Large transient variations in magnetic fields and other parameters often observed on the ground at magnetic latitudes from 70 to 80 (the region of the dayside cusp and cleft), denoted variously as magnetic impulse events or traveling convection vortices, have been considered important because of their probable role in coupling momentum and energy from the magnetopause boundary into the high latitude ionosphere. These often solitary waves have become better understood in recent years, again thanks to both observational studies and modeling efforts.
Because of the spatially extended nature of these transient events,
observational studies have increasingly used data from multiple
locations. Lanzerotti et al. [1991] provided detailed
occurrence statistics on these events, typically observed on the
ground as single-cycle pulsations with periods 
100-500 s, using
conjugate high-latitude data from South Pole and Iqaluit
[4]
(Canada),
and characterized the field-aligned currents associated with them.
Potemra et al. [1992a,b] used data from several satellites
supplemented by ground arrays of magnetometers to show the variety
of ULF waves and transients, including a possible flux transfer
event, stimulated by inferred magnetospheric compressions.
Luehr and Blawert [1994] presented several examples of
travelling convection vortex events observed by large numbers of
magnetometer stations and by the European Incoherent Scatter Radar
(EISCAT), and deduced a typical distribution of vortices as a
function of local time during their main phase. A set of 
5
vortices (and possibly other, very weak ones) is established during
the main phase of such events, with each travelling about 6 hours in
local time before disappearing. This distribution has been amply
confirmed by new data from the Magnetometer Array for Cusp and Cleft
Studies (MACCS) in Arctic Canada [ Hughes et al., 1994].
Figure 2, showing MACCS data from three stations separated in
longitude by 
600 km, indicates multiple vortices grow and decay on
a time scale comparable to the time it takes a vortex to pass a
single site. Both MACCS data and initial data from the Magnetometer
Array in the Greenland Ice Cap (MAGIC) presented by Sitar et
al. [1993] indicate typical azimuthal velocities near 10 km/s.
Theoretical studies of these events include the analytical studies
of Kivelson and Southwood [1991] and Southwood and
Kivelson [1993b] and continuing three-dimensional numerical
simulation work by Lysak and Lee [1992] and Lysak et al.
[1994]. Figure 3, from the simulation of Lysak et al. [1994]
which incorporated an ionosphere with finite conductivity and an
open tail boundary condition, shows three clear vortices in a
simulation of the ionospheric magnetic field resulting from a 240-s
pulse at the magnetopause, with the central vortex the strongest and
in the sense of the region 1 field aligned (Birkeland) currents, at
9:00 magnetic local time (MLT). Their latitudinal and dawn-dusk
signatures are in good agreement with the above observations and
empirical models. Lysak et al. [1994] also added a
B
-dependent magnetic helicity injection, characteristic
of transient reconnection events, which led them to predict a
dawn-dusk asymmetry in vortex occurrence depending on the sign
of the IMF B
component.
There is still disagreement about the cause of the pulses on the
magnetopause that generate these events, however. Sibeck
[1994] has reviewed various mechanisms for both transient and
periodic events in the outer dayside magnetosphere, including solar
wind (and near-Earth) pressure pulses, impulsive plasma penetration,
the Kelvin-Helmholtz (generalized ``wind-over-water'') instability,
and sporadic reconnection (flux transfer
[4] events).
[0]
Sibeck
[1992] presented a case study showing the limited scale lengths of
IMF features upstream or near Earth, and pointed out that variations
in solar wind dynamic pressure applied to the magnetopause depend on
the orientation of the IMF, especially as it determines the location
of the ion foreshock region. In a related study, Song et al.
[1994a] delineated observational characteristics distinguishing
magnetopause surface waves from flux transfer events. Whereas
Sibeck [1992] found evidence of solar wind pressure variations for
each of the events discussed in that study, Konik et al.
[1994], on the basis of a statistical analysis of the events
identified by Lanzerotti et al. [1991], concluded that
sporadic reconnection was responsible for at least 50-70% of these
events, possibly mediated by a Kelvin-Helmholtz instability in the
low latitude boundary layer.