🔭 astrophysics

Prospects of resolving and localising individual supermassive black hole binaries with pulsar timing arrays: the host ranking challenge

This paper simulates future Pulsar Timing Array capabilities to detect individual supermassive black hole binaries and proposes a novel host-ranking framework that, despite challenges from large localisation areas and incomplete galaxy catalogues, can significantly narrow down candidate host galaxies for multi-messenger follow-up observations.

Original authors: Niccolò Veronesi, Maria Charisi, Polina Petrov, Stephen R. Taylor, Jessie Runnoe, Daniel J. D'Orazio, Jacob Pilawa, Chung-Pei Ma

Published 2026-06-02

📖 5 min read🧠 Deep dive

CC0 1.0

Original authors: Niccolò Veronesi, Maria Charisi, Polina Petrov, Stephen R. Taylor, Jessie Runnoe, Daniel J. D'Orazio, Jacob Pilawa, Chung-Pei Ma

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe is a giant, noisy party where billions of black holes are dancing. Most of the time, they dance in such a huge crowd that we can only hear the general hum of the music—a "stochastic background" of gravitational waves. But occasionally, two massive black holes dance together so loudly and rhythmically that we can pick out their specific song from the noise.

This paper is about figuring out how to find the house (the host galaxy) where these two black holes are dancing, once we finally hear their song using a special listening device called a Pulsar Timing Array (PTA).

Here is the breakdown of their "host ranking challenge" in simple terms:

1. The Problem: A Needle in a Haystack

When our listening devices (PTAs) finally hear a specific black hole duo, they can tell us roughly where the sound is coming from. However, the "search area" they give us is huge—hundreds of square degrees of the sky. It's like being told a lost key is somewhere in a city the size of New York, but you don't know the street.

Inside this massive search area, there are hundreds of thousands of potential "houses" (galaxies). The challenge is: Which one of these hundreds of thousands actually contains the dancing black holes?

2. The Simulation: Predicting the Future

The authors didn't wait for the real event; they built a virtual universe.

The Setup: They created 1,000 different versions of the universe, each filled with realistic populations of black hole pairs, consistent with the background noise we already hear.
The Test: They simulated what our listening devices will look like in 20, 25, and 30 years (as we add more "microphones" or pulsars to our array).
The Goal: They picked the "loudest" black hole pair in each simulation and asked: "Can we hear it clearly enough to know where it is?"

The Result:

In the next 5 years, there's about a 38% chance we'll hear a loud enough pair.
In 10 years, that jumps to 51%.
However, being "loud" isn't enough. To pinpoint the location well enough to start a search, the signal needs to be very clear. With that stricter requirement, the chance drops to about 3.8% in 5 years and 14% in 10 years.

3. The Search: Cross-Matching with Galaxy Maps

Once they "heard" a virtual black hole pair, they drew a map of the search area. Then, they overlaid this map onto two massive digital catalogs of the entire sky:

The "Quiet" List: A catalog of normal, old galaxies (Early-Type Galaxies).
The "Active" List: A catalog of galaxies with active, energetic cores (Active Galactic Nuclei or AGN).

The Missing Pieces:
The authors realized their maps aren't perfect. Just like a street map might be blank in a foggy area, these galaxy catalogs have gaps where dust or too many stars block the view (like near the center of our own Milky Way).

They found that in many search areas, there are thousands of "missing" galaxies that aren't in the catalogs. If the real host is one of these missing ones, we might never find it without new telescopes to fill in the gaps.

4. The Solution: A "Host Ranking" System

Since there are so many candidates (about 190,000 quiet galaxies and 40,000 active ones in a typical search area), we can't check them all one by one. The authors created a ranking system to sort the list, putting the most likely suspects at the top.

How the Ranking Works:
Think of it like a detective trying to match a suspect's description to a lineup.

The Clues: The gravitational wave data gives us clues about the black holes' total mass and how far away they are.
The Match: The system looks at every galaxy in the search area and asks: "Does this galaxy's size and distance match the clues?"
- For Quiet Galaxies: They have good data on the galaxy's mass and distance. The system can easily say, "This galaxy is too small" or "This one is too far away," and cross them off the list. This method is very good, eliminating about half of the candidates immediately.
- For Active Galaxies: The data is fuzzier. We only know how bright they look, not exactly how heavy they are. It's like trying to guess a person's weight just by how loud their voice is. Because the clues are vague, the system can't rule out as many of these. It's much less efficient at sorting this list.

5. The Bottom Line

We are getting closer: In the next decade, we have a decent chance of hearing a specific black hole pair.
The search will be hard: Even with a clear signal, the search area will still contain hundreds of thousands of galaxies.
We need a filter: We can't check them all. The authors' ranking system is a tool to help astronomers prioritize. It tells them, "Don't waste time on these 50,000 galaxies; focus on these top 100."
The bottleneck: The system works best when we have detailed maps of galaxy masses. Currently, our maps of "active" galaxies are a bit blurry, making it harder to rank them. If we can get better data on the mass and distance of active galaxies, our ability to find the right host will improve significantly.

In short, this paper provides the instruction manual for the future hunt: it tells us how likely we are to find the signal, how many houses we'll have to check, and how to build a list that puts the most promising houses at the very top.

Technical Summary: Prospects of Resolving and Localising Individual Supermassive Black Hole Binaries with Pulsar Timing Arrays

Problem Statement
Pulsar Timing Arrays (PTAs) have recently detected a stochastic gravitational wave (GW) background, likely originating from a population of supermassive black hole (SMBH) binaries. The next major milestone is the resolution of individual binaries within this background, which would enable multi-messenger astronomy. However, a primary challenge is the limited sky localisation capability of PTAs. The 90% credibility localisation area ( $A_{90}$ ) for the first individually resolved binary is expected to span hundreds of square degrees, potentially containing hundreds of thousands of candidate host galaxies. Furthermore, existing electromagnetic (EM) galaxy catalogues are incomplete in regions of high stellar or dust contamination (e.g., the Galactic plane), creating "missing" host candidates. There is currently no standardized method to efficiently rank these vast numbers of candidates based on GW posteriors and EM properties to prioritize follow-up observations.

Methodology
The authors simulate realistic SMBH binary populations consistent with the GW background measured by the NANOGrav 15-year Data Set (NG15) using the holodeck software. They generate 1,000 realizations of binary populations and select the loudest binary from each population that is detectable in future PTA configurations but undetectable in NG15.

Three PTA configurations are simulated to represent future sensitivity:

IPTA_20: A 20-year baseline with 116 pulsars (similar to the upcoming IPTA 3rd data release).
IPTA_25 (Fiducial): A 25-year baseline with 158 pulsars (adding 42 new pulsars to IPTA_20).
IPTA_30: A 30-year baseline with 200 pulsars.

The authors inject the selected binaries into simulated timing data containing a stochastic GW background and use the QuickCW pipeline to constrain binary parameters (sky location, chirp mass, strain, frequency). They define "localisable" sources as those with a signal-to-noise ratio (S/N) $\ge 8$ , a threshold typically required for well-constrained sky positions.

To identify hosts, the authors cross-match the resulting $A_{90}$ localisation maps with two comprehensive all-sky EM catalogues:

ETG Catalogue: Contains $\approx 19.4$ million quiescent early-type galaxies (ETGs) with photometric redshifts and stellar masses (derived from 2MPZ and WISExSCOS).
AGN Catalogue: Contains $\approx 4.5$ million Active Galactic Nuclei (AGN) candidates (R90 catalogue from Assef et al. 2018) with W1-band apparent magnitudes.

A host ranking system is developed to assign a score ( $H$ ) to each candidate galaxy. The score is the product of the sky position probability ( $p_{sky}$ ) and the probability of the galaxy's EM properties matching the GW posteriors ( $p_{cat}$ ).

For ETGs, $p_{cat}$ is derived from the galaxy's redshift (converted to luminosity distance) and its estimated SMBH mass (derived from stellar mass via scaling relations).
For AGN, $p_{cat}$ is derived from the apparent W1 magnitude, which is a function of the binary's total mass and luminosity distance.
The authors explicitly calculate the number of "missing" hosts by estimating the fraction of the $A_{90}$ area not covered by the EM catalogues.

Key Results

Detectability and Localisation:
- In the next 0–5 years (IPTA_20), there is a $\approx 21\%$ probability of resolving a binary ( $S/N \ge 4$ ), but only a $\approx 0.3\%$ chance of achieving a well-constrained localisation ( $S/N \ge 8$ ).
- In the next 5–10 years (IPTA_25), the probability of resolving a binary rises to $\approx 38\%$ , with a $\approx 3.8\%$ chance of localisation.
- In the next 10 years (IPTA_30), the probability of resolution reaches $\approx 51\%$ , with a $\approx 14.1\%$ chance of localisation.
- Localisable sources are expected to be relatively nearby (median $d_{lum} \approx 400$ Mpc) and very massive ( $M_{tot} \gtrsim 10^{9.8} M_{\odot}$ ).
Localisation Areas and Host Counts:
- For the fiducial configuration (IPTA_25), 30 out of 38 potentially localisable binaries have $A_{90} < 2,000$ deg $^2$ . The median $A_{90}$ is $408$ deg $^2$ .
- These areas contain a median of $\approx 190,000$ ETGs and $\approx 40,000$ AGN.
- In IPTA_30, the median $A_{90}$ shrinks to $212$ deg $^2$ , reducing the median number of hosts by $\approx 40\%$ .
Incomplete Coverage:
- Incomplete EM coverage is a significant limitation. In the fiducial configuration, the median number of missing ETG hosts is $\approx 20,000$ and missing AGN hosts is $\approx 6,000$ .
- For some sources, the $A_{90}$ is partially or completely outside the footprint of the catalogues (e.g., overlapping the Galactic plane), making the true host unidentifiable without new surveys.
Ranking Efficiency:
- The ranking method successfully excludes galaxies with properties inconsistent with GW posteriors.
- ETGs: Approximately 47% (IPTA_25) to 61% (IPTA_30) of ETGs are excluded due to incompatible redshifts or SMBH masses. The ranking efficiency ( $E_{95-50}$ , the ratio of the 95th percentile score to the median) is $\approx 12.4$ for IPTA_25 and $\approx 17.2$ for IPTA_30.
- AGN: Only $\approx 0.3\%$ (IPTA_25) to $3.8\%$ (IPTA_30) of AGN are excluded. The ranking efficiency is significantly lower ( $\approx 3.7$ for IPTA_25), as the apparent magnitude is a degenerate combination of mass and distance, and the chirp mass is poorly constrained by PTAs.

Significance and Claims
The paper claims to provide the first realistic quantification of the number of candidate hosts and the impact of incomplete sky coverage for the first PTA-resolved SMBH binary. The authors demonstrate that while the number of candidates is large, a ranking system based on GW posteriors and EM properties can effectively reduce the search space, particularly for ETGs where stellar mass and redshift are available.

The study highlights that:

Multi-messenger discovery is plausible but challenging: The probability of localising a source within 10 years is non-negligible ( $\approx 14\%$ ), but the localisation areas remain large.
Catalogue completeness is critical: The existence of thousands of missing hosts due to survey gaps (e.g., the Galactic plane) poses a risk of false negatives in host identification.
Data quality dictates ranking success: The efficiency of the ranking method is heavily dependent on the availability of independent redshift and SMBH mass estimates. The current inability to efficiently rank AGN is attributed to the lack of redshift/mass data in AGN catalogues and the poor constraint on chirp mass by PTAs.
Future improvements: The authors suggest that future improvements in pulsar distance measurements (to utilize the pulsar term in GW signals) and the development of AGN catalogues with redshift and mass estimates would significantly enhance localisation and ranking capabilities.

The paper concludes that while the first multi-messenger detection is within reach, success depends not only on PTA sensitivity but also on the completeness and uniformity of EM catalogues and the development of efficient ranking pipelines to prioritize follow-up observations.

1. The Problem: A Needle in a Haystack

2. The Simulation: Predicting the Future

3. The Search: Cross-Matching with Galaxy Maps

4. The Solution: A "Host Ranking" System

5. The Bottom Line

More like this