Lasair: The Transient Alert Broker for LSST:UK

Current and next-generation time-domain surveys will produce more data than ever before. The Zwicky Transient Facility (ZTF; Kulkarni 2018; Bellm et al. 2019) generates up to ∼10⁶ alerts³ (individual photometric data points of time-variable sources) per night (Patterson et al. 2019), while the Large Synoptic Survey Telescope (LSST; LSST Science Collaboration et al. 2009) is expected to produce up to ∼10⁷ alerts per night. However there are significant computational challenges in processing these data into a digestible format, in near real-time, in order to identify interesting transients and trigger follow-up such as spectroscopy and observations in other wavelength regimes. To this end, we present Lasair,⁴ the transient alerts broker for the LSST:UK collaboration. In preparation for LSST's data stream, Lasair ingests the ZTF public alert stream into a relational database, assimilates the alerts into objects, and produces lightcurves and reliable cross-matches to star and galaxy catalogs. Lasair can be viewed and queried through a web browser⁵ and we provide simple example streams of interesting objects, as well as access to a full SQL search engine. Registration to the website is optional, free, and open to all.

Lasair provides a user-friendly interface to access public ZTF transient alerts. The alerts are transmitted by ZTF typically within 13 minutes of the exposure, in Avro/Kafka format. They are ingested into the Lasair database (on hardware in Edinburgh) within 20 minutes.

The online database contains three main SQL tables. The candidates table contains photometric data from ZTF for every alert. The objects table contains metadata for groups of candidates with the same object identification number, assigned by ZTF to all candidates whose positions agree within 15. This metadata includes minimum, maximum and average magnitudes, earliest and latest dates of detection, and mean coordinates. We create an object when there are three or more candidate detections, in order to remove moving objects and reduce bogus detections. Therefore all the ZTF object metadata is stored in the objects table only if there are three or more detections within 15. A third table (sherlock_crossmatches) contains "value-added" contextual classification information provided by Lasair using the software Sherlock.⁶ This consists of nearby sources from a variety of catalogs, with their corresponding photometry, and spectroscopic or photometric redshifts. Sherlock uses star/galaxy separation, distances, and galaxy offsets in order to classify objects as likely supernovae, nuclear transients, or variable stars. Sherlock also matches against known AGN mostly with the Veron (Véron-Cetty & Véron 2010) and Milliquas (Flesch 2015) catalogs and with CV catalogs (Downes et al. 2001; Ritter & Kolb 2003). Any stationary, transient source which is not associated with a cataloged star or galaxy is classified as an "orphan." Lasair provides a reliable way to remove (or select) known variables. We expect that Lasair will provide similar information in the LSST era. Lasair also maintains an up-to-date crossmatch with the IAU Transient Name Server,⁷ so that known transients can be identified in the ZTF stream.

Figure 1 shows the Lasair page for a ZTF object. Each object page displays the ZTF lightcurve and astrometric information, followed by contextual information from the objects and sherlock_crossmatches tables, as well as user comments. Next is a ranked table of likely cataloged crossmatches from Sherlock and an interactive AladinLite display (Bonnarel et al. 2000; Boch & Fernique 2014) of the region. Finally, the relevant candidates information is displayed for every light curve point.

Zoom In Zoom Out Reset image size

Forthcoming functionality will include the highlighting of ZTF transients in high probability regions of LIGO-Virgo gravitational wave skymaps.

The front page of Lasair provides several options to access the data (see Figure 1). The tables can be searched and joined using normal SQL SELECT queries, with an SQL form builder supplied.

Useful SQL queries can be stored by registered users, and either kept private or made public. The latter queries are called "streams"—useful substreams of the data that the Lasair team and Lasair users have provided. Examples of streams are:

• 1.  
  Nuclear transients and TDE candidates: Objects near the core of an inactive cataloged galaxy, rejecting Pan-STARRS stellar matches.
• 2.  
  Supernova candidates: all Objects that are not classified by Sherlock as variable star or AGN or CV, and are not coincident with a Pan-STARRS stellar source.

Lasair provides a simple, single object, cone search by sky position or ZTF object ID.

Often, a scientist would like to keep track of a set of sources, and is most interested in any transient activity associated with these (e.g., to monitor specific AGN activity). Lasair users can input lists of up to a few thousand sources ("watchlists"), which are stored with their account, and ask for a crossmatch at any future time—like a multiple cone search. Watchlist owner alerts are under development, reducing to minutes the time from observation to astronomer alert.

Experienced users can perform more advanced analysis on the results of their queries using Jupyter notebooks, with several examples provided. They can, for example, combine ZTF data with external resources such as machine-learning analysis of light curves.

For acknowledgements, see https://lasair.roe.ac.uk.