5. Variability of semiregular variables
The current data are not sufficient to determine the cause of irregular variations in late type stars. These irregular variations have been found in semiregular, irregular and mira-type variables. The irregularities found in mira velocity curves are of similar size to the velocity variations in the short period SRVs (see the data for Cyg in Hinkle et al. 1982) and might have a common cause.
We showed that both pulsation and convective cells can in principle produce velocity and light changes. The standard pulsation interpretation is in nice agreement with the observed correlation of velocity and light changes. On the other hand it is very difficult to estimate the size of the light and velocity variations due to convective cells without detailed modelling. The interesting feature of convective cells is the irregular variability and the systematic velocity shift they produce.
There are several features found in semiregular variables that cannot be explained without the assumption of stellar pulsation. A number of SRVs like W Hya have periods a few times the timescale estimated for convective cells. Furthermore, their rather large velocity amplitude and the crossing of the center-of-mass velocity by the velocity curve make an explanation with convective cells alone for the long period objects rather unlikely. It has also to be kept in mind that several semiregular variables at least sometimes show hydrogen emission lines of the Balmer series in their spectra. In the miras, and presumably in the SRVs, this emission component originates from shock waves in the stellar atmospheres. These shock waves are the result of global stellar pulsation, giving a strong argument for a pulsational origin for the light curve and velocity variations.
Models predict the existence of large convective cells at the surface of red giants. It is therefore very likely that the observed variability in light and velocity includes a component provided by these surface structures. This component will be found among small amplitude irregulars as well as among miras and may account for the fact that many `nonvariable' red giants are actually variables with small amplitudes (see e.g. Jorissen et al. 1997). If one wants to explain this small amplitude variability with the help of large convective cells it is critical to know the center-of-mass velocity very accurately. Variability due to convective cells only should lead to a measured velocity smaller than the center-of-mass velocity (i.e. always outflow of matter) at all phases.
It is well known that miras do not always reach the same brightness in each light maximum. The differences (see e.g. Fig. 5) can be of the order of one magnitude. Such `irregularities' may be produced by processes on longer time scales than the main stellar pulsation period (e.g. Höfner & Dorfi 1997). This illustrates that pulsation itself, even if only one mode is excited, can introduce irregularities into the light change of long period variables.
At the moment it is not possible to estimate whether convective motion can introduce the observed irregularity into the light and velocity changes alone or not. The variability of the objects with the smallest amplitudes might be explained by convective motion only. For those cases it does not seem necessary to explain the small scale variability with the help of high degree overtone modes as suggested by Percy & Parkes (1998). However, with the actual data available we cannot exclude regular pulsation or chaotic behavior in these objects, but we want to support the possibility of surface structures caused by convective motion as an alternative explanation.
To summarize, we think it is a reasonable approach to see the observed variability in semiregular and irregular variables as a composite of pulsation and an (additional?) irregularity introduced by, e.g., large convective cells. The importance of each of these two components will probably be very different for different objects.
A significant challenge for the late type variables is to define the regions of instability on the H-R diagram. If we want to have a chance to define the limiting parameters for the onset of pulsation and to estimate the influence of convective motion on the variability, a significantly larger sample of light and velocity curves as well as other stellar characteristics (both fundamental as well as atmospheric, for instance the occurrence of emission lines) are needed. In this context the light curves provided by amateur astronomers are of great value but they are not accurate enough to avoid ambiguities in small light curve irregularities. Therefore only photoelectric data as provided e.g. by automatic telescopes can reveal these details in the light curves. Both sources of data are needed to search for possible further components contributing to the light change of semiregular and irregular variables.
© European Southern Observatory (ESO) 2000
Online publication: September 5, 2000