The Hidden Role of Anisotropies in Shaping Structure Formation in Cosmological N-Body Simulations
Francesco Sylos Labini · 2025
Read The Hidden Role of Anisotropies in Shaping Structure Formation in Cosmological N-Body Simulations by Francesco Sylos Labini (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-13765.txt.
Introduction
Francesco Sylos Labini warns that computer-made universes carry a hidden flaw in this 2025 paper. Simulations of cosmic growth start by shifting particles off a neat cube grid, guided by a wavy field. That start matches the target spread on average, yet the grid leaves faint direction bias in the layout.
A direction-aware measure makes the bias plain where averages hide it. In two dark matter runs, gravity did not wash the bias out. Instead it grew, seeding thread-like shapes that last into late cosmic times. The author calls these shapes artifacts of the setup, not true forecasts. Start methods, he argues, need a fresh hard look.
Catalog record
| CREATOR | Francesco Sylos Labini |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 76 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-13765
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The Hidden Role of Anisotropies in Shaping Structure Formation in Cosmological N-Body Simulations Preprint: APS/123-QED Francesco Sylos Labini Abstract Initial conditions in cosmological N N -body simulations are typically generated by displacing particles from a regular cubic lattice using a correlated field derived from the linear power spectrum, often via the Zel’dovich approximation. While this procedure reproduces the target two-point statistics (e.g., the power spectrum or correlation function), it introduces subtle anisotropies due to the underlying lattice structure. These anisotropies, invisible to angle-averaged diagnostics, become evident through directional measures such as the Angular Distribution of Pairwise Distances. Analyzing two Cold Dark Matter simulations with varying resolutions, initial redshifts, and box sizes, we show that these anisotropies are not erased but are amplified by gravitational evolution. They seed filamentary structures that persist into the linear regime, remaining visible even at redshift z = 0 z=0 . Our findings demonstrate that such features are numerical artifacts — emerging from the anisotropic coupling between the displacement field and the lattice — not genuine predictions of an isotropic cosmological model. These results underscore the importance of critically reassessing how initial conditions are constructed, particularly when probing the large-scale, quasi-linear regime of structure formation. In cosmological N -body simulations of Cold Dark Matter (CDM) models, the present-day matter distribution is highly inhomogeneous on small scales — such as those of galaxies and galaxy clusters — but becomes statistically homogeneous and isotropic on sufficiently large scales ( Davis et al. 1985 ) . A central and longstanding question in cosmology is whether the large-scale structures observed in galaxy redshift surveys — extending across tens to hundreds of megaparsecs — are consistent with the predictions of \Lambda CDM N -body simulations ( Huchra and Geller 1982 ; Gott III et al. 2005 ; De Marzo et al. 2021 ; Lopez et al. 2022 ; Lopez et al. 2024 ) . Numerous studies have claimed that CDM simulations, under the standard \Lambda CDM framework, successfully reproduce the large-scale structures observed in the Universe. The evolved mass distribution exhibits a rich and interconnected network of voids, walls, clusters, and filaments — collectively known as the cosmic web ( White et al. 1987 ; Springel et al. 2005 ; Park et al. 2012 ; Sawala et al. 2025 ) . On scales exceeding several hundred Mpc, the distribution becomes increasingly featureless, approaching statistical homogeneity. In both simulations and observations, large-scale structures are commonly identified using Friends-of-Friends algorithms ( Huchra and Geller 1982 ; Davis et al. 1985 ; Lopez et al. 2024 ; Sawala et al. 2025 ) , which group particles or galaxies into structures based on a fixed linking length. This algorithm, while simple and widely used, introduces a scale dependence in structure detection that warrants careful interpretation. Within this framework, it has been argued that structures spanning several hundred Mpc are consistent with CDM predictions — a conclusion that appears puzzling in light of the underlying physics. On small scales ( r\lesssim 10,\mathrm{Mpc}/h ), non-linear gravitational collapse leads to the formation of virialized, quasi-spherical halos ( Navarro et al. 1997 ) . In this regime, the non-linear dynamics erase the memory of the initial conditions (ICs), so that the resulting structures predominantly reflect the bottom-up gravitational assembly characteristic of CDM-like structure formation, rather than preexisting features in the primordial density field. On larger scales ( r\gtrsim 10,\mathrm{Mpc}/h ), the dynamics remain in the linear or mildly non-linear regime, where structures such as filaments can form through the coherent amplification of initial density correlations. Notably, cosmological simulations exhibit filaments that extend far beyond 10 Mpc, in some cases reaching scales comparable to the entire simulation volume — up to several hundred megaparsecs ( White et al. 1987 ; Springel et al. 2005 ; Park et al. 2012 ) . Despite their prominence, the physical origin of these extended structures is often not scrutinized in detail, raising important questions about their relation to ICs, finite-size effects, and numerical artifacts. In standard CDM cosmology, particles typically travel no more than \sim 10,\mathrm{Mpc}/h from their initial positions during the simulation. This implies that structures larger than this scale must emerge from correlations already present in the ICs. At early times, the two-point correlation function \xi(r) of the CDM density field exhibits a clear transition: it follows a power-law decay at small scales, while at intermediate scales ( 10\lesssim r\lesssim 100,\mathrm{Mpc}/h ) it approaches an exponential decay Gabrielli et al. 2005 . The characteristic scale of this exponential behavior — usually defined as the correlation length in statistical physics Gabrielli et al. 2005 — is approximately r_{c}\approx 18\,\mathrm{Mpc}/h ( Sylos Labini 2011 ) . Beyond this scale, \xi(r) has negligible amplitude, and no long-range correlations or coherent structures of even small amplitude are expected to be amplified under linear gravitational evolution. The two-point correlation function \xi(r) crosses zero at approximately 120\;\mathrm{Mpc}/h and remains negative beyond this scale. This alternation between positive and negative correlations reflects the super-homogeneous nature of the distribution Gabrielli et al. 2002 ; Gabrielli et al. 2003 ; Torquato and Stillinger 2003 ; Gabrielli et al. 2005 ; Torquato 2018 ; Philcox and Torquato 2023 . Super-homogeneous systems are characterized by suppressed mass fluctuations compared to a random (Poisson) distribution: the variance of mass within a spherical region grows proportionally to the surface area rather than the volume. Examples of super-homogeneity include all perfect crystals, perfect quasicrystals, and exotic amorphous states of matter Torquato 2018 . While standard analyses typically focus on the non-linear amplification of \xi(r) at small scales, comparatively little attention has been paid to the regime in which \xi(r) remains positive but exhibits an exponential decay — corresponding to the linear growth phase. Yet, this is precisely the range where filamentary structures are frequently observed in simulations. Although some of these features may result from chance alignments, potentially producing structures that extend far beyond the expected correlation length, their systematic appearance across different simulations suggests the presence of a more fundamental underlying mechanism. We argue that such large-scale filamentary structures on scales larger than 10-20 Mpc /h are artifacts of how the ICs are generated. In most N -body codes, ICs are produced by displacing particles from a regular lattice (the so-called pre-initial conditions) using a correlated Gaussian displacement field derived from the initial power spectrum, often via the Zel’dovich approximation (see, e.g., Carlberg and Couchman 1989 ; Baugh et al. 1995 ; Couchman et al. 1995 ; Crocce et al. 2006 and references therein). While this procedure ensures that spherically averaged statistics, such as the power spectrum P(k) or the two-point correlation function \xi(r) , match the theoretical model reasonably well over a finite range of spatial scales, it simultaneously introduces subtle but non-trivial anisotropies into the particle distribution. These anisotropies arise from the intrinsic spatial correlations and geometric symmetries of the regular grid used as the pre-initial configuration. Crucially, they are not fully erased by the small-amplitude, correlated displacement field applied during the generation of the ICs. These residual anisotropies are typically not captured by angle-averaged statistics. However, they become evident when analyzed using directional statistics, such as the Angular Distribution of Pairwise Distances (ADPD), which is sensitive to anisotropic clustering. The ADPD provides