1: \begin{abstract}

2:  We discuss the current implementation of the ALI method into our

3: HYDrodynamical RAdiation (Hydra)

4: code for rapidly expanding, low density 

5: envelopes commonly found in core collapse and thermonuclear 

6: supernovae, novae and WR stars.

7:  Due to the low densities, non-thermal excitation by high 

8: energy photons (e.g. from radioactive decays) and the time 

9: dependence of the problem, significant departures from LTE are common 

10: throughout the envelope even at large optical depths. 

11:  

12:  ALI is instrumental for both the coupling of the statistical equations 

13:  and the hydrodynamical equations with the radiation transport (RT). 

14:  We employ several concepts to improve the stability, and convergence 

15: rate/ control including the {\sl concept of leading elements}, 

16: the use of net rates, level locking, reconstruction of global 

17: photon redistribution functions, equivalent-2-level approach, and 

18: predictive corrector methods. For appropriate conditions, 

19: the solution of the time-dependent rate equations can 

20:  be reduced to the time-independent problem plus the (analytic) 

21: solution of an ODE 

22:  For the 3-D problem, we solve the radiation transport via the moment equations. 

23: To construct the Eddington tensor elements, we use a Monte Carlo scheme to 

24: determine the deviation of the solution for the RT equation from the 

25: diffusion approximation (ALI of second kind). 

26:  

27:  At the example of a subluminous, thermonuclear supernova (SN99by), we show 

28: an analysis of the light curves, flux and polarization spectra and discuss 

29: the  limitations of our approach.

30: \end{abstract}

31: