The Geospace Environment Modeling (GEM) substorm challenge event of November 24, 1996, has been used as a test case to investigate the influence of different data and model inputs on the assimilative mapping of ionospheric electrodynamics (AMIE) outputs. During that period the interplanetary magnetic field (IMF) went from northward to southward and then returned to northward. In the later part of the day a moderate substorm with AL~-600 nT took place. The AMIE convection patterns derived from the Super Dual Auroral Radar Network (SuperDARN) data alone are generally similar to those derived using ground magnetometer alone, especially during the relatively stable southward IMF period. However, some differences are found during the northward IMF period and during the substorm; namely, the reversed convection configuration near local noon imaged by SuperDARN is absent in the magnetometer observations while the strong convection on the nightside recorded by the magnetometers during the substorm expansion phase is not seen by the radars. Different conductance models do not seem to have a big effect on the large-scale distributions of ionospheric convection and Joule heating, but they do alter the cross polar cap potential drop and the hemispheric integrated Joule heating rate by nearly a factor of 2. When the AMIE-derived electric potential drop was compared with the in situ measurements along the satellite track, it is found that AMIE underestimated the potential drop by 20 kV, amounting to a 22% underestimation. The analysis of the AMIE results based on the SuperDARN and ground magnetometer data reiterates the view that the response of ionospheric convection to a sudden IMF southward turning is global and nearly simultaneous. ¿ 2001 American Geophysical Union