One month's worth of Polar ultraviolet imager (UVI) data were subjected to
a "blob" analysis to determine the statistical dynamics of substorm feature
s observed in the Lyman-Birge-Hopfield long (LBWL) band (152-188 nm). Adapt
ed from similar DoD analyses of target images, the analysis consists of fin
ding, on a frame-by-frame basis beginning at substorm onset, the following
aspects of an individual auroral feature: peak power (i.e., power of precip
itating electrons), total power, centroid location magnetic local time (MLT
) and magnetic latitude (MLAT), and speed of centroid. Over 120 individual
auroral features were successfully acquired at onset and tracked until diss
ipation during January 1997. The power in the peak pixel and total power we
re random in time but displayed transient spikes that lasted 5-10 min. Over
the course of a substorm, the total energy of blobs averaged similar to 2.
0 x 10(4) GJ. A histogram of these energies suggests no preferred energy bu
t that lower energies were, more common than higher energies. Analysis of t
he blob positional dynamics generally supports a poleward and westward move
ment. During the course of a substorm, 90% of the blobs moved poleward, whi
le over 60% moved westward. However, these movements were not steady and di
splayed random components. Furthermore, a sizable minority (similar to 35%)
of the blobs moved eastward, which does not agree with the conventional pi
cture of auroral surges. Blob speeds varied from essentially zero up to sev
eral kilometers per second. However, during the January substorms the blobs
did appear to have a preferred speed of 0.84 +/- 0.34 km s(-1).