M33 / HST13706 BOT clearing, timing, and diffuse object investigation

We take a maximally conservative approach, under the assumption that the sources are non-varying in time and the GALEX FUV NGS image is accurate. The GALEX beam is ~6". The NGS image we have (much deeper than the standard AIS GALEX image) has 1.5" pixels. If we gather all the flux in 5x5 (7.5"x7.5") areas, for every possible pixel center, we can assume the worst-case hypothesis that all of this flux comes from a single point source that will fall into the COS aperture. In this case we calculate the fraction of the PS flux inside the 7.5"x7.5" pixel to be about 73% of the total flux, so we multiply our final results by 1/0.73. We then look in an area around each source 100'' on a side. At each 1.5" pixel we add up all the counts in the 5x5 (7.5"x7.5") area surrounding it, and convert from, counts to flux using this GALEX webpage, to get a map of maximal flux to COS for each position within the 100" box. We then report the maximum of that value within each box as the most flux one could conceivbly get into the COS aperture, if one pointed the telescope in the worst direction within that box.

The code for this procedure can be found here

Each of these 7 objects has a FUSE spectrum of realtively high qualtity. They can be seen here:

All have realtively flat spectra, such that extrapolating from 1180 angstroms to the COS wavelengths is not unreasonable.

Neither of these represents a good guess as to the actual flux from the source. The BOT clearing method is taking a number of maximal assumptions in the interest of safety. The FUSE fluxes are a more reasonable upper limit, but in the case of diffuse structures are likely not at all a good estimate of the flux. This is because the aperture is much, much larger (30"x30") than COS's.

To get a our best estimate we use the same NGS GALEX FUV map of M33 as described above. We then place FUSE apertures on each position, as well as COS apertures. We then take the ratio of the two to get a guess as to how much decreased the COS flux should be. In regions with lots of diffuse flux, we get very high ratios of expected FUSE to expected COS; in regions that are dominated by a central source, we get more similar answers. We use the fact that UIT-236 has both COS and FUSE spectra to give ourselves a check -- we find we do pretty well at the 50% level; i.e. the ratio of COS to FUSE is pretty similar to the COS-aperture-on-GALEX to FUSE-aperture-on-GALEX.

The code for this procedure can be found here