ASTs in regions with different exposure times and/or filters

[lea-hagen]

After conversation with @karllark and @k-gilbert today, we have determined that a full set of ASTs is necessary for each region that has a different selection function (and therefore noise model). For simple cases like METAL, where one field of view is slightly smaller, that's just two sets: one in the middle and one around the border. For LUVIT, there are more Frankensteinian (TM) arrangements, with overlapping WFPC2 pointings, offset UV and IR, etc. Each of those combinations of overlaps will need its own ASTs.

Also relevant to these, as well as PHAT, is the chip gap (and its many positions). They're fairly obvious in PHAT MegaBEAST maps, indicating that they need their own noise model.

[benw1]

Exposure Map of one half-brick run. All six bands are here. The number of exposure levels is large, perhaps making it necessary to take the map into account when running ASTs. Areas with significantly different exposure will have different completeness and different noise parameters.

[benw1]

Old stellar density map of part of Brick 2. The chip gaps and corners of the IR detector show up.

[benw1]

Same as above for new photometry. Now the IR detector edge overlaps are visible, along with the exposure effects, but the chip gap is no longer apparent.

[benw1]

Stellar density map for part of Brick 22 from old photometry.

[benw1]

Same as above for new photometry. Now the IR detector edge overlaps are visible, along with the exposure effects, but the chip gap is no longer apparent. The IR detector edge effects are not as strong in the less crowded regions.