Mass loss rate formulae are derived from observations or from suites of models. For theoretical models, the following have all been identified as factors greatly influencing the atmospheric structure and mass loss rates: Pulsation with piston amplitude scaling appropriately with stellar L; dust nucleation and growth, with radiation pressure and grain-gas interactions and appropriate dependence on temperature and density; non-grey opacity with at least 51 frequency samples; non-LTE and departures from radiative equilibrium in the compressed and expanding flows; and non-equilibrium processes affecting the composition (grain formation; molecular chemistry). No one set of models yet includes all the factors known to be important. In fact, it is very difficult to construct a model that can simultaneously include these factors and be useful for computing spectra. Therefore, although theoretical model grids are needed to separate the effects of M,L,R and/or $T_{mathrm{eff}}$ or Z on the mass loss rates, these models must be carefully checked against observations. Getting the right order of magnitude for the mass loss rate is only the first step in such a comparison, and is not sufficient to determine whether the mass loss formula is correct. However, there are observables that do test the validity of mass loss formulae as they depend directly on $dlog dot M/dlog L$, $dlog dot M/dlog R$, or $dlog dot M/dlog P$.
Download