We develop a prescription for characterizing the strengths of metal lines associated with Ly \alpha forest absorbers ( LYFAs ) of a given neutral hydrogen column density N _ { HI } and metallicity [ O/H ] .
This Line Observability Index ( LOX ) is line-specific and translates , for weak lines , into a measure of the equivalent width .
It can be evaluated quickly for thousands of transitions within the framework of a given model of the Ly \alpha forest , providing a ranking of the absorption lines in terms of their strengths and enabling model builders to select the lines that deserve more detailed consideration , i.e .
those that should be detectable in observed spectra of a given resolution and signal-to-noise ratio .
We compute the LOX for a large number of elements and transitions in two cosmological models of the Ly \alpha forest at z \sim 3 derived from hydrodynamic simulations of structure formation .
We present results for a cold dark matter universe with a cosmological constant ; an \Omega = 1 cold dark matter model yields nearly identical results , and we argue more generally that the LOX predictions are insensitive to the specific choice of cosmology .
We also discuss how the LOX depends on redshift and on model parameters such as the mean baryonic density and radiation field .

We find that the OVI ( 1032 Å , 1038 Å ) doublet is the best probe of the metallicity in low column density LYFAs ( N _ { HI } \approx 10 ^ { 14.5 } { cm } ^ { -2 } ) .
Metallicities down to [ O/H ] \sim -3 yield OVI absorption features that should be detectable in current high-quality spectra , provided that the expected position of the OVI feature is not contaminated by HI absorption .
The strongest transitions in lower ionization states of oxygen are OV ( 630 Å ) , OIV ( 788 Å ) , and OIII ( 833 Å ) .
These absorption lines are all predicted to be stronger than the OVI feature , but even at redshifts 3 - 4 they will have to be observed in the ultraviolet , and they are extremely difficult to detect with present UV instruments , such as the Space Telescope Imaging Spectrograph ( STIS ) .
At lower redshifts , detection of these lines may be possible in STIS spectra of the very brightest QSOs , while one may have to wait for next-generation instruments such as the Cosmic Origins Spectrograph ( COS ) to detect such lines in a number of high-redshift QSOs .

The strongest metal lines with restframe wavelength larger than 912 { \AA } associated with higher column density LYFAs at z \sim 3 are CIII ( 977 Å ) and SiIII ( 1206.5 Å ) , which peak at N _ { HI } \sim 10 ^ { 17 } { cm } ^ { -2 } .
Of the lines with rest wavelengths \lambda _ { r } > 1216 { \AA } , which can potentially be observed redwards of the Ly \alpha forest , the CIV ( 1548,1551 ) doublet is expected to dominate in all LYFAs , regardless of the value of N _ { HI } .
We argue that CIV and CII absorption may peak in different spatial regions , and that comparison of single-phase models of the CIV/CII ionization ratios with observed CIV/CII column density ratios can lead to an overestimate of the ionization parameter in the central parts of the absorbers .