AN ANALYSIS OF VISIBILITY-BASED CONTINUUM SUBTRACTION

Continuum emission is often a viewed as a contaminant in radio interfe rometric spectral-hue imaging. At best, it makes the analysis of the s pectral-line images difficult. At worst, it limits the dynamic range o f the images. In recent papers, van Langevelde & Cotton (1990) introdu ced visibility-based continuum subtraction technique to radio interfer ometry, and Cornwell et al. (1992) analysed its properties. The techni que is based on modelling the continuum by a first-order polynomial in the visibility spectrum. We present a quantitative analysis of the te chnique, both in terms of the residual continuum and the sensitivity t o noise. The analysis is for polynomial fits of arbitrary order. A les s complete analysis of a related, image-based, continuum subtraction t echnique is also presented. We show that higher order polynomials can be useful for data with large field size and bandwidth, and that stabi lity in the presence of noise is not a problem, Some errors in the pap ers cited above are corrected.