Low-redshift quasar Lyα absorption-line systems associated with galaxies

In this paper we present semi-analytic models and Monte-Carlo simulations of QSO Lyabsorption line systems which originate in gaseous galactic haloes, galaxy discs and dark matter (DM) satellites around big central haloes. The aim is to estimate the number density per unit redshift of Lyabsorption lines related with galaxies and to investigate the properties of the predicted galaxy/absorber systems, such as equivalent widths Wr, projected distances �, galaxy luminosities LB, as well as absorber redshifts z. It is found that for strong Lyabsorption lines (Wr � 0.3u galactic haloes and satellites can explain � 20 per cent and 40 per cent of the line number density of the HST QSO Absorption Line Key Project respectively. The population of DM satellites is adopted from numerical simulations by Klypin et al. (1999). If big galaxies indeed possess such large numbers of DM satellites and they possess gas, these satellites may play an important role for strong Lylines. However the predicted number density of Lyman-limit systems by satellites is � 0.1 (per unit redshift), which is four times smaller than that by halo clouds. Including galactic haloes, satellites and HI discs of spirals, the predicted number density of strong lines can be as much as 60 per cent of the HST result. The models can also predict all of the observed Lyman-limit systems. For strong lines the average covering factor within 250h 1 kpc is estimated to be � 0.36, which is in good agreement with observations. And the effective absorption radius of a galaxy (with unit covering factor) is estimated to be � 150h 1 kpc. There exist correlations of Wr versus �, LB and z. The models predict Wr / � � L � (1+z) with � � 0.5,� � 0.15, � 0.5. To compare with results of imaging and spectroscopic surveys, we study the se- lection effects of selection criteria similar to the surveys. We simulate mock observa- tions through known QSO lines-of-sight and find that selection effects can statistically tighten the dependence of line width on projected distance. This result confirms pre- vious suggestions in the literature. After applying selection criteria, the models can predict similar distributions of Wr, �, LB, absolute magnitudes and absorber red- shifts to those of imaging and spectroscopic surveys. Finally we find that the total redshift interval of present observations (� 5) is not large enough for the models to reveal the real relationships if adopting the selection criteria. An adequate total red- shift interval might be � 10. This may conciliate contraditious conclusions about the anti-correlation of equivalent widths versus projected distances by different authors.