Imaging systems based on a narrow-band tunable filter are used to obtain Doppler velocity maps of solar features. These velocity maps are created by taking the difference between the blue- and red-wing intensity images of a chosen spectral line. This
method has the inherent assumption that these two images are obtained under identical conditions. With the dynamical nature of the solar features as well as the Earths atmosphere, systematic errors can be introduced in such measurements. In this paper, a quantitative estimate of the errors introduced due to variable seeing conditions for ground-based observations is simulated and compared with real observational data for identifying their reliability. It is shown, under such conditions, that there is a strong cross-talk from the total intensity to the velocity estimates. These spurious velocities are larger in magnitude for the umbral regions compared to the penumbra or quiet-sun regions surrounding the sunspots. The variable seeing can induce spurious velocities up to about 1 km/s It is also shown that adaptive optics, in general, helps in minimising this effect.
In a previous paper we described a method of estimating the single-measurement bias to be expected in astrometric observations of targets in crowded fields with the future Space Interferometry Mission (SIM). That study was based on a simplified model
of the instrument and the measurement process involving a single-pixel focal plane detector, an idealized spectrometer, and continuous sampling of the fringes during the delay scanning. In this paper we elaborate on this ``instrument model to include the following additional complications: spectral dispersion of the light with a thin prism, which turns the instrument camera into an objective prism spectrograph; a multiple-pixel detector in the camera focal plane; and, binning of the fringe signal during scanning of the delay. The results obtained with this improved model differ in small but systematic ways from those obtained with the earlier simplified model. We conclude that it is the pixellation of the dispersed fringes on the focal plane detector which is responsible for the differences. The improved instrument model described here suggests additional ways of reducing certain kinds of confusion, and provides a better basis for the evaluation of instrumental effects in the future.
