Swift Supernovae -- What are bad targets?

Which supernovae make good candidates for Swift observations is a very science-dependent question and has different answers to different people.  So I will start with things that make SNe poor targets for Swift as they are independent of the science.




Too faint:  if we can't detect it, are the upper limits obtained still useful?  This depends on the SN type/intrinsic luminosity/distance/extinction.  This has roughly been translated into a distance limit of z<0.03 for SNe Ia, but UV bright, luminous SNe II can be observed to farther distances (even z~0.2) 

High extinction: A_V >~ 0.5  This is just a rough guide, as it also depends on source brightness/distance/exposure time.

UV wavelengths are a sensitive probe of extinction.  There is a certain range of reddening which is significant enough for the UV to constrain the amount and wavelength dependence of the reddening.  But too much and we don't detect it at all and the expected brightness may be so much fainter than our upper limits that they don't constrain anything.  This applies to both MW and host galaxy reddening.  The MW reddening can be known ahead of time via NED.  The host galaxy reddening isn't usually known in advance, but if it is described as heavily reddened based on the optical data, than it is probably already too much for the UV.  This is of course distance and SED dependent.  A very nearby type II SN can be more heavily reddened than a distant Ic and still be detected.

Contamination:  while galaxy light can be subtracted off, if the underlying galaxy is too bright it contributes to the coincidence loss in a way that may not be correctable.  The cutoff value I use is 5 counts/second within a 5" aperture.  That value may not mean anything to you if you don't do UVOT photometry, but that threshold is frequently crossed for very nearby SNe (z~0.005) within their host galaxy.  Because of the lower throughput of the UV filters it rarely affects the UV light.  So severe galaxy contamination can be a very bad thing, but we should also be careful of moderate galaxy contamination because we never know how long Swift will be around and if we can get template images to subtract the galaxy.  For good targets we shouldn't not observe them because of that fear, but if we are choosing between two possible targets or creating selection criteria for a sample for which we can be choosy, then low galaxy contamination is preferred.

Nearby Star Contamination: similar to galaxy light, a nearby star can contribute to coincidence loss in a way that is hard or impossible to correct for.  This was the case for SN2005am.  Avoid m<15 stars or galaxy nuclei within 8" or so.  Most stars are fainter in the UV so this becomes less of an issue.   If the UV data still looks useful, we will usually continue taking data in the optical filters as well (sometimes with different modes to ameliorate some of the problems) in case they can be fixed in the future.  Some data will turn out to be useless, but for transient objects there is no way to go back and take data you skipped the first time around.