The dependence of high-latitude ionospheric convection on the orientation of the Interplanetary Magnetic Field (IMF) is an important part of understanding the fundamental processes which determine our space weather. Many studies have addressed how the convection responds to a change in the orientation of the IMF [e.g., Nishida and Maezawa, 1971; Lockwood et al., 1986; Etemadi et al., 1988; Saunders et at., 1992; Taylor et al., 1998, and references therein]. A standard model of the convection response to a change in the IMF has emerged, whereby reconnection of the new IMF state initiates a change in the plasma flow at the ionospheric footprint of the new X line. The new convection flow then expands dawnward and duskward in a twin-vortex pattern at a phase speed of 5 km s-1 resulting in a newly-established global convection pattern after 15 min [Cowley and Lockwood, 1992].
Recently several studies have reported observations that apparently contradict this model [Ridley et al., 1997, 1998; Ruohoniemi and Greenwald, 1998]. These studies seem to indicate that the ionospheric convection response to a sudden change in the IMF is globally instantaneous, i.e., that the entire polar ionosphere responds nearly simultaneously (<2 min) to reconnection of the new IMF state at the magnetopause.
In all of these studies the IMF is monitored using a spacecraft, typically upstream of the Earth's bow shock. Changes in the IMF are propagated to the Earth's bow shock and through the magnetosheath to the subsolar magnetopause, where reconnection is believed to initiate. Several minutes (1-3) are added to the subsolar magnetopause impact time to account for the communication between newly reconnected field lines at the magnetopause and their ionospheric footprints. The lagged (positive or negative) IMF is then compared to the observations of the ionospheric plasma convection, obtained by a variety of different techniques, and conclusions about the ionospheric response are inferred. Critical to the validity of this method is the ability to determine precisely when the IMF arrives at the magnetopause. Ridley et al.  describes in detail some of the uncertainties involved in such techniques and conclude that typical time estimates may be uncertain to 8 min.
We believe that additional difficulties exist in determining when a change in the IMF impacts the magnetopause and that the uncertainty can be even greater. A potentially important element missing from this type of analysis is the detailed plasma flow in the magnetosheath, in particular, the extent to which field lines are draped over the magnetopause when reconnection of the new IMF state occurs. Instead of trying to precisely measure the time when a change in the IMF impacts the subsolar magnetopause, which may or may not be the site of reconnection, we are suggesting that the amount of field line draping may significantly influence the nature of the convection response. In some cases field lines of the new IMF state may be draped over a significant portion of the dayside magnetopause and reconnection may occur over a larger region of the dayside than was previously thought.
In order to address this issue we have selected an event which was marked by a large and rapid (<2 min) southward turning of the IMF that was observed by the WIND spacecraft upstream of the bow shock and the GEOTAIL and IMP8 spacecraft in the magnetosheath. To determine the convection response we use the Northern Hemisphere array of SuperDARN radars which monitor a large portion of the dayside high-latitude ionosphere. The approach we take is to observe the IMF change at the various satellites and estimate the degree to which field lines of the new IMF state are draped over the magnetopause at the time when new reconnection is believed to occur, or 2 min before the first significant ionospheric convection response is observed.
Our findings for this event are consistent with field lines of the new IMF state draping over the entire dayside magnetopause. Such a configuration of field lines at the magnetopause would help to explain the observed nearly simultaneous (2 min) response over the dayside ionosphere with the SuperDARN array, similar to the recent results presented by other researchers [Ridley et al.,1997, 1998; Ruohoniemi and Greenwald, 1998]. The draping provides a mechanism for the rapid, large-scale onset of convection over the dayside ionosphere.