1. In Section 3.4, you classify absorbers into three discrete phases (cold, cool, warm) based on specific ions and compare them to CLOUDY PIE models. However, the CGM is known to be multiphase and dynamically complex, with overlapping ionization conditions.
Why did you assume a strict one-to-one mapping between ions and phases, and how do you justify the exclusion of mixed-phase or non-equilibrium ionization scenarios (e.g., turbulent mixing layers, non-equilibrium cooling) that could significantly alter inferred cloud densities, temperatures, and especially cloud sizes (Fig. 12)?
Given that many components are saturated and likely blend multiple unresolved clouds, does your assumption of single-phase PIE models risk underestimating cloud sizes and misrepresenting the true physical state of the gas?
2. Your sample is selected based on QSOs within ~3.5 RHIRHI of galaxies with FUV-bright backgrounds.
How do you account for the potential bias introduced by excluding sightlines with poor data quality (e.g., J0235–0925) or those contaminated by intervening systems (e.g., J1024+2422)?
Could this selection artificially enhance the perceived correlation between metal detection and small ρ/RHIρ/RHI, particularly since most non-detections lie at larger radii? Additionally, given the small sample size (31 galaxies), how robust is your conclusion that ρ/RHIρ/RHI is a better tracer than ρ/Rvirρ/Rvir, especially when comparisons with COS-Halos/GASS rely on heterogeneous data and differing selection criteria?