ar X iv : a st ro - p h / 02 02 38 4 v 1 2 1 Fe b 20 02

Resonance line-profile calculations based on hydrodynamical models of cataclysmic variable winds. ABSTRACT We describe a method of calculating synthetic line profiles using a generalized version of the Sobolev approximation. We apply this method to calculate line profiles predicted by the models of two-dimensional line-driven winds from luminous disks due to Proga, Stone & Drew. We describe the main properties of the model line profiles and compare them with recent HST observations of the cataclysmic variable IX Vel. The model wind consists of a dense, slow outflow that is bounded on the polar side by a high-velocity stream. We find that these two wind components produce distinct spectral features. The fast stream produces profiles which show features consistent with observations. These include the appearance of the classical P-Cygni shape for a range of inclinations, the location of the maximum depth of the absorption component at velocities less than the terminal velocity, and the transition from net absorption to net emission with increasing inclination. However the model profiles have too little absorption or emission equivalent width compared to observed profiles. This quantitative difference between our models and observations is not a surprise because the line-driven wind models predict a mass loss rate, mostly due to the fast stream, that is lower than the rate required by the observations. We note that the model profiles exhibit a double-humped structure near the line center which is not echoed in observations. We identify this structure with a non-negligible redshifted absorption which is formed in the slow component of the wind where the rotational velocity dominates over expansion velocity. We conclude that the next generation of disk wind models, developed for application to CVs, needs to yield stronger wind driving out to larger disk radii than do the present models. Subject headings: accretion disks – outflows – novae, cataclysmic variables – methods: numerical – 2 –