawn cell. Similar ionospheric effects are seen in other global
simulation models when realistic ionospheric
conductance distributions are included (22). The
fast flows in the plasmasheet are concentrated
in the premidnight sector (Fig. 3, bottom center
panel). The final simulation (Fig. 3, right panels)
has a time-independent, artificially depleted Hall
conductance in a band centered on the magnetic
pole. This simulation illustrates how an axisym-metric Hall conductance gradient controls rotation of the convection pattern [counterclockwise
(CCW) 6° in contrast to the CW rotation in the
causal conductance run] and moves the concentration of fast flows to the postmidnight sector.
Although the distributions of conductances in
the uniform (Fig. 3, left) and auroral depletion
(Fig. 3, right) simulations are unrealistic, their
comparison with the simulation using a more
realistic, causally regulated conductance (Fig. 3,
center) illustrates the important effect of ionospheric Hall conduction in controlling convection in the coupled M-I system. The simulation
with causally regulated conductance produces
plasmasheet flows with an average distribution
similar to that of satellite observations (Fig. 1)
and ionospheric convection resembling empirical patterns (Fig. 2).
We performed additional simulations to isolate the effects of the night-to-day conductance
gradient from EUV ionization and from the auroral enhancement in Hall conductance due to
electron precipitation. Including the EUV contribution only produces a weak CCW rotation
and a large asymmetry in the polar cap potentials
with the dusk potential ≈35% larger in magnitude than the dawn potential. A band of enhanced
Hall conductance distributed similarly to the Hall-depletion case in Fig. 3 produces rotation and
asymmetry opposite to that of the Hall-depletion
case: CW with larger dawn than dusk polar cap
potentials. Combining EUV and a contrived Hall-enhancement band produces a configuration resembling the causal-empirical results (Fig. 3,
center). A meridional gradient in Pedersen conductance does not produce rotation or asymmetry.
The ratio of background Pedersen to Hall conductance has a modest effect on the asymmetry and
significantly diminishes the asymmetry only for
unrealistically large ratios. More controlled experiments are needed to fully resolve cause and
effect, but the comparisons indicate that the meridional gradient in Hall conductance at the dayside convection throat (ionospheric projection of
dayside x-line) regulates the asymmetry in polar
cap potentials and dayside rotation, whereas the
meridional gradient in Hall conductance at the
nightside convection throat regulates the nightside rotation. These two effects combine to determine the distribution of magnetotail reconnection
and the plasmasheet transport of magnetic flux.
We derived the simulated reconnection potential FR by integrating the reconnection electric
Fig. 2. Polar distribution of ionospheric convection and electric potential. Dependence on magnetic
latitude and local time (MLT) is derived from the widely used Weimer empirical model (27) with 0° dipole
tilt for solar wind and interplanetary magnetic field input variables indicated in GSM coordinates below
Table 1. Asymmetry in reconnection rates.
Run Average rate (mV/m)
T0 = MLT at FR = 0 T0 – 1 hourT0 + 1 hour
Causal (T0 = 23.91) 0.96 0.83
Depletion (T0 = 24.30) 0.68 0.93