Difference in contour area between sky maps and circulars

From a discussion in the emfollow channel from @nicolas.arnaud : https://chat.ligo.org/ligo/pl/88j1iy1n83yrf8nnkyj34r8jpo

For https://gracedb.ligo.org/superevents/S240430ca, there is a one degree^2 difference (out of about 5,000) for the 90% area between https://gracedb.ligo.org/api/superevents/S240430ca/files/bayestar.png and the auto-generated circular text -- see https://gcn.nasa.gov/circulars/36353. At yesterday's SRC (UTC TZ, https://wiki.ligo.org/LSC/JRPComm/SRC20240506AgendaMinutes) it was decided that I would report that feature here to see if it is worth investigating/fixing.