The bispectrum of the Lyman α forest at z ∼ 2-2.4 from a large sample of UVES QSO absorption spectra (LUQAS)

We present a determination of the bispectrum of the flux in the Lyman alpha forest of quasistellar object (QSO) absorption spectra obtained from a large sample of Ultraviolet Echelle Spectrograph (UVES) QSO absorption spectra (LUQAS), which consists of spectra observed with the high-resolution UVES. Typical errors on the observed bispectrum as obtained from a jack-knife estimator are similar to50 per cent. For wavenumbers in the range 0.03<k<0.1 s km(-1) the observed bispectrum agrees within the errors with that of the synthetic absorption spectra obtained from numerical hydro-simulations of a LambdaCDM model with and without feedback from star formation. Including galactic feedback changes the bispectrum by less than 10 per cent. At smaller wavenumbers, the associated metal absorption lines contribute about 50 per cent to the bispectrum and the observed bispectrum exceeds that of the simulations. At wavenumbers k<0.03 s km(-1), second-order perturbation theory applied to the flux spectrum gives a reasonable (errors smaller than 30 per cent) approximation to the bispectra of observed and simulated absorption spectra. The bispectrum of the observed absorption spectra also agrees, within the errors, with that of a randomized set of absorption spectra where a random shift in wavelength has been added to absorption lines identified with VPFIT. This suggests that for a sample of the size presented here, the errors on the bispectrum are too large to discriminate between models with very different 3D distribution of Lyman alpha absorption. If it were possible to substantially reduce these errors for larger samples of absorption spectra, the bispectrum might become an important statistical tool for probing the growth of gravitational structure in the Universe at redshift z greater than or similar to 2.