Astron. Astrophys. 337, 591-602 (1998)

5. Discussion

We will now summarize the agreements and disagreements between the predicted and the observed spectra. There is agreement that the spectrum of a large and deeply penetrating bolide is generated by a superposition of thermal continuum, molecular emission bands and atomic emission lines. The general shape of the continuum in the panchromatic region, i.e. nearly no dependence on wavelength, is also confirmed. All predicted atomic lines are present. Finally, the effective excitation temperature as determined from the atomic lines is about 5000-6000 K in both predicted and observed spectra (the main component).

The major difference is the lower column density of the Fe I atoms and the high ratios of Ca I /Fe I and Na I /Fe I in the theoretical spectra. The ratios can be explained by a different degree of ionization. Both calcium and sodium have low ionization potentials and their low abundance in the observed spectrum is explained by their high ionization. The effect of incomplete evaporation of calcium is of minor importance here. A higher ionization for the same temperature means a lower density. The column density of Fe I is, however, higher. It seems that there is more vapor in the radiating region but with lower density, so the radiating region must be much larger than modeled.

The lower density can also explain the lower continuum level because a large part of the continuum is produced by free-free emission and this grows with the square of density. A lower optical thickness of the continuum could also allow the radiation produced by the Si II lines in the vicinity of the shock wave not to be absorbed and to be present in the spectrum.

It seems that a substantially larger volume occupied by the radiating vapor would explain most of the differences between the theoretical and the observed spectra. Indeed Borovicka & Spurný (1996) interpreted the Beneov spectrum as being produced in a cylinder of a diameter about 20 meters, i.e. at least 10 times the diameter of the body. We therefore computed the spectrum of a 10 m wide layer of vapors with a temperature of 5000 K and a density of - g cm^-3. The same spectral opacities as for the modeled spectra were used. The shape of the resulting spectrum proved to be closer to the observation.

However, the gas dynamical modeling shows that all vapor in the vicinity of the body must be confined by the shock wave and the diameter of the allowed region is no more than twice the diameter of the body (see Fig. 2). From the gas dynamics point of view, it is impossible to produce a 20 meter diameter vapor region in the vicinity of a 2 meter diameter single body. The sum of a number of individual smaller bodies produces spectra closer to the observation than a single body but still worse than an extended cloud of vapors.

Again, meteoroid fragmentation seems to be the most plausible explanation of this discrepancy. If the bolide is formed not by a single big body but by a swarm of bodies separated by several meters, the interaction between the shock waves and the vapor may lead to the formation of a common radiating volume at some distance from the head of the bolide. The interaction process has been investigated by Artem'eva & Shuvalov (1996) for 17-27 fragments. The shape of the luminous volume becomes more like a cylinder with the ratio of the radius to the length larger than for a single body with the summary mass. Of course, the radius depends on the distance between the fragments. However, we cannot use the results of these simulations for a more detailed analysis of the spectra as the radiation has not been included into the model yet.

Although the fragmentation is certainly important, the problem with the size of the luminous volume is also present at altitudes around 60 km, where the meteoroid was probably not fragmented. We have to consider the possibility that the main radiating region does not surround the body but is located behind the body. This might occur, if the material is ablated from the body in the form of tiny solid fragments or liquid drops and evaporated later. The vapor then tends to fill the cavern formed by the rarefied wake of the main body and the vapor volume increases and the density decreases. This has been confirmed by a 2D purely hydrodynamic simulation (V.V. Shuvalov, personal communication). More detailed 2D radiation hydrodynamic simulations of such a scenario should be conducted and the resulting spectrum should be compared with observations.

There also remains the question of the primary mechanism which drives the ablation and excitation of radiating atoms. Whether it is the radiation of the shock wave as supposed in many theoretical papers, e.g. Nemtchinov et al. (1994), or the thermal collisions in the hot gas as supposed by Borovicka (1993), or both. In the theoretical model under consideration (Golub' et al. 1996a) both mechanisms are actually acting. The first one (radiation) dominates for large bodies or (and) at low altitudes, the second (collisions, or molecular and electronic thermal transfer) dominates at high altitudes and for small bodies. The observed spectrum does not resolve the question of which mechanism dominated in the Beneov case because the 5000 K vapors can be formed both ways and the exact origin of the continuum is not clear. Moreover, there is another mechanism, a turbulent heat transfer, which should additionally be taken into account in the future.

Some observed features are beyond the scope of the present model. The wake radiation at high altitudes needs non-equilibrium modeling. For the description of the persistent radiating cloud at the position of the bright flare at 24 km the quasi-stationary ablating piston model should be replaced by a non-stationary model of an expanding cloud. The recombination of "reexisting" molecules of oxides and an additional oxidation of metals by the air due to turbulent diffusion should be taken into account.

The spectral analysis of the Beneov bolide confirmed the important role of meteoroid fragmentation revealed already from the analysis of the dynamics and light curve in Paper I. Moreover, it was shown that the radiation is not produced by individual independent fragments but by a more complicated configuration. The details of the ablation mechanisms both before and after the fragmentation and the mutual interaction of the fragments are essential for the description of the radiation. This is important not only for the understanding of the meteoroid-atmosphere interaction but also for the derivation of the mass, density, structure and composition of meteoroids from the observation of bolides. The Beneov-like meteoroids are intermediate in size between smaller meteoroids where the conventional analysis gives reliable results (e.g. Ceplecha 1996, Ceplecha et al. 1996) and very large meteoroids where the simple pancake model may probably be a usable approximation.

Our physical model requires further improvement. First, the model probably underestimates the ablation rate from the rough surface of real meteoroids with non-uniform structure (Popova & Nemtchinov 1996). Second, the model underestimates the role of the kinetic energy of small fragments and liquid droplets thrown away from the rough surface of the meteoroid. Third, turbulent diffusion of the vapor in the air increases the size of the volume occupied by the vapor and decreases vapor density. All these factors should be more correctly incorporated into future theoretical models and codes. We do not anticipate that all these factors may qualitatively change our general description of the Beneov meteor phenomenon and our estimates of the meteoroid mass.

© European Southern Observatory (ESO) 1998

Online publication: August 17, 1998
helpdesk.link@springer.de