Velocities of Morning Sector Patchy Auroral Forms

Using 6 years (1993-1998) of CANOPUS Gillam All-Sky Imager (ASI) 558 nm data we identified 80 occurrences of morning sector patchy aurora. For these events, we calculated the velocity field for each ASI image frame using a machine vision optical flow technique. An optical flow computation determines the velocity field of an image based on changes in the brightness pattern between two consecutive frames separated by a small time interval, and is subject to the constraints that (1) the brightness of a given object point remains relatively constant over time and (2) only small motion changes occur between two consecutive images. Only patchy auroral data that reasonably fit these two constraints, and for which the optical flow field represented the motion reasonably well (37 occurrences which produced > 1700 vector fields corresponding to > 4.4 x 10⁵ velocities) were considered. From the 16 x 16 vector fields obtained in the optical flow computations, the average vector of the inner 4 x 4 vector field (inner 64 x 64 pixels of the ASI image) was computed for each image. Average velocity occurrence statistics show that auroral patches flow eastward with speeds on the order of ∼ 55 m/s. We present the results focussing on a comparison of these motions with convection on a statistical basis, and where possible on an event basis. Our working hypothesis is that these auroral features are due to co-existing hot and cold plasma populations in the trough region. The hot electrons are precipitated by wave particle interactions in the vicinity of the cold drifting ``blobs'' of plasma, possibly of plasmaspheric origin. Establishing whether or not these patches move with the convection velocity would be valuable, both in terms of identifying the source mechanism for this type of precipitation as well as providing an additional means of remote sensing convection.