3.3.2 The Accretion Disk Model
The standard Shakura-Sunyaev disk model is difficult to reconcile with the almost simultaneous variability in UV, optical and X-rays in NGC 4151 and NGC 5541, indicating that illumination by the hard X-ray source may give rise to a large fraction of the UV-optical flux through reprocessing. 99, 107However, it could be that these sources are atypical examples, as pointed out before. Indeed, there seems to be no correlation between variations in UV-optical and X-rays in most AGN. Ulrich et al. 98 suggested instead that the varying part of the optical flux is the low-energy tail of the UV-bump, thereby producing simultaneous UV-optical variability. They also argued that the latter may be due to disk instabilities in the inner region.
However, one advantage of the illumination model is that the LIL may
be explained this way. Collin-Souffrin and collaborators (e.g., Ref.
108) have
developed such a model, whose basic assumption is an
extended accretion disk (to at least R ~
104rg). One problem with the
model has been the condition of negligible self-gravity, which imposes
an upper limit on 
. In F9 this
would translate into a highly
supersonic turbulence, which seems unphysical. A recent version
32, 103
avoided this problem by adopting a small -value. The LIL emission
region was confined to R ~
102-104rg, since line saturation
for R 
102rg and inadequate heating for R
104rg inhibit line formation in
other regions. The result seems able to meet quite a number of
observational constraints. Thus, the line widths may be explained in
terms of kinematics, since
vK(103rg) ~
103 km s-1, where vK is the
Keplerian rotational velocity. The large column density and covering
factor required to explain the intensity of the LIL and the formation
and confinement problems may also be accounted for. Furthermore, the
constancy of line ratios may be explained by quasi-thermalization of
the lines.
The model requires that a sizeable fraction (~ 1/2) of the
bolometric luminosity is emitted in hard X-rays or 
-rays. The bulk of
the UV luminosity should then be emitted by the accretion disk, of
which a part is due to reprocessing. By combining the fitting
procedures for the UV continuum and line emission from the disk, the
obtained accretion rate was much lower than the one estimated from the
bolometric luminosity. One explanation may be that the ``real''
accretion rate is higher, but that only part of the liberated
gravitational energy comes out in optical-UV. The rest may be
dissipated in a hot, optically thin phase. as in the Haardt and
Maraschi model discussed above.
The applicability of these results to other AGN such as OSOs remains an open issue. Even though rapid broad line variability has been detected in QSOs, the under-sampling usually precludes definite conclusions. The BLR lines also seem to respond to a continuum change in a very complicated fashion. Furthermore, some sources show rapid continuum variability, but no line variability. Such null-results may yield interesting constraints on, e.g., the BLR geometry.
To summarize, a picture of a much smaller BLR and brighter continuum source than previously thought has emerged. The most promising method to map out the BLR may involve further developments of the cross-correlation technique, such as the echo mapping (1D, 2D or 1D + time) proposed by Welsh and Horne 109 and further discussed by Horne.110 The number of unsettled issues is still large, some of which include: