Annu. Rev. Astron. Astrophys. 1997. 35: 445-502 Copyright © 1997 by Annual Reviews. All rights reserved

5.4. Summary, Perspectives, and Emerging Fields in Emission Line Variability Studies

THE MOST IMPORTANT RESULTS

1. The dimension of the line emitting region can be derived from the time delays between continuum and line variations. The delay depends on the line and on the velocity.

2. The velocity field of the broad line gas appears to be dominated by virialized motions in the gravity field of the black hole (the fastest moving gas is closest to the center; the blue and the red wings of the emission lines, to first order, vary simultaneously, as expected if radial motions are not important). The black hole mass can then be derived from the observed motions.

3. The variability observations can be best understood in the frame of the "accretion disk plus wind" model. ⁽¹⁾ The results of variability studies strengthen or are consistent with this model. Specifically, the first-order similarity between the responses of the blue and red wings of the lines supports the model. A second-order effect is observed in three AGN in the form of some asymmetry between the blue- and red-wing responses, with the red wing leading the blue; such a difference has a natural explanation in wind models. Finally, hints for different values of the transfer function at zero lag in HIL and LIL have been found in several AGN. If this effect is confirmed in those and other AGN, it will add strong support to this model.

In conclusion, the gas in the radio-quiet AGN central region has a roughly ordered velocity field: rotation in a disk for the low-ionization medium and rotation plus outflow for the high-ionization medium, which is pulled out from the disk by magnetic or radiative forces. The details of the disk plus wind structure are complex. For example, changes in the HIL and LIL transfer functions over several years are best understood as changes in the distribution of the HIL and LIL gas on time scales roughly commensurate with the dynamical time of the inner BLR in low-luminosity AGN. Together with the complexity of the line profile variations (Figure 5), they suggest the presence of inhomogeneities on the disk surface and in the outflowing medium.

THE LIMITS AND THE REMEDIES     Derived transfer functions should in general be regarded with caution in spite of the successful challenge of Horne (Section 5.2). The physical situation is bound to be complicated by the gas inhomogeneities, local overlap in velocities (e.g. Done & Krolik 1996), and imperfect correlation between the measured continuum and the ionizing continuum. The calculation of the transfer function itself is limited not by the techniques but by the data. "Improving the Signal-to-Noise ratio and sampling by modest factors of 3 would greatly sharpen the velocity-delay maps" (Horne 1994).

Such an improvement represents a rather tall order but appears to be necessary in order to reach firm conclusions on the distribution and velocity field of the gas and the anisotropy of the line emission in each cloud (Horne 1994). In the future, such excellent data will allow full utilization of the information in the profile variations (something that has not yet been done systematically) and will motivate the development of inversion methods incorporating more specific geometry and physics into the models, such as photoionization codes. Meanwhile, the universality of results obtained predominantly from a few low-luminosity radio-quiet AGN selected as targets because they were the most active AGN about 10 years ago should be regarded with caution.

THE EMERGING FIELDS     A number of avenues are ripe for future variability investigations. First, the entire parameter space defined by the black hole mass and the accretion rate should be explored. Several subsets of AGN populate the faint end of the optical/UV luminosity function. Intrinsically low-luminosity broad-line AGN (evidently with small accretion rates but with unknown black hole masses) have been discovered in 13% of nearby galaxies (Ho et al 1995). Their continuum and line variability is essentially unknown, and what little we know is puzzling, as evidenced by the fact that in M81, after about 15 years of no detectable change in the Balmer lines, a broad double-peaked component has recently appeared (Peimbert & Torres-Peimbert 1981, Ho et al 1996, Bower et al 1996).

From HST observations, LINERS (galaxies with low-ionization narrow emission-line regions whose spectra resemble neither an HII region nor the narrow-line spectrum of a broad-line AGN; Heckman 1980) have recently been found to harbor point-like nuclear UV sources in 25% of the cases (Maoz et al 1995). The absence of a nuclear UV source in 75% of the observed LINERS may reflect their duty cycle (Eracleous et al 1995a). The study of the continuum variability and the search for broad lines will greatly improve our understanding of LINERS.

Finally, for the NLS1 - where the accretion rate may be exceptionally high (Section 2.2) - it is important to determine the central mass through systematic investigations of line variability.

All of the above AGN have in common that the broad lines are weak or only moderately strong. Their spectral variability studies require the use of HST or of large or very large optical telescopes. The question of the duty cycle of AGN, and the passage of a given AGN from one subset to another (from LINERS to "classical" broad-line Seyfert 1 to NLS1), requires good statistics on the number of AGN in each subset and long-term monitoring to witness the passage from one class to another, as occurred with Pictor A and NGC 1097 (Section 5.3, Appearance and Disappearance of Broad Emission Components). In fact, with the development of electronic archives, one could imagine light curves extending several centuries or more. (And we would not even be pioneers because records of supernova can be found in Chinese and other archives that are a millenium or older).

At the bright end of the optical/UV luminosity function, the monitoring of line and continuum variability will allow us to follow the dependence of the line response on the AGN luminosity (cf the non-detection of significant Ly variations in 3C 273, Section 5.1), and perhaps determine the central mass in the brightest objects in the universe. Powerful large field multi-object spectrographs (for example, installed on Schmidt telescopes) should be the instruments of choice.

Second, the recent observations of the profile and variability of the Fe K line (Tanaka et al 1995, Yaqoob et al 1996) and of the absorption lines originating in the hottest part of the warm absorber (Iwazawa et al 1997) are new ways to explore the innermost regions of AGN and will develop into a flourishing field with the Advanced X-Ray Astrophysics Facility, the X-Ray Multi-mirror Mission, and other X-ray missions.

The third promising avenue is that of theoretical investigations. This is hardly an emerging field, but new data will elicit new models and calculations. The accretion disk plus wind model appears particularly promising as it offers explanations for the variability of the broad lines, the two phases of the BLR gas, the blueshift of the emission and the absorption lines, and the emission of the UV continuum. This model and others should be pursued to the point where they can predict line intensity and profile variability patterns that can be compared with observations.

The effect of the certain presence of the central star cluster should also be investigated. The stars can be accreted (cf NGC 4552, Renzini et al 1995), collide with one another (Courvoisier et al 1996), or be trapped by the disk (Artymowicz et al 1993), and their atmospheres contribute to the BLR (Alexander & Netzer 1994, Armitage et al 1996). All these processes cause line and continuum variations.

¹ This model is the simplest architectural sketch explaining the two basic results of spectroscopic studies: 1. The HIL and the LIL come from two media with distinctly different physical conditions and 2. these media have different velocity fields since the high ionization lines are blueshifted (a shift that is small compared to the full width of the lines). This model attributes the emission of the LIL to the accretion disk, whereas the HIL are emitted by a hotter more diffuse outflowing medium; that the disk is opaque is implicit in this model. Back.