Modeling Globular Cluster Stellar Streams with a Basis-Expansion N-body Code

1. Problem Statement

Globular cluster (GC) stellar streams are critical probes for understanding galaxy formation and the distribution of dark matter substructure. However, modeling their evolution presents a significant computational trade-off:

• Direct N-body codes (e.g., NBODY6++GPU): These are the "gold standard" for accuracy, directly summing forces between all particles. However, they scale as $O(N^2)$, making them prohibitively expensive for realistic GCs ($N > 10^5$) over cosmological timescales (Hubble time).
• Particle-spray methods: These approximate the escape process by injecting stars near Lagrange points based on tuned distribution functions. While computationally fast, they often assume a fixed, spherically symmetric progenitor and ignore the internal collisional dynamics and time-dependent evolution of the cluster. This can lead to inaccuracies in stream morphology and kinematics, particularly for clusters subject to strong tidal forces.
• Monte Carlo methods (e.g., CMC): These approximate collisional evolution efficiently ($O(N \log N)$) but typically assume spherical symmetry and rely on post-processing to model tidal debris in asymmetric host potentials, introducing potential errors in escape physics.

The Gap: There is a need for a method that accurately models the evolving, collisional dynamics of a GC and its resulting tidal debris in a realistic host potential, with a runtime significantly faster than direct N-body but more accurate than current particle-spray approximations.

2. Methodology: The KRIOS Code

The authors introduce KRIOS, a new hybrid N-body code designed to bridge this gap. KRIOS combines a Self-Consistent Field (SCF) method for collisionless gravity with a statistical treatment of two-body relaxation.

• Self-Consistent Field (SCF) Potential:
  • The gravitational potential is expanded using a set of basis functions (Zhao 1996) that are tunable to the cluster's density profile.
  • The potential is updated dynamically as the cluster evolves, allowing for the resolution of non-spherical structures and time-dependent tidal fields.
  • Force calculations scale as $O(N)$ and are parallelizable.
• Collisional Dynamics (Relaxation):
  • Two-body relaxation is modeled using a 3D version of the Hénon (1971) "super-encounter" method, where cumulative effects of many encounters are approximated by single effective interactions.
  • This captures gravothermal evolution and core collapse without the $O(N^2)$ cost.
• Time-Stepping Strategy:
  • KRIOS employs two distinct timesteps to handle different physical scales:
    1. System Timestep ($\Delta t_{sys}$): For two-body relaxation and full SCF recomputation.
    2. Integration Timestep ($\Delta t_{int}$): Determined by the tidal tensor of the host galaxy. This allows the code to resolve rapid changes in the cluster's potential and mass loss caused by tidal shocks (e.g., disk passages) that occur on dynamical timescales, faster than the relaxation timescale.
• Reference Frames:
  • The code uses an accelerating, non-rotating reference frame for the cluster to avoid explicit rotational pseudo-forces.
  • A test particle integration tracks the cluster's center of mass through the host potential, allowing for accurate interpolation of the cluster's position at arbitrary times.

3. Key Contributions

1. Development of KRIOS: A new N-body code that integrates collisional stellar dynamics with a self-consistent, time-varying potential in an external tidal field.
2. Validation against Gold Standards: The authors demonstrate that KRIOS reproduces the internal evolution (Lagrange radii, mass loss) and stream properties of NBODY6++GPU simulations with high fidelity but in a fraction of the wall-clock time (e.g., ~11 hours vs. ~6 months for equivalent $N$).
3. Systematic Comparison with Particle-Spray: The paper provides a comprehensive comparison between KRIOS and various particle-spray methods (e.g., Fardal15, Chen25) across 10 different orbits mimicking known Milky Way streams.
4. Identification of Failure Modes: The study identifies specific orbital regimes where particle-spray methods fail:
   • Orbits with strong tidal forces (small perigalacticon).
   • Orbits where the progenitor is tightly bound to the host (low energy, high angular momentum).
5. Analysis of Progenitor Physics: The authors investigate the importance of progenitor modeling, showing that while spherical symmetry is often sufficient, the treatment of mass-loss rates and the time-evolution of the progenitor's potential are critical for accurate stream kinematics.

4. Results

• Performance: KRIOS runs are significantly faster than NBODY6++GPU. For $N=2 \times 10^5$, KRIOS took a median of 11 hours compared to much longer times for direct N-body, while maintaining accuracy.
• Internal Evolution: KRIOS matches NBODY6++GPU in tracking Lagrange radii and mass loss rates, even for eccentric orbits subject to disk shocks.
• Stream Morphology and Kinematics:
  • Integrals of Motion: KRIOS reproduces the distribution of energy ($E$) and angular momentum ($L_{z'}$) of stream particles much more accurately than particle-spray methods, particularly for eccentric orbits.
  • Discrepancies: Particle-spray methods show significant deviations in stream density ($\rho(\phi_1)$) and velocity dispersion ($\sigma(\phi_1)$) near the progenitor and in the tidal tails for orbits with strong tidal interactions.
  • Specific Orbits: For orbits like "Orbit 3" (strong tidal forces), particle-spray models systematically underestimated velocity dispersion and mispredicted density fluctuations. For "Orbit 5" (weaker tides), agreement was better but still showed systematic offsets.
• Progenitor Assumptions:
  • Spherical Symmetry: Enforcing spherical symmetry ($\ell_{max}=0$) in the KRIOS potential (mimicking CMC) resulted in only mild disagreements in stream properties, suggesting spherical approximations are often sufficient for escape dynamics if the potential is updated correctly.
  • Mass Loss Rate: A constant mass-loss rate (common in particle-spray) is a poor fit for clusters undergoing disk shocks. Variable mass-loss rates, particularly those tied to the tidal tensor's strength, are necessary to replicate the bursty nature of mass loss seen in KRIOS/NBODY6++GPU.

5. Significance

• Dark Matter Probes: Accurate modeling of stellar streams is essential for detecting dark matter subhalos. If stream substructure (gaps, spurs) is misattributed to modeling artifacts (e.g., incorrect escape physics or mass-loss rates) rather than dark matter encounters, it could lead to false positives or negatives in dark matter searches.
• Efficiency for Large-Scale Studies: KRIOS enables the generation of large grids of stream models necessary for Bayesian inference and population studies, which are currently impossible with direct N-body codes.
• Refining Stream Models: The paper provides a roadmap for improving particle-spray codes, emphasizing the need for time-dependent progenitors, realistic mass-loss prescriptions, and the inclusion of strong encounters, rather than just refining the initial distribution functions.
• Future Cosmology: The authors plan to integrate KRIOS with time-dependent host potentials from cosmological simulations (e.g., FIRE), paving the way for realistic catalogs of stellar streams in a cosmological context.

In summary, KRIOS represents a major step forward in stellar dynamics, offering a "sweet spot" between the prohibitive cost of direct N-body simulations and the oversimplifications of particle-spray methods, thereby enabling more robust studies of galaxy formation and dark matter physics.