The flare occurred in the northwest quadrant, as indicated in the following SXT image (click to enlarge).
The image was taken during the long decay of the flare. The cusp is clearly seen, which points to the west. The relation between the orientation of the cusp and the associated CME was surveyed here. But in this nugget, there seems something else that is unusual. It is seen in the following movie (click the blue thumbnail).
These frames are long-exposure images of the 5.2 arcmin field of view. The central part is heavily saturated, but expanding loops are seen toward west. An interesting thing that follows is the sudden contraction or collapse of outer loops. SXT observations of an inward motion in the form of blobs have been reported, which was seen in the early decay phase of a flare (see McKenzie and Hudson 1999.) But the timing of the contraction in this event is noteworthy.. See the following time plot (click to enlarge).
After the peak of 06:08 UT, there was a gradual increase in the light curves around 06:30 UT, and the contraction took place at the start of this increase (see the arrow). This increase is seen up to the 33-53 keV channel of HXT, so it should be associated with very hot plasma or nonthermal electrons. Similar gradual variations were seen earlier (between 04:30 and 05:10 UT), but the inward motion was not seen in SXT images. Does this inward motion or contraction mean anything to the reconnection process or is it visible only because of some geometrical effects? An interesting thing is that HXT image reconstruction completely failed for the second gradual peak, suggesting that a large structure is involved in hard X-ray emission.
20 November 1999 Nariaki Nitta (firstname.lastname@example.org)