The detection of gravitational waves and light from a single source was one of the most important discoveries of the gravitational wave era. Can upcoming data from Vera C. Rubin Observatory help astronomers repeat that feat?
These images from the Hubble Space Telescope show the fading light of the kilonova associated with the gravitational wave event GW170817. [NASA and ESA Acknowledgment: A. Levan (U. Warwick), N. Tanvir (U. Leicester), and A. Fruchter and O. Fox (STScI)]
Not Yet Repeated
When two neutron stars merge, the collision sends ripples through spacetime and generates an electromagnetic signal called a kilonova. In 2017, researchers detected gravitational waves and kilonova emission from colliding neutron stars — the first and, to date, only electromagnetic signal definitively associated with a gravitational wave event.
In a research article published this week, a team led by Simon Stevenson (Swinburne University of Technology; OzGrav) considered how the much-anticipated Legacy Survey of Space and Time (LSST), carried out by Rubin Observatory, can enhance our ability to detect kilonovae and pair them with their gravitational wave counterparts.
Survey Simulations
When a new gravitational wave signal is detected, researchers rush to search for an electromagnetic counterpart to the signal. While Rubin Observatory can aid this type of reactive search, Stevenson’s team focused on a different strategy: spotting kilonovae during routine survey operations and using these detections to trigger a targeted hunt through gravitational wave data, searching for signals missed by automated detection algorithms.
Over the 10-year course of LSST, Rubin will scan the visible southern sky every few days, uncovering a wide variety of transient sources such as supernovae, novae, and tidal disruption events. Because Rubin will amass a huge amount of data each night, astronomers worldwide will use data brokers to pass along the most promising signals. Stevenson’s team used for their analysis a bespoke and modular data broker called Fink. Fink currently handles 200,000 alerts per night from the Zwicky Transient Facility and will be scaled up to handle 10 million alerts per night as Rubin requires.
Examples of simulated kilonova light curves for different redshifts and models (Kasen and Bulla). Click to enlarge. [Adapted from Stevenson et al. 2026]
Searching for Signals
Though Rubin is a transient-tracking powerhouse, it’s not ideally suited to discovering kilonovae specifically; these events evolve so rapidly that Rubin may only be able to collect a few data points before the event fades from its view. To gauge Rubin’s efficacy as a kilonova detector, Stevenson and collaborators simulated kilonova light curves and identified detections with a signal-to-noise ratio greater than 5; these detections will be passed to the data broker. They found that Rubin will detect about 42 kilonovae per year, but only a few will have a strong enough signal to be passed to the broker.
For the handful of detections that are passed to the broker, what are the prospects for tracking down the associated gravitational wave signals? Due to the survey cadence, most kilonova detections occurred 1–2 days after the gravitational waves from the neutron star collision would have reached Earth, but delays of up to 5 days were possible. This means that searches for gravitational wave signals must sift through multiple days of data — a computationally intensive prospect that might require the development of new search techniques. The computational requirements of this search are increased by the possibility of contaminants (signals misclassified as kilonovae and passed on by the data broker), which will require additional followup.
Despite these challenges, it’s clear that Rubin’s upcoming detections of kilonovae will advance our study of these rare events. With LSST expected to get underway early this year, the next multi-messenger signal may be just around the corner.
Citation
“Strategy for Identifying Vera C. Rubin Observatory Kilonova Candidates for Targeted Gravitational-Wave Searches,” Simon Stevenson et al 2026 ApJ 998 8. doi:10.3847/1538-4357/ae244e