
Geotail
12s average low energy particle (LEP) data and magnetic
field (MGF) data is used to identify plasma
sheet fast flows

We
identify periods of fast flows by using a criterion of
Vx_{perp} > 300 km/s, where Vx_{perp}
is the X component of the
plasma
flow perpendicular to the ambient magnetic field. If two
fast flows occur within 15 min, they are considered one
event. We have identified 921 events from the 2year
period of the data. The starting time of the event is
designated as the time of the first data point with 
Vx_{perp} > 300 km/s.

We
also impose a constraint of ß
>
0.5 for covering at least 85% duration of the event on
the selection of the fast flow events, where ß
is
the ratio of the ion pressure to the magnetic pressure.
This constraint ensures that Geotail is mainly in the
central plasma sheet, not in the plasma sheet boundary
layer. Moreover, we select events with Geotail locations
around the noonmidnight meridian (20
< Y <
20 R_{E})
to avoid some magnetosheath events. We only use fast
flows events that occurred between 2100 and 0300 MLT. 
Calculating
the Integrated Auroral Power Rate of Change

We
examine the availability of Polar UVI LBHlong images
taken every 3 min for a period of 30 min centered at the
starting time of the fast flow events, then we have 11
corresponding Polar UVI images for each event. In real
situations, some images are missing or of bad quality.
We discard events with < 9 corresponding images. Also
we discard events with a change of field of view of the
Polar UVI camera which occurred during the events. 
We
integrate auroral power over a region of 6080
degree MLAT
and 20000400 MLT for each of the UVI images for the 30
min period of each fast flow event. We do not integrate
the auroral power for events with < 30% field of view
over the integration region. The estimation of the
auroral power can be used to evaluate the largescale
impact of the localized fast flow in terms of auroral
energy transport. 
To
quantitatively investigate the auroral energy change
during which fast flows occur, we fit a straight line to
all the auroral power calculated for each event by using
the leastsquares method. The slope of the straight line
is interpreted as the auroral power rate of change. 
A
total of 149 events are left through a series of event
selection criteria.

For
the scientific results derived from these events, readers are referred to
Shue
et al. [2003] (J.
Geophys. Res., 108(A6),
1231, doi:10.10292002JA009794) and Shue
et al. [2008] (J. Geophys. Res., 113, A02205, doi:10.1029/2007JA012456). 
Figure
1.
Observations of Polar Ultraviolet Imager
(UVI) auroral images and Geotail plasma and magnetic
fields for the 27 May 1997, 2249 UT event. (a) A
sequence of Polar UVI LymanBirgeHopfield long band images have been shifted to
the starting time of fast flows. The
auroral images with negative (positive) timings marked to the left show these images
were observed before (after) the first
fast flow occurred. A diamond on each image indicates the foot point of Geotail,
mapped by using the Tsyganenko [1989] model. (b) Auroral
power for each image is integrated and
plotted against timing. A vertical bar indicates the standard error of the mean of
the auroral power. Data points for
the auroral power are fitted to a red straight line by a
linear leastsquares
method. The slope of the line is marked on the top
of the plot. (c) Geotail data have been shifted to the starting time of the fast
flows marked by the vertical dashed line.
The horizontal blue line on
 Vx_{perp}
and ß
indicates a threshold
value of 300 km/s and 0.5, respectively. 
