LICK OPTICAL-SPECTRA OF QUASAR HS-1946+7658 AT 10 KILOMETERS PER SECOND RESOLUTION - LYMAN-ALPHA FOREST AND METAL ABSORPTION SYSTEMS

We present optical spectra of the most luminous known QSO HS 1946+7658 (z(em) = 3.051). Our spectra have both full wavelength coverage, 3240-10570 angstrom, and in selected regions, either high signal-to-noise ratio, SNR congruent-to 40-100, or unusually high approximately 10 km s-1 resolution, and in parts of the Lyalpha forest and to the red of Lyalpha emission they are among the best published. We find 113 Lyalpha systems and six metal-line systems, three of which are new. The metal systems at z(abs) = 2.844 and 3.050 have complex velocity structure with four and three prominent components, respectively. We find that the system at z(abs) = 2.844 is a damped Lyalpha absorption (DLA) system, with a neutral hydrogen column density of log N(H I) = 20.2+/-0.4, and it is the cause of the Lyman limit break at lambda approximately 3520 angstrom. We believe that most of the H I column density in this system is in z(abs) = 2.8443 component which shows the strongest low-ionization absorption lines. The metal abundance in the gas phase of the system is [M/H] congruent-to -2.6 +/- 0.3, with a best estimate of [M/H] = -2.8, with ionization parameter log GAMMA = -2.75, from a photoionization model. The ratios of the logarithmic abundances of C, O, Al, and Si are all within a factor of 2 of solar, which is important for two reasons. First, we believe that the gas abundances which we measure are close to the total abundances, because the ratio of aluminum to other elements is near cosmic, and Al is a refractory element which depletes very readily, like chromium, in the interstellar medium. Second, we do not see the enhancement of O with respect to C of [O/C] congruent-to 0.5-0.9 reported in three partial Lyman limit systems by Reimers et al. (1992) and Vogel & Reimers (1993); we measure [O/C] = -0.06 for observed ions and [O/C] congruent-to 0.2 after ionization corrections, which is consistent with solar abundances. We see C II*(lambda1335) offset by 15 km s-1 with respect to C II(lambda 1334), presumably because the gas density varies from 2 to 8 cm-3 with changing velocity in the DLA system. These densities imply that the damped component is 6-25 pc thick, which is reasonable for a single cloud in a cold spiral disk. They also imply that the cloud is relatively highly ionized with more C III than C II, more O III than O I, and log N(H I) = 20.72, which is 3 times the H I column. The system at z(abs) = 1.7382 is also believed to be damped with N(H I) approximately 10(21) cm-2, because we see Cr II, but its Lyalpha line will never be seen because it is below the Lyman limit of the other DLA system. We see a 2.6 sigma lack of Lyalpha forest lines well away from the QSO redshift, which may be a chance fluctuation. We also see a correlation between column density N(H I) and Doppler parameter b for 96 unsaturated Lyalpha forest absorption lines, and although this correlation persists in the 36 Lyalpha lines which lie in regions where the SNR congruent-to 8-16, we agree with Rauch et al. (1993) that it is probably a bogus effect of low SNR. The same applies to lines with very low b values: in regions where SNR less-than-or-equal-to 8 we see many Lyalpha lines which appear to have 10 less-than-or-equal-to b less-than-or-equal-to 20, but when 8 less-than-or-equal-to SNR less-than-or-equal-to 16 we see only one line with b less-than-or-equal-to 15 km s-1, and two others which we believe have b less-than-or-equal-to 20, with values of 20 and 16 km s-1. Traditional Lyalpha line samples which include all lines which have W/sigma(W) greater-than-or-equal-to 4 are not adequate to explore the distribution of the properties of individual clouds, because we need much higher W/sigma(W) and SNR to avoid the strong biases.