Some remarks about nova spectroscopy and photometry

In this paper we explore the sensitivity of stellar atmosphere modeling of early nova spectra to the input parameters. Our results identify spectral regions where the continuum spectra are insensitive to the input parameters and certain wave bands where they are sensitive to the input parameter. Thus, we identify the best spectral region for determining elemental abundances, effective temperatures, and luminosities. We find in some cases that nova atmospheres have a continuum extinction coefficient that is dominated by scattering. The continuum scattering and the resulting continuum extinction are for these cases nearly gray. We apply and extend the small parameter Chandrasekhar expansion and we use the properties of the gray opacity, to study the effects of small residual absorption in the atmosphere. Two cases are treated: the case of a static extended atmosphere and the case of an extended atmosphere in motion. We solve, self-consistently, the transfer equation for these physical conditions. The assumption of radiative equilibrium allows us to also obtain a self-consistent temperature profile which validates our approximation a posteriori. Using these solutions, we (a) derive conditions for the required signal-to-noise ratio (or the relative accuracy of the observe fluxes) and the required observed frequency range needed to obtain a prescribed accuracy for those nova parameters obtained from a comparison of observed with synthetic spectra; (b) assess and establish the importance of spectroscopic (in contrast to photometric) measurements; (c) show that the use of synthetic spectra is an essential part, in the optically thick phase, of the determination of nova characteristics from the observations; and (d) define a new property continuum similarity and show how and when nova spectra possess this property. We find that the only wavelength region which is not affected by the physical conditions used to arrive at these conclusions is the ultraviolet. Therefore, we conclude that the ultraviolet is the most important wavelength region to use when deriving nova parameters from synthetic spectra.