Discussion

The observed flux of at 1455 Å is the brightest ever recorded for NGC 4151. The differences between the Astro-1 and the Astro-2 spectra are most clearly seen in the ratio of the Astro-1 spectrum to the mean Astro-2 spectrum shown in Figure 2. The continuum is five times brighter on average, obviously more blue than in 1990, and dramatically brighter approaching the Lyman limit. Ignoring the airglow features, the most prominent differences are the deeper absorption lines in the high ionization lines C IV , N V , and O VI and in the hydrogen Lyman series. The narrow emission cores of He II and also stand out. While they are brighter now than in 1990, they have not brightened as much as the continuum or the cores of the broad emission lines, consistent with the suggestion of Clavel et al. (1987)Clavel87 that they form outside the BLR at a radius of . Perhaps the most interesting features are the broad, sharply peaked structures visible on the red wing of C IV , Ly, and, to a lesser extent, O VI . The peak redward of C IV matches the ``satellite" lines L'2 and L2 first reported by Ulrich et al. (1985)Ulrich85, and the profile is quite similar to the C IV emission line itself.

Clavel et al. (1987)Clavel87 noted correlated variability between the C IV satellite lines although they do not correlate with the continuum. Given these characteristics and the observed red and blue shifts, they suggested the lines are formed in jets of material ejected from the nucleus. The L'2 and L2 features are present in both our Astro-1 and Astro-2 spectra, but the L1 line is not. While they vary between our two observations, they are not in step with the rest of the broad line. The shape of the L'2+L2 feature in the ratio of the Astro-1 to the Astro-2 data offers an intriguing possibility --- the satellite line is not intrinsic jet emission, but rather a reflection of the broad line region.

Features that are noticeably unchanged since Astro-1 are the absorption lines of the lowest ionization species, e.g. Si II and C II . Examining the velocities of the absorption lines in Table 1 shows two distinct groupings in velocity, one near and another near . The latter is comprised entirely of the lowest ionization species, and its velocity places it at rest relative to the Galaxy. The lack of variation (a point strongly argued by Veron et al. 1985 using IUE data) and the zero redshift suggest that these lines are in our own interstellar medium rather than in NGC 4151.

Even though the Lyman lines show much stronger absorption in 1995 compared with 1990, the apparent column density of the gas responsible for the bulk of the absorption is now much lower. This implies that the hydrogen is more highly ionized, and that turbulent velocities have increased, effectively increasing the Doppler parameter of the absorbers. The decrease in the neutral hydrogen column of the broad component of the Lyman lines compared to Astro-1 is consistent with photoionization of a warm absorber. An increase of a factor of five in the continuum intensity would lead to a five times decrease in the neutral column if the gas were optically thin in the Lyman continuum. Comparison with Astro-1 columns of at a lower Doppler parameter of (Kriss et al. 1992) implies a decrease of at least 20% in the neutral column.

Given these conditions, is there a warm absorber model that can match the high X-ray columns and the observed UV ionization states? Ignoring the metal lines, one can certainly match the hydrogen column and the X-ray absorbing medium. Models of warm absorbers both in high density clouds near the ionizing source ([Netzer 1993]) and in a more distant, lower density medium ([Krolik & Kriss 1995]) can produce appropriately high total columns of with low neutral columns. Photoionization models similar to those discussed by Krolik & Kriss (1995)KK95 with , temperatures of a few times K, and total columns of produce X-ray absorption typical of that historically seen in NGC 4151 (e.g., [Weaver et al. 1994b]) and match the absorption seen in C IV, N V, and O VI. However, whether the Doppler parameter is low and typical of thermal gas, or turbulent and high as in the Lyman lines, it is hard to match simultaneously these high ionization lines with the relatively high neutral hydrogen column and the high columns observed for lower ionization ions like C III and Si IV.

Once again this implies an absorbing medium with multiple zones. While the complexity is inelegant, this is certainly not unexpected. Kriss et al. (1994b)Kriss94b argued that the UV absorption occurs in a thin atmosphere above the obscuring torus. One expects the medium in the vicinity of the photoionized skin of the torus to be turbulent and inhomogeneous ([Balsara & Krolik 1993]). One could imagine clumps of high density, lower temperature material suspended in a surrounding medium of higher temperature, lower density, more highly ionized gas. As material flows away from the torus and is accelerated into the X-ray heated wind, these higher density clumps will eventually evaporate into the surrounding hotter, lower density medium. Given the low opacity of the broad, turbulent component of the absorbing gas, it is unlikely that it collimates the radiation in NGC 4151. However, our data allow gas with thermal Doppler parameters of to be present in sufficient quantity to be opaque to the ionizing radiation.

This picture implies that our line of sight to NGC 4151 passes close to the surface of the obscuring torus itself since the most profound difference between the warm absorber in NGC 4151 and others such as 3C 351 is the presence of the lower ionization C III lines, particularly the transition indicative of high density gas. The UV-absorbing lines and O VIII edges seen in the X-ray in these other warm absorbers have lower total columns that may represent lower density gas more distant from the surface of the torus. Our future discussion of the shorter timescale variations seen in the UV-absorbing lines of NGC 4151 among the individual Astro-2 observations will give additional insights into the possible structure and location of the absorbing gas.