An efficient method for dealing with line haze in stellar spectra

When analyzing any wavelength region, stellar spectroscopists must deal with the multitude of weak lines which depress the true continuum and thus form the pseudocontinuum. Calculating the effect of these lines, known collectively as the “line haze,” under conventional synthetic spectrum analysis is CPU intensive; an excessive amount of time is often spent calculating the opacity profiles of lines which are too weak to have even a cumulative effect on the line haze. We have developed a method of identifying and eliminating these “ultraweak” lines from synthetic spectrum computations. We treat the line opacity as a perturbation to the continuum opacity and generate tables of equivalent width coefficients which are a function of the model atmosphere, species, excitation potential, and wavelength region. Through interpolation of these tables and knowledge of the gf value for each line, we rapidly and accurately calculate the equivalent width of any line on the linear part of the curve of growth. We then determine the cumulative line blocking as a function of equivalent width. Through the use of this function and a specified line-blocking error, we eliminate from further consideration all lines having little or no effect on the pseudocontinuum. The reduction in the number of lines used in the spectrum synthesis results in a significant savings in computation time. We present the method for $314\,000$ atomic and molecular lines in the 864-878 nm wavelength region, effective temperatures of 4000-6000 K, $\log g$ of 1.5-4.5 (cgs units), and solar abundances.