Planet DSMC Code: Methods & Applications

• Planet DSMC Code is an ambiguous label referring to distinct computational approaches in seismology, gas dynamics, and exoplanet inference, each defined by its context.
• In planetary seismology, the DSpecM1D code employs a displacement-based spectral method to solve the frequency-domain seismic wave equation with sub-percent benchmarking accuracy.
• For astrophysical gas dynamics and exoplanet inference, implementations range from hybrid DSMC–EPSM methods for ram-pressure flow to Bayesian orbital fittings using Differential Evolution MCMC.

“PLANET DSMC Code” does not denote a single, uniformly identified software package in the cited literature. Instead, the expression is associated with several technically distinct computational traditions: a planetary whole-body seismology code based on the Direct Solution Method, namely $\texttt{DSpecM1D}$, which is explicitly not a Direct Simulation Monte Carlo code (Myhill et al., 9 Mar 2026); astrophysical rarefied-gas solvers built on Direct Simulation Monte Carlo, such as the hybrid DSMC–EPSM code coupled to EXP for ram-pressure dynamics (Weinberg, 2013); and planetary-system Monte Carlo inference codes such as RUN DMC, which perform Bayesian orbital fitting rather than gas-kinetic simulation (Nelson et al., 2013). The term is therefore best understood as an ambiguous label whose meaning depends on whether the context is seismology, rarefied-flow kinetics, or exoplanet parameter inference.

1. Nomenclature and domain ambiguity

The most important fact about the phrase is that the cited papers do not identify a code literally called “PLANET DSMC Code.” They instead document several nearby usages of “DSM,” “DSMC,” and “DMC,” and they repeatedly warn that the acronym is context-dependent. In seismology, “DSM” denotes the Direct Solution Method; in gas dynamics, “DSMC” denotes Direct Simulation Monte Carlo; and in exoplanet inference, “DMC” appears in the name RUN DMC but refers to Differential evolution Markov chain Monte Carlo, not rarefied-gas simulation (Myhill et al., 9 Mar 2026, Nelson et al., 2013).

Interpretation	Code or method	Scope
Planetary direct solution seismology	$\texttt{DSpecM1D}$	Synthetic seismograms in self-gravitating, spherically symmetric planets
Astrophysical DSMC gas dynamics	Hybrid DSMC–EPSM + EXP	Ram-pressure stripping and multiphase gas–gravity interaction
Planetary-system Monte Carlo inference	RUN DMC	Radial-velocity fitting with $N$-body DEMCMC

This ambiguity is not merely terminological. It changes the governing equations, numerical discretization, output observables, and even the meaning of “particle.” In $\texttt{DSpecM1D}$, the unknown is a frequency-domain seismic displacement field reduced to radial ODE systems (Myhill et al., 9 Mar 2026). In astrophysical DSMC, particles are simulation molecules advancing under the collisional Boltzmann equation (Weinberg, 2013). In RUN DMC, the “particles” are absent altogether; the method samples posterior distributions over orbital parameters using an ensemble MCMC proposal mechanism (Nelson et al., 2013).

2. $\texttt{DSpecM1D}$: the planetary DSM code in seismology

If the expression is intended to mean a planetary direct solution / direct spectral method code, the relevant system is $\texttt{DSpecM1D}$, introduced in “The direct spectral element method for the calculation of synthetic seismograms in self-gravitating, spherically symmetric planets” (Myhill et al., 9 Mar 2026). The code is defined as

$$\texttt{DSpecM1D} \equiv \text{Direct Spectral Element Method 1D}.$$

Its purpose is to solve the frequency-domain seismic wave equation in self-gravitating, spherically symmetric, non-rotating, anelastic, transversely isotropic planet models, and to recover time-domain synthetic seismograms by inverse Fourier–Laplace transformation (Myhill et al., 9 Mar 2026). The paper is explicit that the method is not Earth-specific, even though its examples are Earth-centered.

The formulation starts from the weak form

$$-\omega^2 \mathcal{P}(\mathbf{u}', \mathbf{u}) + \mathcal{H}(\omega,\mathbf{u}', \mathbf{u}) = \mathcal{F}(\mathbf{u}'),$$

then reduces the problem by generalized spherical harmonic expansion into independent 1D radial systems for each spherical harmonic degree $l$. The displacement is split into toroidal and spheroidal subsystems, with spheroidal radial functions $U_{lm}(r),V_{lm}(r),\phi_{lm}(r)$ and toroidal radial function $\texttt{DSpecM1D}$0. A central methodological point is that the code uses a displacement formulation in both solids and fluids, rather than a fluid potential formulation, which allows inclusion of full self-gravitation, solid-fluid coupling, and arbitrary fluid stratification in a single variational framework (Myhill et al., 9 Mar 2026).

This displacement-based treatment in fluids is the key reason the code is relevant to planetary whole-body seismology. The method handles solid-only planets, planets with fluid cores, Earth-like mantle–outer-core–inner-core structures, and in principle other spherically symmetric bodies with arbitrary radial stratification. The paper is explicit that the method does not require neutral stratification in the fluid. Supported constitutive complexity includes transverse isotropy in solids and attenuation/dispersion through frequency-dependent Love parameters $\texttt{DSpecM1D}$1 (Myhill et al., 9 Mar 2026).

Radially, the planet is partitioned into spectral elements $\texttt{DSpecM1D}$2, with GLL Lagrange polynomials on each element and GLL quadrature for integration. After Galerkin discretization, each $\texttt{DSpecM1D}$3 problem becomes

$\texttt{DSpecM1D}$4

The resulting matrix is stated to be symmetric but not Hermitian, and the implementation uses sparse LU decomposition with the Eigen library (Myhill et al., 9 Mar 2026). The problem is embarrassingly parallel in both $\texttt{DSpecM1D}$5 and $\texttt{DSpecM1D}$6, and the implementation parallelizes over degree $\texttt{DSpecM1D}$7.

Validation is extensive. $\texttt{DSpecM1D}$8 is benchmarked against $\texttt{DSpecM1D}$9 and $N$0. For the Bolivia 1994 benchmark, the vertical spectra comparison against $N$1 yields average relative misfit $N$2 and maximum $N$3; three-component displacement and acceleration benchmarks show mean differences below $N$4, with maximum differences of a few percent in attenuating/dispersive PREM tests (Myhill et al., 9 Mar 2026). A first version of the library is available at the GitHub repository explicitly given in the paper. In this sense, $N$5 is simultaneously a methodological paper, a code paper, and a benchmarking paper.

3. Direct Simulation Monte Carlo in astrophysical gas dynamics

If “PLANET DSMC Code” is intended in the literal Direct Simulation Monte Carlo sense, the relevant cited astrophysical example is the hybrid DSMC–EPSM gas solver coupled to the $N$6-body code EXP in “Direct Simulation Monte Carlo for astrophysical flows: II. Ram pressure dynamics” (Weinberg, 2013). That paper does not name a code called PLANET, but it does describe a concrete DSMC-based implementation for galaxy–ICM/IGM interaction.

The solver is a parallel hybrid DSMC/EPSM method merged with EXP. DSMC advances the collisional Boltzmann equation by alternating collisionless streaming and a Monte Carlo collision step. In dense regions, where explicit DSMC collisions become expensive because the mean free path is very small, the method switches to EPSM (Equilibrium Particle Simulation Method), replacing local states by an equivalent thermalized random distribution that conserves energy and momentum and is designed to recover the Navier–Stokes limit when mean free paths are much smaller than the flow scale (Weinberg, 2013).

The astrophysical coupling is one-way in the reported runs: stars and dark matter are self-gravitating and mutually interacting, gas feels their gravitational field, but gas self-gravity is omitted in those simulations. The target regime is explicitly transitional and multiphase rather than purely continuum: hot/cold, dense/rarefied interfaces, shocks, ablation, Kelvin–Helmholtz instability, and ram-pressure stripping. The paper emphasizes that many such astrophysical flows are “not quite hydrodynamic” and are therefore natural applications for DSMC (Weinberg, 2013).

Implementation-level details are present but incomplete. The code uses collision cells, monitors local Knudsen number and ballistic path relative to cell size, and tracks the fraction of cells in the EPSM regime. The most explicit resolution guidance given is that an ideally tuned DSMC simulation simultaneously maintains approximately 10 particles per simulation volume whose length scale is approximately 1 mean free path (Weinberg, 2013). Boundary conditions are problem-specific: periodic in the slab plane for the slab test, vacuum vertically, and periodic transverse boundaries for finite-column wind–galaxy impact calculations.

The paper is best viewed as an application paper rather than a software manual. It does not provide source code, pseudocode, collision-selection formulas, or detailed parallel decomposition. Nonetheless, it is an unambiguous example of DSMC in an astrophysical setting, and it demonstrates that a “planetary-adjacent” interpretation of DSMC refers to rarefied gas kinetics rather than to seismological DSM.

4. RUN DMC and the exoplanet inference interpretation

A third reading arises from the conjunction of “planet” and “Monte Carlo.” “RUN DMC: An efficient, parallel code for analyzing Radial Velocity Observations using N-body Integrations and Differential Evolution Markov chain Monte Carlo” presents a planetary-system inference code, but it is not a DSMC gas-dynamics solver (Nelson et al., 2013). Its domain is exoplanet radial-velocity fitting when mutual interactions make a Keplerian approximation inadequate.

RUN DMC combines self-consistent $N$7-body integration with Differential Evolution MCMC. The posterior is explored in a high-dimensional parameter space, at least $N$8 dimensions for a coplanar $N$9-planet system, using an ensemble proposal of the form

$\texttt{DSpecM1D}$0

with $\texttt{DSpecM1D}$1 and default $\texttt{DSpecM1D}$2 (Nelson et al., 2013). The method is designed to handle correlated posteriors in resonant or strongly interacting multi-planet systems.

The code uses a time-symmetric 4th-order Hermite $\texttt{DSpecM1D}$3-body integrator, broad separable priors, Gaussian likelihood with an explicit jitter term, and parallelization via OpenMP and Swarm-NG. The authors recommend $\texttt{DSpecM1D}$4, note that $\texttt{DSpecM1D}$5 is usually appropriate, and show that larger ensembles can matter substantially for systems with more planets or more parallel hardware (Nelson et al., 2013).

This interpretation matters because the acronymic overlap can be misleading. RUN DMC is a planetary Monte Carlo code, but its “particles” are MCMC chains, not gas molecules, and its outputs are posterior samples over orbital parameters rather than flowfields or seismograms.

5. Directions in DSMC code development relevant to a production codebase

Several papers describe capabilities that a production DSMC codebase would need if “PLANET DSMC Code” denotes a rarefied-flow solver rather than $\texttt{DSpecM1D}$6. These developments cluster around boundary modeling, chemistry, multiscale acceleration, and hybrid coupling.

For surface physics, “Radiative equilibrium boundary condition and correlation analysis on catalytic surfaces in DSMC” introduces a radiative-equilibrium wall treatment coupled to finite-rate surface chemistry in DSMC. The core local boundary condition is

$\texttt{DSpecM1D}$7

with convective and chemical heating sampled directly from particle impacts and surface reactions (Ko et al., 22 Apr 2025). The paper shows that linear interpolation between two isothermal DSMC runs can underpredict catalytic radiative-equilibrium heat flux by up to $\texttt{DSpecM1D}$8 over $\texttt{DSpecM1D}$9, demonstrating that catalytic wall temperature and heat flux are nonlinearly coupled.

For multiscale steady-state acceleration, “Enhancing DSMC simulations of rarefied gas mixtures using a fast-converging and asymptotic-preserving scheme” proposes direct intermittent GSIS–DSMC coupling, in which DSMC is periodically augmented by synthetic macroscopic equations built from DSMC-sampled stress, heat flux, and inter-species exchange terms (Luo et al., 25 Aug 2025). At $\texttt{DSpecM1D}$0, the reported speedups relative to SPARTA range from about $\texttt{DSpecM1D}$1 to $\texttt{DSpecM1D}$2, depending on mixture model and mass ratio, while preserving agreement with fine-mesh DSMC. Closely related ideas appear in the turbulent-rarefied DIG-SST framework, which couples DSMC moments to synthetic equations and a $\texttt{DSpecM1D}$3-$\texttt{DSpecM1D}$4 SST turbulence model so that the method asymptotically converges either to SST or to DSMC results, depending on regime (Luo et al., 13 Mar 2025).

For hybridization and noise suppression, two complementary strands are represented. “DSMC-LBM mapping scheme for rarefied and non-rarefied gas flows” formulates a Hermite/Gauss–Hermite map between DSMC and LBM, so that DSMC moments can be projected to lattice populations and LBM populations can be reconstructed into a continuous truncated distribution for particle resampling (Staso et al., 2015). “Multilevel radial basis function surrogates for noise-robust DSMC-CFD coupling” develops a surrogate-based DSMC–CFD framework, using SPARTA as the micro solver and OpenFOAM as the macro solver, with multilevel RBFs replacing earlier global basis functions in MMS-Sparse (Kamal et al., 27 Apr 2026). A plausible implication is that modern DSMC infrastructures increasingly treat coupling and denoising as first-class numerical components rather than as postprocessing.

For chemistry and internal energy, “A harmonically-coupled-anharmonic-oscillator approach for polyatomic chemistry modeling in DSMC” extends anharmonic oscillator modeling from diatomics to polyatomic molecules in PICLas. Stretching modes are treated as local anharmonic Morse oscillators, bending modes as harmonic oscillators, and dissociation is made mode-specific through local stretching excitation and zero-point-energy bookkeeping (Lauterbach et al., 30 Jun 2026). This directly targets planetary-entry and other high-enthalpy nonequilibrium applications where uncoupled harmonic vibrational models become inadequate.

6. Validation practice, limitations, and recurrent misconceptions

Across these literatures, validation is not ancillary. $\texttt{DSpecM1D}$5 is benchmarked against $\texttt{DSpecM1D}$6 and $\texttt{DSpecM1D}$7 with sub-percent agreement for elastic and anelastic PREM-based tests (Myhill et al., 9 Mar 2026). The astrophysical DSMC/EPSM work cites standard shock tube and Kelvin–Helmholtz benchmarks in its companion paper, while the application paper itself emphasizes physically resolved shock formation, ablation, and angular-momentum transport (Weinberg, 2013). Hybrid and multiscale DSMC papers validate against SPARTA, analytical limits, or benchmark DSMC datasets rather than treating acceleration alone as sufficient (Luo et al., 25 Aug 2025, Kamal et al., 27 Apr 2026).

The main limitations are equally explicit. $\texttt{DSpecM1D}$8 is restricted to non-rotating, spherically symmetric, 1D radial models and is not a 3D arbitrary-geometry solver (Myhill et al., 9 Mar 2026). The ram-pressure DSMC paper omits gas self-gravity in the reported runs and does not disclose a reusable code package (Weinberg, 2013). RUN DMC is not a gas-dynamics code at all; it is a Bayesian exoplanet RV sampler (Nelson et al., 2013). In DSMC itself, several studies underscore that low-Kn regimes are expensive, that statistical noise is intrinsic to cellwise particle sampling, and that hybridization is often necessary. The comparative study using SPARTA and OpenFOAM concludes that DSMC is superior for $\texttt{DSpecM1D}$9, whereas Navier–Stokes CFD is more accurate and efficient in the continuum regime $\texttt{DSpecM1D}$0, with a practical continuum-breakdown threshold $\texttt{DSpecM1D}$1 for hybrid decomposition (Peravali et al., 2024).

The most persistent misconception is therefore acronymic. In the cited literature, DSM in planetary seismology is the Direct Solution Method, DSMC in gas dynamics is Direct Simulation Monte Carlo, and DMC in RUN DMC refers to differential evolution Markov chain Monte Carlo. The phrase “PLANET DSMC Code” can point to any of these only by contextual inference, not by stable nomenclature. The seismological code most closely matching a “planetary DSM code” is $\texttt{DSpecM1D}$2 (Myhill et al., 9 Mar 2026). The astrophysical rarefied-flow example most closely matching a literal DSMC code is the hybrid DSMC–EPSM implementation coupled to EXP (Weinberg, 2013). The exoplanet inference code is RUN DMC (Nelson et al., 2013).