The number density of 0.6<z<1.7 Mg II systems from CORALS:: Observational biases at intermediate redshift

The goal of the Complete Optical and Radio Absorption Line System (CORALS) survey is to quantify the potential impact on QSO absorber statistics from dust in intervening galaxies. Dust may introduce a selection bias in surveys that are based on magnitude-limited QSO samples, leading to an underestimate of absorber number density, n(z). Here we present the results of the second phase of the CORALS survey, which extends our previous work on z > 1.8 damped Lyalpha systems (DLAs) to search for strong metal line systems (candidate DLAs) in the range 0.6 < z < 1.7. We have identified 47 Mg II systems with rest-frame equivalent widths EW(Mg II lambda2796)> 0: 3 8 in our sample of 75 radio-selected quasars. The total redshift path covered by the survey is Deltaz = 35.2, 58.2, and 63.8 for EW(Mg II lambda2796) > 0.3, 0.6, and 1.0 Angstrom thresholds, respectively (5 sigma). Our principal and most robust result is that the n(z) value of low-redshift Mg II systems determined for the CORALS survey is in excellent agreement with that of optically selected, magnitude-limited QSO samples. We use empirically determined Mg II equivalent width statistics to estimate the likely number of DLAs in this sample. The statistically inferred number density of DLAs, n(z) 0.16(-0.06)(+0.08), is consistent with other low-redshift samples, although the large 1 sigma error bars permit up to a factor of 2.5 more DLAs in CORALS. However, confirmation of the DLA candidates, precise evaluation of their n(z) values, and measurement of their H I column densities awaits UV observations with the Hubble Space Telescope. Finally, we report an excess of intermediate-redshift Mg II systems observed toward bright QSOs, which could be due to a lensing amplification bias. However, there is also evidence that this excess could simply be due to more sensitive EW detection limits toward brighter QSOs. We also emphasize that absorber statistics determined from magnitude-limited surveys reach a steady value if the completeness limit is significantly fainter than the fiducial value of the quasar luminosity function.