Abstract
We present the first simulation of the signature of baryonic acoustic oscillations (BAO) in Lyman alpha forest data containing 180,000 mock quasar sight-lines. We use eight large dark-matter only simulations onto which we paint the Lyman alpha field using the fluctuating Gunn-Peterson approximation. We argue that this approach should be sufficient for the mean signature on the scales of interest. Our results indicate that Lyman alpha flux provides a good tracer of the underlying dark matter field on large scales and that redshift space distortions are well described by a simple linear theory prescription. We compare Fourier and configuration space approaches to describing the signal and argue that configuration space statistics provide useful data compression. We also investigate the effect of a fluctuating photo-ionizing background using a simplified model and find that such fluctuations do add smooth power on large scales. The acoustic peak position is, however, unaffected for small amplitude fluctuations (<10%). Larger amplitude fluctuations make the recovery of the BAO signal more difficult and may degrade the achievable significance of the measurement.
Introduction
To understand the nature of Dark Energy, cosmologists must map the expansion history of the universe. The primary tool for this is the Baryon Acoustic Oscillation (BAO) signal—a relic sound wave from the early universe that leaves a characteristic imprint on the clustering of matter at a scale of approximately 150 Megaparsecs (Mpc). While this signal is well-measured in the Cosmic Microwave Background and relatively recent galaxy surveys, a significant data gap exists at high redshifts ($z \sim 2–3$). Mapping millions of galaxies at these distances is prohibitively expensive. This 2009 paper by Slosar et al. explores whether the Lyman-alpha (Ly$\alpha$) forest—absorption lines in quasar spectra caused by intergalactic neutral hydrogen—can serve as a reliable and more accessible tracer of the BAO signal.
Methodology and Datasets
The research team utilized a suite of eight large-volume N-body simulations, each representing a cube 1.5 $h^{-1}$Gpc on a side with $3000^3$ particles. To model the gas without the computational cost of full hydrodynamics, they employed the Fluctuating Gunn-Peterson Approximation (FGPA), which maps Ly$\alpha$ flux onto the dark matter density field via a power-law relationship ($\tau \propto \rho^{1.5-2.0}$).
The study generated 180,000 mock quasar sight-lines, mimicking the density expected in surveys like the Baryon Oscillation Spectroscopic Survey (BOSS). The data were analyzed using both 3D correlation functions (configuration space) and cross-power spectra (Fourier space).
Key Insights: The IGM as a Dark Matter Tracer
The simulations revealed that on the large scales relevant to BAO (~100 Mpc), the intergalactic medium (IGM) is a faithful tracer of underlying dark matter. Despite non-linearities on small scales, the gas behaves linearly on large scales. Furthermore, at $z=2.5$, the universe is sufficiently smooth that non-linear gravitational collapse only broadens the BAO peak by roughly 4%, allowing linear theory to remain highly accurate.
A critical finding was the superiority of configuration space analysis for this type of data. Because quasar sight-lines are essentially sparse "pencils" through space, Fourier analysis is prone to severe aliasing. The 3D correlation function proved to be a far more robust method for data compression and signal recovery.
Results: Detecting the BAO Signal
The primary success of the study was the clear recovery of the BAO "ridge" at the expected scale of $r \sim 105 \, h^{-1}$Mpc in the mock flux correlation functions. The measurement of "bias"—the degree to which gas clusters relative to dark matter—was found to be approximately $b \simeq 0.2$ and scale-invariant on large scales.
Additionally, the study confirmed that redshift space distortions (the Kaiser effect) are well-described by standard linear prescriptions. Tests on fluctuating UV backgrounds showed that while such fluctuations add smooth noise, they do not shift the critical position of the BAO peak.
Limitations and Future Outlook
While groundbreaking, the study had limitations. The dark-matter-only simulations did not resolve the Jeans scale, which could affect error bar calculations. The model also assumed parallel quasar sight-lines, a simplification of real-world wide-angle survey geometries. Potential systematics like metal contamination and quasar continuum fluctuations remain challenges for observational data. Nevertheless, this work provided the theoretical foundation for the subsequent detection of the BAO peak in Ly$\alpha$ forest data by the BOSS collaboration, cementing the forest's role as a premier cosmological tool.
Summary updated