We performed spectroscopic observations of the small-separation lensed quasar SDSS J1001+5027, whose images have an angular separation $\theta \sim 2.^{\!\!\prime\prime}86$, and placed constraints on the physical properties of gas clouds in the vicinity of the quasar (i.e., in the outflowing wind launched from the accretion disk). The two cylinders of sight to the two lensed images go through the same region of the outflowing wind and they become fully separated with no overlap at a very large distance from the source ($\sim 330$ pc). We discovered a clear difference in the profile of the CIV broad absorption line (BAL) detected in the two lensed images in two observing epochs. Because the kinematic components in the BAL profile do not vary in concert, the observed variations cannot be reproduced by a simple change of ionization state. If the variability is due to gas motion around the background source (i.e., the continuum source), the corresponding rotational velocity is $v_{rot}\geq 18,000$ km/s, and their distance from the source is $r\leq 0.06$ pc assuming Keplerian motion. Among three MgII and three CIV NAL systems that we detected in the spectra, only the MgII system at $z_{abs} = 0.8716$ shows a hint of variability in its MgI profile on a rest-frame time scale of $\Delta t_{rest}$ $\leq 191$ days and an obvious velocity shear between the sightlines whose physical separation is $\sim 7$ kpc. We interpret this as the result of motion of a cosmologically intervening absorber, perhaps located in a foreground galaxy.

We study the internal structure of the Circum-Galactic Medium (CGM), using 29 spectra of 13 gravitationally lensed quasars with image separation angles of a few arcseconds, which correspond to 100 pc to 10 kpc in physical distances. After separating metal absorption lines detected in the spectra into high-ions with ionization parameter (IP) $>$ 40 eV and low-ions with IP $<$ 20 eV, we find that i) the fraction of absorption lines that are detected in only one of the lensed images is larger for low-ions ($\sim$16%) than high-ions ($\sim$2%), ii) the fractional difference of equivalent widths ($EW$s) between the lensed images is almost same (${\rm d}EW$ $\sim$ 0.2) for both groups although the low-ions have a slightly larger variation, and iii) weak low-ion absorbers tend to have larger ${\rm d}EW$ compared to weak high-ion absorbers. We construct simple models to reproduce these observed properties and investigate the distribution of physical quantities such as size and location of absorbers, using some free parameters. Our best models for absorbers with high-ions and low-ions suggest that i) an overall size of the CGM is at least $\sim$ 500 kpc, ii) a size of spherical clumpy cloud is $\sim$ 1 kpc or smaller, and iii) only high-ion absorbers can have diffusely distributed homogeneous component throughout the CGM. We infer that a high ionization absorber distributes almost homogeneously with a small-scale internal fluctuation, while a low ionization absorber consists of a large number of small-scale clouds in the diffusely distributed higher ionized region. This is the first result to investigate the internal small-scale structure of the CGM, based on the large number of gravitationally lensed quasar spectra.

While narrow absorption lines (NALs) are relatively stable, broad absorption lines (BALs) and mini-BAL systems usually show violent time variability within a few years via a mechanism that is not yet understood. In this study, we examine variable ionization state (VIS) scenario as a plausible mechanism, as previously suspected. Over three years, we performed photometric monitoring observations of four mini-BAL and five NAL quasars at $z_{\rm em}$ $\sim$ 2.0 - 3.1 using the 105 cm Kiso Schmidt Telescope in $u$, $g$, and $i$-bands. We also performed spectroscopic monitoring observation of one of our mini-BAL quasar (HS1603+3820) using the 188-cm Okayama Telescope over the same period as the photometric observations. Our main results are as follows: (1) Structure function (SF) analysis revealed that the quasar UV flux variability over three years was not large enough to support the VIS scenario, unless the ionization condition of outflow gas is very low. (2) There was no crucial difference between the SFs of mini-BAL and NAL quasars. (3) The variability of the mini-BAL and quasar light curves was weakly synchronized with a small time delay for HS1603+3820. These results suggest that the VIS scenario may need additional mechanisms such as a variable shielding by X-ray warm absorbers.

We study the geometry and the internal structure of the outflowing wind from the accretion disk of a quasar by observing multiple sightlines with the aid of strong gravitational lensing. Using Subaru/HDS, we performed high-resolution ($R$ $\sim$ 36,000) spectroscopic observations of images A and B of the gravitationally lensed quasar SDSS J1029+2623 (at $z_{em}$ $\sim$ 2.197) whose image separation angle, $\theta$ $\sim$ 22$^{\prime\prime}\!\!$.5, is the largest among those discovered so far. We confirm that the difference in absorption profiles in the images A and B discovered by Misawa et al. (2013) remains unchanged since 2010, implying the difference is not due to time variability of the absorption profiles over the delay between the images, $\Delta t$ $\sim$ 744 days, but rather due to differences along the sightlines. We also discovered time variation of C IV absorption strength in both images A and B, due to change of ionization condition. If a typical absorber's size is smaller than its distance from the flux source by more than five orders of magnitude, it should be possible to detect sightline variations among images of other smaller separation, galaxy-scale gravitationally lensed quasars.