Ionizing radiation fluctuations and large-scale structure in the Lyman-alpha forest

TL;DR

This paper investigates large-scale fluctuations in the hydrogen ionizing background at $z=3$ and their impact on the Ly$\alpha$ forest by coupling large-volume $\\Lambda$CDM N-body simulations with a high-resolution hydrodynamic run to calibrate optical depths. The authors implement a Monte Carlo raytracing framework to propagate quasar-derived radiation through an inhomogeneous IGM, incorporating lightcone effects, finite quasar lifetimes, and anisotropic beaming, and they generate Ly$\alpha$ spectra to quantify changes in the flux power spectrum. They find that radiation fluctuations suppress the Ly$\alpha$ flux power spectrum by about $10$–$12\%$ on scales $k\sim0.05$–$0.5\,h\,\mathrm{Mpc}^{-1}$, with lightcone effects playing a larger role than beaming; a uniform extra background further reduces this suppression. In parallel, they predict a strong foreground proximity effect around foreground QSOs, and compare to Sloan Digital Sky Survey data, which show no such effect but instead enhanced absorption near QSOs, highlighting tensions between simple quasar-radiation models and observations. The work has important implications for interpreting Ly$\alpha$ forest data in cosmological analyses and provides a framework for refining radiative-transfer modeling in the high-redshift IGM.

Abstract

We investigate the large-scale inhomogeneities of the hydrogen ionizing radiation field in the Universe at redshift z=3. Using a raytracing algorithm, we simulate a model in which quasars are the dominant sources of radiation. We make use of large scale N-body simulations of a LambdaCDM universe, and include such effects as finite quasar lifetimes and output on the lightcone, which affects the shape of quasar light echoes. We create Lya forest spectra that would be generated in the presence of such a fluctuating radiation field, finding that the power spectrum of the Lya forest can be suppressed by as much as 15 % for modes with k=0.05-1 Mpc/h. This relatively small effect may have consequences for high precision measurements of the Lya power spectrum on larger scales than have yet been published. We also investigate another radiation field probe, the cross-correlation of quasar positions and the Lya forest. For both quasar lifetimes which we simulate (10^7 yr and 10^8 yr), we expect to see a strong decrease in the Lya absorption close to other quasars (the ``foreground'' proximity effect). We then use data from the Sloan Digital Sky Survey First Data Release to make an observational determination of this statistic. We find no sign of our predicted lack of absorption, but instead increased absorption close to quasars. If the bursts of radiation from quasars last on average < 10^6 yr, then we would not expect to be able to see the foreground effect. However, the strength of the absorption itself seems to be indicative of rare objects, and hence much longer total times of emission per quasar. Variability of quasars in bursts with timescales > 10^4yr and < 10^6 yr could reconcile these two facts.

Figures (17)

• Figure 1: The effective optical depth $\overline{\tau}_{\rm eff}$ seen by photon crossing through the neutral hydrogen associated with dark matter with the mass of 8 dark matter simulation particles, as a function of density in units of the mean. We show results from the d250 simulation (green points) and d500 simulation (red points), as well as parametric fits (solid lines).
• Figure 2: A test of the raytracing method in a model of a single quasar embedded in a uniform medium. We plot the mean radiation intensity at a given distance $r$ from the quasar. The solid line shows the theory curve, assuming that the radiation attenuation length at 912 Å is $r_{\rm att}=111 \;h^{-1}{\rm Mpc}$ (measured from gasdynamical simulations). The points show the radiation intensity measured in volume filled with poisson distributed particles. A single quasar was placed in the center of the volume and the raytracing algorithm (§ 3.1) used to calculate the intensities. We average the intensity in cubical cells, and the error bars on the points are the standard deviation among cells which fall in the same $r$ bin.
• Figure 3: The intensity as a function of angle in rings at a distance of $10, 20, 40, 80, 160 \;h^{-1}{\rm Mpc}$ from a single quasar in the simulation density field. The straight lines show the uniform attentuation approximation. Normalized so that $J=1$ at $r=r_{\rm att}$ for uniform approximation.
• Figure 4: Clustering of quasars measured in real space. We show the correlation function for two different quasar lifetimes and for isotropic and anisotropic emission.
• Figure 5: The radiation intensity field ($J$, in units of the mean) in $25 \;h^{-1}{\rm Mpc}$ thick slices through one of the d250 simulation volumes (box side length $250 \;h^{-1}{\rm Mpc}$) at $z=3$. All panels show the same region of space, but with different assumptions/ approximations for the radiation emitted from quasars. The left panels are for quasars with a lifetime $t_{q}=10^{7}$ yrs and the right panels for $t_{q}=10^{8}$ yrs. Panels (a) and (b) are not output on the lightcone and have no quasar beaming. Panels (c) and (d) include lightcone effects (the observer's line of sight is the y-axis) and isotropically emitted radiation. In panels (e) and (f), radiation is emitted from quasars in a cone of opening angle 90 degrees, and lightcone effects are included (see text). Quasars visible to an observer are shown as points.