For how long will forced photometry be run on DIAObjects, after discovery?


(Stephen Smartt) #1

Question on how forced photometry will be run on images taken after a DIAObject is detected. Again, using LSE-163 Data Products and Definitions, Juric et al. 2019-07-29. - section 3.2.1 states :

“For all DIAObjects overlapping the field of view … forced photometry will be performed on the difference images. Those measurements will be stored as DIAForced- Sources. No alerts will be issued for these DIAForcedSources, but the DIAForcedSource measurements will be included in any future alerts triggered by a new DIASource at that location.”

I take this to mean that a DIASource which is found with through the DIAForcedSource photometry with S/N>5 will have an alert issued.

But what if the DIAForcedSource produces a measurement S/N < 5 ? Where will this information be stored - the Prompt Products Database ?

If so, then 3 questions

  1. on what timescale will it be available, 24hrs ?
  2. For how long will such forced photometry be run ? The whole survey ?
  3. Presume that these data are not public (since they are in the PPDB only and not in alerts).

A strong science case for accessing forced photometry after detection of a transient is fast declining transients e.g. kilonovae, NS-WD mergers. A detection followed by a non-detection is often as interesting as the other way round.


(Еrіс Веⅼⅼⅿ) #2

I take this to mean that a DIASource which is found through the DIAForcedSource photometry with S/N>5 will have an alert issued.

Not quite. Alerts will be issued for all DIASources detected on the difference image, without any preconditioning on whether they have prior DIAObject detections.

Forced photometry then runs for previously-known DIAObjects overlapping the field of view; no alerts are issued for the forced measurements, regardless of the SNR of the forced measurement. They are simply stored in the PPDB.

  1. on what timescale will it be available, 24hrs ?

Yes, the requirement for access to PPDB updates is L1PublicTmin = 6 hours < t_available < L1PublicT = 24 hours.

  1. For how long will such forced photometry be run ? The whole survey ?

This is currently undefined; the ticket to resolve it is DM-15605. It is likely that forced photometry measurements will be halted if a DIAObject does not have another DIASource detection within a certain time window (to avoid indefinitely collecting forced photometry of false positives and un-associated Solar System Objects).

  1. Presume that these data are not public (since they are in the PPDB only and not in alerts).

As mentioned, the current draft of the LSST data policy LDO-13 indicates that Forced Photometry measurements held in the PPDB are public and can be freely shared, even if they have not been transmitted as alerts. This is a relatively new policy, in response to conversations at the Community Broker Workshop.


(Stephen Smartt) #3

Very good. Understand all now.

I’ve looked at some of the useful info we extract in our ATLAS lightcurves (we continue to force for all “Good” objects, so no false positives, no time window). As @ebellm notes in the ticket DM-15605 , 30 days is a reasonable proposal. Looks about right from our low redshift z objects in ATLAS - after the last 5 sigma detection, very little real information after 30 days (probing ~1 mag difference, getting to ~2 sigma). I would say the 30 day post-discovery window is a good proposal (30 days after the last 5 sigma detection … in any filter).

Ideally, preserving the time window for higher z objects (time dilated) would be great. 50 days would help (hence preserving to z ~ 0.7).

So 30 days is good (from my quick assessment of low z transient samples), if you can do 50 days even better and there is some justification for that. More quantitative justification could be done, but don’t think we would end up with a different answer. Just make sure the “in any filter” is applied, since that will catch the high-z (red) objects.