The power spectrum of the flux distribution in the Lyman-alpha forest of a Large sample of UVES QSO Absorption Spectra (LUQAS)

TL;DR

The LUQAS study delivers a high-precision measurement of the Lyman-$\alpha$ forest flux power spectrum from 27 UVES QSOs, enabling detailed comparisons with Croft et al. (2002) and McDonald et al. (2000). By employing multiple flux estimators and careful treatment of continuum, DLAs, and metal lines, the authors show that the 1D spectrum is smooth on small scales while the 3D spectrum is noisier, with a peak amplitude near $k\approx 0.03\,\mathrm{s/km}$ consistent with prior work. After accounting for mean flux evolution and using simulations to calibrate the flux-to-matter mapping, the LUQAS results align with previous measurements in amplitude and display redshift evolution indicative of gravitational growth. The work highlights the importance of continuum fitting and metal-line contamination at large scales and underscores the role of simulations in translating flux statistics into constraints on the matter power spectrum and cosmological parameters.

Abstract

The flux power spectra of the Lyman-alpha forest from a sample of 27 QSOs taken with the high resolution echelle spectrograph UVES on VLT are presented. We find a similar fluctuation amplitude at the peak of the ``3D'' flux power spectrum at k ~ 0.03 (km/sec)^(-1) as the study by Croft et al. (2002), in the same redshift range. The amplitude of the flux power spectrum increases with decreasing redshift if corrected for the increase in the mean flux level as expected if the evolution of the flux power spectrum is sensitive to the gravitational growth of matter density fluctuations. This is in agreement with the findings of McDonald et al. (2000) at larger redshift. The logarithmic slope of the "3D" flux power spectrum, P_F(k), at large scales k < 0.03 (km/sec)^(-1), is 1.4 +- 0.3, i.e. 0.3 shallower than that found by Croft et al. (2002) but consistent within the errors.

Figures (9)

• Figure 1: Number of spectra for which the Lyman-$\alpha$ absorption region covers a given redshift (bottom axis) and wavelength (top axis). The median redshift of the whole sample is $z=2.25$. The total redshift path is $\Delta z=13.75$ .
• Figure 2: Effect of continuum fitting on the 1D flux power spectrum. Circles are for the continuum fitted spectrum $F2= e^{-\tau}/<e^{-\tau}>-1$, solid triangles are for the not continuum fitted spectrum $F3 = I_{\rm obs}/<I_{\rm obs}>-1$. Dashed and solid line are for spectra smoothed with a Gaussian window of 25 and 50 Å width, respectively, as described in the text. Error bars denote the $1\sigma$ errors of the mean values. Open triangles show the flux power spectrum of $F3$ for the region redwards of the Lyman-$\alpha$ emission line (1265.67 - 1393.67 Å).
• Figure 3: Effect of metal lines. The solid curve shows the metal line contribution to the total flux power spectrum of the Lyman-$\alpha$ forest region. Metal-line only spectra have been generated from the line lists of 13 QSOs fitted with VPFIT. The mean redshift of this subsample is $<z>=2.36$.
• Figure 4: The 1D and "3D" flux power spectrum of the full sample for the flux estimator $F1=\exp(-\tau)$. Errors bars are jackknife estimates. Note that the LUQAS points (diamonds) in the left panel are different from the one in the published version in Kim et al. (2004) MNRAS 347, 355 (where the $k$ values of the 1D flux power spectrum had been erroneously shifted by half a bin size in $\log k$).
• Figure 5: "3D" power spectrum of two different flux estimators. Circles are for $F2= e^{-\tau}/<e^{-\tau}>-1$ calculated from continuum fitted spectra and triangles are for $F3_{s25} = I_{\rm obs}/<I_{\rm obs}>_{s25}-1$ calculated from not continuum fitted spectra. $F3$ is smoothed with a Gaussian window of width 25 Å. See Section \ref{['conteff']} for details. Error bars are jackknife estimates calculated from the sample of 27 QSOs.