In 1994 Wendy Calvin of USGS Flagstaff identified these bands as being due to condensed molecular O2 (Spencer et al., 1995):
The condensed O2 is too volatile to be exposed on Ganymede's surface, and is most likely trapped in bubbles within the surface water ice. A likely origin is dissociation of surface H2O by Jovian plasma bombardment, followed by escape of the hydrogen (Calvin et al. 1996).
Since then, we have used FOS and WFPC2 on the Hubble Space Telescope to map the O2 distribution on Ganymede, finding that the absorption bands are concentrated at low latitudes (Calvin and Spencer 1998).
The low-latitude, trailing-hemisphere concentration of the O2 bands is not well understood, particularly in light of the discovery of Ganymede's magnetic field, which should protect the trailing hemisphere from low-energy particle bombardment.
Recent observations have shown that Saturn's icy satellites do not show the O2 band (Spencer 1998), despite the detection of O3 on Rhea and Dione by HST (Noll et al. 1997).
More recent HST/STIS (1999) data which map the strength of the O2 band across most of Ganymede's surface are still being analyzed.
We have also discovered very weak O2 signatures on Europa and Callisto (Spencer and Calvin 2002).