ON THE ESTIMATION OF INTENSITY FOR LOW S N RATIO NARROW EMISSION-LINES/

The aim of this paper is to investigate the errors involved by the int ensity measure of low signal-to-noise ratio narrow emission lines. We also studied the errors associated with line intensity ratios. The spe ctral data acquisition was simulated by a Monte-Carlo program. These d ata were tested for the presence of a line and the lines detected were measured by a chi2 fitting program. This procedure generated a sample of intensities for a given signal-to-noise ratio. From these samples, we have established that the ''measured'' intensities are strongly bi ased towards overestimated values and their distribution is well repre sented by a log-normal distribution. If this phenomenon is not taken i nto account, erroneous astrophysical conclusions may be derived. In ad dition, we determined that a line ratio follows also a log-normal dist ribution. Finally we provide some charts able to give the error bars a ttached to a line intensity measurement and also to a line ratio. Thes e charts are given in terms of signal-to-noise ratio, and can be easil y converted into line intensities if the energy of the noise is known.