This is the first study in which a physical ionospheric model (time-de
pendent ionospheric model (TDIM)) has been driven through a substorm u
sing self-consistent magnetospheric convection electric field and auro
ral electron precipitation inputs. Both of these were generated from a
simulation of a red substorm event using the MHD model [Fedder et al.
, 1995b]. Interplanetary magnetic field (IMF) data were available for
1.5 hours until the substorm breakup. Hence the substorm growth and ex
pansion dynamics is captured in a 1.5-hour time period. As a reference
against which to compare this TDIM substorm simulation, a typical cli
matological TDIM simulation was carried out using standard statistical
representations of the convection electric field and auroral oval. No
te that these statistical representations are driven by the K-p index.
This is a 3-hour index, yet the substorm growth and expansion occurs
in 1.5 hours. Hence a static convection electric field and auroral ova
l are used for: the TDIM reference simulation. From the comparison of
these two simulations, we find, as expected, the E region densities ar
e different. However, these differences lead to factors of 2-4 differe
nces in the integrated Half and Pedersen conductivities, These conduct
ivities, in turn, are crucial as an ionospheric boundary condition for
magnetospheric MHD modeling. The F region spatial and temporal respon
ses are complex and exhibit large differences, from tens of percents t
o factors of 4 in density and up to +/-70 km in h(m)F(2). These differ
ences are all larger than typical experimental uncertainties. The days
ide and cusp variabilities are very sensitive to the convection patter
n and are not well correlated to magnetic indices, such as the 3-hourl
y K-p index, In the polar cap, the differences in the location of the
tongues of ionization and the polar holes readily lead to factors of 2
-4 in local density differences, Differences in the locations of ''bou
ndaries'' in the plasma convection and auroral precipitation lead to l
arge differences in the local F region densities and in the locations
of strong density gradients, both of which are relevant to space weath
er applications.