Semi-Analytic Modeling of Dark Matter Subhalo Encounters with Thin Stellar Streams: Statistical Predictions for GD-1-like Streams in CDM

This paper uses semi-analytic modeling to predict that dark matter subhalos with masses between $2\times 10^6$ and $10^8 M_{\odot}$ are most likely to create observable gaps in GD-1-like stellar streams, typically occurring at a rate of three per Milky Way-like host with characteristic widths of 5–27 degrees and underdensities of 10–30%.

The Gist

The Big Picture: Invisible Ghosts in the Night

Imagine the Milky Way galaxy is a giant, swirling city of stars. Hidden within this city are billions of "ghosts" called dark matter subhalos. These ghosts are clumps of invisible matter that have no light of their own, so we can't see them with telescopes.

The scientists in this paper wanted to answer a big question: How often do these invisible ghosts bump into the stars, and what kind of damage do they leave behind?

To find out, they didn't look at the ghosts directly. Instead, they looked at stellar streams.

The Analogy: The Stream as a Long, Thin Ribbon

Think of a stellar stream (like the famous GD-1 stream) as a long, thin ribbon of stars stretching across the sky. This ribbon was once a tight bundle of stars (a globular cluster) that got stretched out over billions of years as it orbited the galaxy.

Because this ribbon is so thin and the stars are moving in perfect unison, it is extremely fragile. If a dark matter "ghost" (a subhalo) flies past it, it acts like a rock thrown into a calm pond. The gravitational pull of the ghost tugs on the ribbon, creating a gap (a missing chunk of stars) or a spur (a little branch sticking out).

How They Did the Study: A Digital Simulation Factory

Since we can't see the ghosts, the researchers built a massive digital simulation to see what would happen if they were there.

1. Building the City (SatGen): They used a computer program called SatGen to generate 400 different versions of the Milky Way. In each version, they randomly placed thousands of dark matter ghosts with different sizes and speeds, following the rules of our current best theory of the universe (Cold Dark Matter).
2. Dropping the Ribbon (Gala): They then simulated a ribbon of stars (like GD-1) falling into each of these 400 virtual cities.
3. Watching the Collision: They watched to see which ghosts flew close enough to the ribbon to cause a disturbance. They filtered out the "weak" bumps and focused only on the ones strong enough to tear a hole in the ribbon.
4. Measuring the Damage (Gala): For the interesting collisions, they ran a high-definition simulation to measure exactly how wide the gap was and how deep the hole in the ribbon was.

What They Found: The "Goldilocks" Ghosts

The study revealed some specific rules about which ghosts are the most destructive:

• The Right Size: The ghosts that cause the biggest gaps aren't the tiniest ones or the massive ones. They are the "Goldilocks" size: roughly 2 million to 100 million times the mass of our Sun.
  • Analogy: If a ghost is too small, it's like a pebble skipping on water—no big splash. If it's too huge, it might be rare or move too fast to leave a clean gap. The medium-sized ones are just right to rip a hole in the ribbon.
• The Speed and Distance: These ghosts usually fly past the ribbon at about 200–400 km/s (very fast!) and pass within about 0.1 to 1.5 kilometers (in astronomical terms, this is a very close shave).
• The Frequency: On average, a ribbon like GD-1 gets hit by a ghost big enough to make a noticeable gap about 3 times during its entire life.
• The Resulting Gaps: When a gap forms, it is usually 5 to 27 degrees wide (that's huge in the sky, about 10 to 50 times the width of the full moon) and has a "depth" where the star density drops by 10% to 30%.

The "Mass" Mystery: Dark vs. Bright

One of the most interesting findings is about what these ghosts are made of.

• The ghosts that cause these gaps likely started their lives with a mass between 20 million and 1 billion suns.
• The paper notes that objects this small are usually too small to hold onto gas and form stars. This means most of the ghosts causing these gaps are completely dark (invisible).
• However, a few might be small, faint dwarf galaxies that do have some stars. So, looking at these gaps helps us hunt for both invisible dark matter and faint, hidden galaxies.

Does the Size of the Galaxy Matter?

The researchers also tested what happens if the Milky Way is twice as heavy as we think it is.

• Result: It didn't change the story much. Even in a heavier galaxy, the gaps look roughly the same size and happen at roughly the same rate. This suggests that our method of finding these gaps is robust, regardless of exactly how heavy our galaxy is.

The Bottom Line

This paper provides a statistical "recipe" for what we should expect to see in the sky. It tells us that if we look at thin streams of stars like GD-1, we should expect to see a few large, clean gaps. These gaps are the fingerprints of invisible dark matter clumps passing by.

By measuring the size and shape of these gaps in real data, astronomers can eventually figure out exactly how much dark matter exists and what it is made of, essentially using the stars as a giant detector for the invisible universe.

Technical Summary

Technical Summary: Semi-Analytic Modeling of Dark Matter Subhalo Encounters with Thin Stellar Streams

Problem Statement
In the $\Lambda$-Cold-Dark-Matter ($\Lambda$CDM) paradigm, structure formation is hierarchical, predicting a vast population of bound dark matter (DM) subhalos below the mass scale required for galaxy formation ($M \lesssim 10^8\text{--}10^9 M_\odot$). These low-mass subhalos are expected to be devoid of luminous matter due to reionization and tidal stripping, rendering them invisible to traditional optical surveys. Consequently, detecting and characterizing this substructure requires indirect methods. Stellar streams, particularly "thin" streams originating from tidally disrupted globular clusters, serve as sensitive dynamical probes. Their low velocity dispersions ($\sim 0.1\text{--}1$ km s$^{-1}$) make them susceptible to perturbations from passing subhalos, which can manifest as density gaps or off-track features (spurs). While previous analytic estimates and simulations have suggested that streams like GD-1 should exhibit gaps, a robust statistical framework connecting the underlying subhalo population of a Milky Way (MW)-mass host to concrete predictions for gap formation rates and properties has remained incomplete.

Methodology
The authors develop a modular, semi-analytic framework to statistically model the interaction between a GD-1-like stellar stream and the DM subhalo population of a MW-mass host. The workflow, illustrated in Figure 1, integrates three primary components:

1. Subhalo Population Generation (SatGen): The study utilizes the semi-analytic code SatGen to generate realizations of the CDM subhalo population around MW-mass hosts. This code constructs merger trees and evolves subhalo structural parameters (mass, concentration) and orbital histories, accounting for tidal stripping, dynamical friction, and tidal tracks. The authors generated 400 "fiducial" realizations with a host virial mass of $10^{12} M_\odot$ and 200 "high-mass" realizations ($2 \times 10^{12} M_\odot$) to assess systematic uncertainties.
2. Stream Modeling and Event Selection (Galpy & Impulse Approximation): A mock GD-1-like stream is generated using the streamdf module in Galpy, evolving 2002 test particles along an eccentric orbit synchronized with the SatGen cosmic time. To identify potential gap-forming encounters, the authors calculate the "parallel impulse curve" ($\delta v_\parallel$) for each stream-subhalo encounter. Encounters are pre-selected based on impact parameters (within 5 times the subhalo's scale radius) and duration. A secondary filter requires the impulse curve to exhibit exactly two extrema with a specific separation and amplitude ($\Delta \delta v_\parallel \ge 0.1$ km s$^{-1}$), indicative of gap formation rather than adiabatic orbital changes.
3. Gap Simulation and Characterization (Gala): Selected candidate events are simulated using the N-body code Gala. The stream is modeled using the particle spray method, and the subhalo is represented as a Plummer sphere with mass and scale radius matching the SatGen subhalo at the time of encounter. The simulation evolves the system from pre-encounter to post-encounter times. Gap properties (width $\Delta \phi_{\text{gap}}$ and depth $f_{\text{gap}}$) are quantified by comparing the density of the perturbed stream against a baseline unperturbed stream, fitting an inverted top-hat function to the density ratio using a modified BoxLeastSquares algorithm.

Key Contributions and Results
The study provides statistical predictions for gap formation in GD-1-like streams within the $\Lambda$CDM framework:

• Gap Frequency: The authors find that a GD-1-like stream is expected to form an average of $\sim 3$ gaps per host realization over its lifetime (specifically, $\langle N \rangle \approx 2.8$ for gaps with depth $f_{\text{gap}} > 0.1$). For gaps resembling the specific feature observed in GD-1 (width $\sim 9$ deg, depth $\sim 0.25$), the rate is approximately 0.3 events per realization.
• Subhalo Mass Scale: The subhalos most likely to produce significant gaps have instantaneous masses at the time of encounter in the range $2 \times 10^6 M_\odot - 10^8 M_\odot$. This corresponds to infall masses of approximately $2 \times 10^7 M_\odot - 10^9 M_\odot$. This mass range implies that gap formation can be driven by both dark subhalos (below the star-formation threshold) and luminous dwarf satellites.
• Gap Properties: Typical gaps have widths of $\sim (5\text{--}27)$ degrees and fractional underdensities of $\sim (10\text{--}30)\%$. The study identifies positive correlations between gap width and subhalo mass, impact parameter, and lookback time. Conversely, gap depth shows weak to no correlation with these parameters, suggesting degeneracies in the flyby geometry.
• Kinematics: Encounters responsible for gaps typically have impact parameters of $(0.1\text{--}1.5)$ kpc and relative velocities of $\sim (200\text{--}410)$ km s$^{-1}$.
• Host Mass Dependence: Increasing the host halo mass (to $2 \times 10^{12} M_\odot$) increases the subhalo number density and encounter velocities. While this leads to a modest increase ($\sim 10\%$) in the gap formation rate, the statistical properties of the gaps (width and depth distributions) remain largely insensitive to the host mass, consistent with previous analytic work.
• Temporal Evolution: The encounter rate exhibits secular growth as the stream lengthens over time. Furthermore, the rate peaks during the stream's pericenter passages, where the stream is both longest and fastest, maximizing the volume of phase space swept.

Significance and Claims
The paper claims to establish a robust, modular framework for quantifying stream-subhalo encounters that relaxes assumptions common in previous studies (e.g., circular orbits, simplified subhalo populations). By combining semi-analytic subhalo generation with detailed N-body stream simulations, the authors provide a statistically rigorous prediction for the frequency and morphology of gaps in thin streams.

The authors emphasize that their results suggest gaps in streams like GD-1 are "uncommon but not in a statistically significant way" given current observational constraints, noting that the presence of specific features like spurs may further narrow the parameter space of viable perturbers. They assert that their method is adaptable to other DM paradigms (e.g., Self-Interacting Dark Matter or Warm Dark Matter) and can be extended to study other stream features (such as spurs) or populations of streams. The work is positioned as a necessary step in preparing for upcoming large-scale surveys (Vera Rubin Observatory, Roman Space Telescope) to maximize the science return regarding dark matter substructure.

Limitations
The authors acknowledge several limitations:

• The stream modeling uses $\sim 5 \times 10^5$ test particles, far exceeding the number of observed stars in GD-1, which simplifies noise but may not capture real observational biases.
• Gap identification relies on a comparison with an unperturbed stream model, a standard practice that defines gaps relative to the simulation rather than absolute observational data.
• Encounters are treated as independent events; the cumulative effect of multiple subhalo interactions on a single stream trajectory is not modeled.
• The study does not model the stellar population of the stream or the observability of gaps in real data, focusing instead on the dynamical formation of underdensities.