Hello community:
LASAIR website sent me here for help. The big picture is that I’m trying to get information on past transients so I can plan my filters well. I’m not finding good information on famous, well-documented ZTF transient hosts. What I’m hoping for is the host galaxy is _____, the link to the full catalog info is _____, and the ____colored mark on the aladdin image is what we’re referring to.
Here is what I can actually get:
For the most famous transient I can think of, “The Koala”, ZTF18abvkwla,
LASAIR - No diff image, a broken link to a host catalog name (no response at SDSS search; no idea how to find that)
ANTARES-No host info or diff image
ALerCE - Yes, diff image!, confusing set of host matches
ATserver - no host info
OTHER ALERTS:
ZTF19abvizsw, AT2019pim - nothing ANTARES, LASAIR, ALerCE
ZTF20aajnksq, AT2020blt - nothing ANTARES, LASAIR, ALerCE
ZTF 21aaeyldq/AT 2021any - ANTARES covers, no host info (LASAIR does NOT cover)
ZTF 21aayokph/AT 2021lfa- ANTARES covers, no host info, (LASAIR has no difference image, claims no host)
So, where do I go to get good host info? What am I doing that makes LASAIR so incomplete compared to other brokers? Did LASAIR only get going after 2021?
Thanks for joining the Forum and using it to request assistance from the Lasair developers. As Forum admin, I’ve updated the title a bit to be more representative of the topic of host association and also moved this post to the Lasair sub-category, so they’re sure to see it.
I’ll let the Lasair representatives address any potential differences between what Lasair covers, with respect to other brokers (attention @roy).
But as far as the broken link to the SDSS host on the ZTF18abvkwla page in Lasair, that’s an easy fix. I can see that there’s a typo in the hotlink:
Perhaps I can also help to address the question “So, where do I go to get good host info?”. I work on supernovae too, and host association is a scientifically challenging endeavor. I’m not sure what ZTF provides in the alert packets in terms of cross-matches to static-sky sources for brokers to use. But I can say that in the future, LSST will provide the nearest three stars, nearest three galaxies, and the nearest low-z galaxy (DMTN-151). I think it’s also the case that most brokers are planning their own static-sky cross-matches, and that at the moment, this service is probably a work in progress for all brokers. However, especially at higher redshifts and for transients without spectroscopic classifications, host association can always have some level of ambiguity that probabilistic analyses will need to account for.
First let me say that Lasair is built for rapid response to emerging transients. The one you quote – ZTF18abvkwla – is over six years old, and a lot has changed with the ZTF survey and schema in that time, and there has been trouble with cutout images. I am sorry that we are not aware of what you think is a better name (Koala): we do not know which transients are “famous” or “well-documented”.
I have made you a sample filter here which finds recent bright supernovae that are in TNS, with full information about the host galaxy. You can do “Run Filter” for results, or “Duplicate Filter” to make your own version. There is more information about building filters here.
But perhaps you want supernovae coincident with a specific set of galaxies that you already know? You can make a “watchlist” of up to a million, build a filter, and have coincident alerts streamed to you, see here
Hello: regarding the host information: OK, yes, when I applied your simple fix, I was brought into a fabulous world of information i didn’t know about - thanks.
But troubles with LASAIR host info continue.
AT2023fhn - I cannot make heads or tails or missing space from the sherlock association "
In fact, I cannot even get a position of the host to check it myself. Why don’t I just look at the survey displayed on the page through Aladin, and search that? Oops, it doesn’t say what the survey is - when I click on the “stack” it is not listed.
OK, another one: [AT2023vth] {the first and ONLY one with a diff image! Yay!)
Sherlock: " The transient is possibly associated with 117652691427871888 ; an r=21.31 mag galaxy found in the PS1 catalogue. Its located 2.95" N, 1.90" E from the galaxy centre."
-There was no link this time.
So, I went to the pann-starrs search page, PanSTARRS Catalog Search and there isn’t really a way to search on that ID number. So, I searched on the coordinates… and that failed too.
So, I am 0% success for hosts, even when sherlock seems to find something. It would be really really helpful if LASAIR could provide a working link, or if not, clear information so I could find host info myself. Or, more hints on how to get that info out of the page?
Thanks for your honest feedback. Feel free to send us more examples. We need to hear from dissatisfied users as well as satisfied ones, and we’re keen to address any problems if possible.
Here are a few more responses to your questions…
ZTF21aaeyldq is a singleton alert. Although the data is in our backend database, we don’t deliver search output for this object because at the time we didn’t track objects with just one detection, on the assumption that these were either solar system objects, artificial satellites or noise. (This policy has changed, though I’d need to double check exactly when that change was made. Possibly mid 2022.) ANTARES states the number of alerts is 24, but if you examine the lightcurve you will see only one actual detection (Num. Mag. Values: 1).
ZTF21aayokph is in Lasair. It claims no host because there is (in the current list of catalogues we search) no viable host. Note that ANTARES also claims no host. Are there brokers that identify the host? Feel free to zoom out on the Aladin-Lite widget. I can’t see any viable host there in the Pan-STARRS imagery. There may be a host in a deep catalogue that is not part of our search suite at the moment.
ZTF emits both public and PRIVATE alerts. Lasair is only able to consume the public alerts. If it’s not in these 3 brokers, it was probably a private alert. Note that in BOTH cases, the reporting group on TNS is ZTF. Feel free to point us to a broker that contains the info.
The “stack” error you are seeing in the AladinLite widget is a known issue. It was fixed in the LSST implementation of Lasair, but it has not been backported into Lasair-ZTF. A fix would also help you to see the the Pan-STARRS catalogue info. The complaint about Pan-STARRS not finding the object ID is a problem for the Pan-STARRS consortium, not Lasair.
As for the presence of image data. A flaw in the way the filesystem is able to handle the image data means we had to remove links to old image data. As new alerts come in for old objects the images get rewritten. It would be useful to us to know what exactly you need the image data for. (LSST will not provide cutouts at all except for a tiny 6 arcsec stamp around the object.)
Obviously we will do what we can to fix broken links. Note that we are currently working on both Lasair-ZTF (legacy) and Lasair-LSST (not yet released, since there is no LSST data yet). Priority is always given to addressing fixes in Lasair-LSST, but where the backports are straightforward we will endeavour to apply them.
Finally, I should point out that Lasair has been operating since mid 2018.
Thank you very much for your reply - I really learned something. In fact, my jaw dropped on the floor when you told me there are such a thing as PRIVATE alerts! How is that possible? Doesn’t that fundamentally violate something in the NSF funding contract?
Just on the technical aspects: ZTF subtracts the images, but I believe you are saying that ZTF also declares the difference image sources detected, i.e. declares an alert. So, I suppose they could set a public threshold of 3.5 sigma (and e.g. a certain number of connected pixels, etc.) and a private one of 3.0 sigma, for example, so the broker users would never see some number of alerts.
On the face of it, however, this is completely nuts. Why wouldn’t it be the data brokers job to provide the filtering facilities so that each investigator could decide on their own alert criteria?
Since a huge fraction of the sources I’m interested are private, I need to know everything about this. Who do I talk to at ZTF to get the details on this? What could I really be missing, and how do I find out?
Hello Again, I have remaining questions on the host information and association for at2023vth:
First though, please let me offer an apology: Ken said in one post, “We need to hear from dissatisfied users as well as satisfied ones,”. I hope that everyone understands that these are meant as requests for help, not negative comments. I hope I don’t come across as “dissatisfied”. It is easy to criticize something, find imperfections, but very difficult to create something. I recognize this, and brokers are a truly monumental task with wonderful features that reflect a lot of hard work by a lot of folks. Thanks for that! I just want to do what I can to take advantage of the good stuff here to further my project, so I must hunt these issues down.
Allow me to summarize this discussion on host information (and this is, of course, from the viewpoints of my needs):
*In one case that meant a misformatted link (koala), and so the fix is:
-check for obvious formatting issues (spaces surrounding host ID links); Thanks to Melissa Graham
*In one case at2023vth Sherlock gave a host ID for PS1 that I could not figure out how to access: Sherlock: " The transient is possibly associated with 117652691427871888 ; an r=21.31 mag galaxy found in the PS1 catalogue." When I went to the Pan-STARRS page a search on this number I got nothing. When I searched in a cone search, careful to search PS DR1, I found that same ID number, but the r was off by .4 mag (21.72, not the 21.31 mag sherlock said), so I was never sure I had the right host.
UNSOLVED:-------------> What did I do wrong that I got the host with the wrong brightness?
Can you give a straight-up procedure for accessing the catalog info on associated hosts? a) interactively, and b) non-interactively?
I did more digging: Sherlock: “The transient is possibly associated with 117652691427871888; an r=21.31 mag galaxy found in the PS1 catalogue. Its located 2.95” N, 1.90" E from the galaxy centre."
So, I think that means the transient (OT) is located ~ 3" N and ~2"E of the host galaxy center.
The LASAIR aladdin display shows a pink cross/circle identified as the transient, on top of a big, fuzzy, elliptical shape, obviously a galaxy. A green square down and left of the OT pink cross/circle, 2 FWHM off the center of a fainter, smaller round galaxy, is identified in the window as the associated host, consistent with the distance to the host given.
big galaxy r=21.09, PS1 ID 117652691437722372; the OT is on top, vaguely 0.75" NE of my by-eye center.
little gal r=21.73 PS1 ID 117652691427871888; This is the correct distance relative to the OT.
Sherlock gives the wrong r magnitude for its associated host, and ignores the big galaxy that the OT is on top of. Obviously Sherlock is complicated, and I went over the description on our web pages, there is more to it (maybe it rejects the big galaxy on the basis of colors or distance; maybe what I think is a galaxy is a mess of otherwise known points souces?). I could not find a distance for any of these objects, nor classifications, which are not in the PS1 data in Vizier.
Given the obvious nature of the bigger galaxy, also listed in PS1, does this sound like Sherlock acting like you expect it? (If so, any insight there would be welcome.)
UNSOLVED-----------------> Sherlock’s association looks crazy; is it me or him (it)?
Thanks,
-Bruce
PS
It seems that if you move the Aladdin Pan STARRs locations North about 0.8", W about 0.1", they work much better. The same locations are indicated on your LASAIR page and an independent vizier display. So that’s likely a vizier/aladdin issue.
It’s possible to use the Lasair API to recompute the full Sherlock crossmatch record for a particular object of interest which will give you all the possible associations rather than just the top one. The process is described here: Lasair API and Client
You can do this for ZTF23ableqsp/AT2023vth using a command like this:
I won’t list the full output here because it’s quite long, but it does give your preferred host of 117652691437722372 as its second possible match. I don’t know enough about Sherlock to say why it ranks them the way that it does.
We don’t process and store this information for every transient, mostly for efficiency reasons, but you can get it through the API if you want it.
