Review and improve offering on LIGO-Virgo and Fermi-GBM skymaps

[smarttgit]

We are nearly there in making this scientifically useful, but could do with a few enhancements. At present nobody uses it as it not scientifically functional, but with some effort this could be a good feature for V2.0 release

https://lsst-uk.atlassian.net/wiki/spaces/LUSCSWG/pages/128516170/UK+Science+requirements+for+the+transient+server+Lasair+for+Level+1+Alerts

This thread will document what we want

[smarttgit]

https://lasair.roe.ac.uk/skymap/

This is a useful compilation list. Could we make it into a a table instead, with the rows being the events (as listed) and each of them having 5 columns. With the info propagated from the information on each of the objects pages. We have this info and we display it, we just need to put in this top table.

Distance ISO Date: Julian date Percent probability max prob RA,Dec

This will give uses a rapid overview

Note that we need to include "Terrestrial" probability. For some reason that is not displaying

[smarttgit]

On each skymap page :

Need a button to download an ascii file of the Glade galaxies contained within the 90% contour. Can it return the following : Name, RA, DEC, redshift or distance , angular size, magnitude, absolute magnitude, probability contour within which the galaxy lies, and the probability ranking (so, yes 2 columns of probability).

Note the 2 probabilities requested here :

The first is straightforward (Note that QUB (Dave Young) has code for the contour calculation if you need it. This is NOT the probability in the healpix pixel at that position.

The second requires a weighting algorithm of some sort.

The best version of the probability ranking calculation takes into account the galaxy mass (luminosity). Basically an algorithm like this: https://github.com/Lanasalmon/HOGWARTs

[smarttgit]

Can you add another Mollweide projection (done with Aitoff), with RA and DEC grid lines (degrees please!)

And the clickable ZTF coverage should be boxes the size of the camera - not the size of the number of transients in the field. Currently it is pretty but not useful

Fine to keep this colour image, but we want something else to show coverage with RA and DEC marked. Stephen will post examples.

[smarttgit]

The tool to add the ZTF transients that are -X to +Y days from trigger is very useful. And nice to see them pop up. But it must be changed to OBJECTS, not candidates

Currently it selects any candidate with any detection in that time window - which is not useful. We need to select only OBJECTS which have their FIRST detection within that time window.

And as well as the plot, we need to return a table of these OBJECTS

We either need to remove the Sherlock classified stars VS and AGN or give buttons to say "remove VS objects" and/or "remove AGN"

In the Table, give the probability contour within which they are contained, and Sherlock redshift (that helps to filter the objects by likelihood) and the Sherlock type

[smarttgit]

Is it possible to add functionality to plot specific points on the sky map? E.g. by providing a list of coordinates via text upload. Would be very useful for plotting specific known candidates – especially since normally, the candidates produced by ZTF for most GW events are not released as public alerts but are released in GCNs.

[smarttgit]

Stephen to write to ZTF to see if we can get an LSST dry-run type of trigger. They will not do this for NSBH or BNS events, but may do it for us a BBH.

For example, ZTF trigger their public telescope time on a BBH and tile out the region, we test how fast we can get the objects into Lasair and make all of the above functionality work. The top level quesiton is how fast can a user access objects once a GW event has had the telescope triggered. I think ZTF might do this for us as a test, with a BBH.

[RoyWilliams]

(1) Many GLADE galaxies do not have redshift, but this seems to be an essential component. Should we just ignore all those, so that every galaxy has a redshift? If not, how do we effectively marginalise over redshift?

SJS reply :

If they fall within the 2D skymap but have no redshift then list them, but no redshift or absolute mag

(2) Which ZTF transients to show? -- objects not candidates -- objects with jdmin from -X to +Y relative to the event -- sherlock join with no VS, BS, AGN, CV

SJS reply :

Yes

(3) Terrestrial probability. The LIGO people keep adding more classifications. I can add this one easily.

(4) Coverage map. I can add a box for each field showing its real size.

(7) The probability that I listed on the web page is relative to the sum of the probabilities of all the other GLADE galaxies. Foe each one, we compute the 3D skymap probability, then normalise by the sum of all of these. Foe example https://lasair.roe.ac.uk/skymap/GW170817/ shows the correct galaxy (NGC 4993) with probability of 9.11%.

SJS reply :

This is good, also useful to add this second probability of the contour within which the object lies. This would be useful for both objects and galaxies.

(8) (a) We already have Aitoff projection. But the contour lines wrap round sometimes and it looks bad with horizontal lines (ask Ken). (b) I have asked Thomas Boch about making a RA/Dec grid on the Aladinlite and have not received and answer.

SJS reply :

OK, let us (QUB) give you some examples

(9) We can make a table of possible ZTF transients, but what is the sort order? How to find the most likely ZTF transient? The one with the highest 3d skymap probability, like the galaxy list?

SJS reply :

Highest 3D skymap probability would be ideal, but not all objects will have a redshift, since Sherlock may associate them with a galaxy without a photo-z or spec-z. So yes, I think IF a redshift is available from Sherlock, then then report it and compute the 3d probability and rank them that way. If no spec-z or phot-z then list these at the bottom. We could have two columns one is 3D probability (which requires redshift of host of the object) and the other is just the 2D. Hence any transient with no z (from any source) is listed after the 3D probabilities, and those are ranked by 2D probability contour.

[mnicholl]

Agree with all Stephen's suggestions. To me, the most useful version of this page would have two tables right below the sky map. Both should have download buttons.

For the first table : Objects table (ONLY for objects with jdmin > LIGO alert), ranked by probability. At least the following columns: ZTF name | TNS Name | First detection (MJD) | most recent detection (MJD) | peak g | peak r | g-r colour | Absolute Mag | Sherlock host name | Sherlock host z | Specz or Photz flag | 3D probability | 2D probability | GW Abs Mag

• Quote (g-r) at peak, or close to it when both magnitudes have less than 2 day difference in time (e.g. taken 1 night apart).
  For the first Absolute Mag (normal Abs Mag) : if Sherlock has a Specz or Photz then calculate the Absolute Mag, including foreground extinction (use the brighter of g or r) ** Two probabilities - see comments above, 3D for those objects with a Sherlock redshift, 2D for those without. Needs to be two separate columns. For the second "GW absolute mag", use the LIGO distance at that healpix position. Use the brighter of g/r. Quoting 2 absolute mags will be very useful to immediately check for consistency. This "GW absolute mag" can be calculated for everything, not just those with a redshift.

For second table : GLADE galaxies, calculated immediately the GW map is received (see above).

(Note - I deleted the requirement for transients appended to the Glade table, since the Glade Table will be computed within minutes/seconds of the GW trigger and there will be no transients at that stage. We would need to be constantly updating the Table, but I don't think that is necessary, since the transient object selection listed above will dynamically do that).

[tms-epcc]

As agreed at the 28/APR/21 Lasair internal review, @andyxerxes has written a report for the A3.09 activity. At the above meeting it was agreed to close all tickets in the A3.09 current activity.