Astron. Astrophys. 317, 942-948 (1997)

3. Data reduction

We first selected the spectrograms exhibiting the best spatial resolution. These were first reduced in the usual fashion: The average dark signal was subtracted and a gain correction was derived from the flat field spectrograms and was applied. The latter eliminates possible effects from vignetting, varying width of the spectrograph slit along the spatial direction, and dust on the optical surfaces and on the chip. Next, since the two oppositely polarized light beams pass through the spectrograph along slightly different paths, correction for different optical aberrations had to be applied. For this the and subframes had to be adjusted to have the same orientation of the spatial and the wavelength axes and to have the same spatial and wavelength scale. This is necessary to avoid spurious V signals upon subtraction of the spectrograms. Noise reduction was achieved by low-pass Fourier filtering in wavelength direction and by taking running means in spatial direction over 3 CCD pixels, limiting the spatial resolution to 05.

The addition and subtraction of subframes yielded spectrograms of the Stokes I and V components, respectively, which were normalized to the continuum intensity . The polarimetric accuracy is of 0.5% of , essentially due to photon noise and to inaccuracies of the gain corrections. Fig. 1 depicts profiles from regions with strong V signals, i.e. from magnetic regions. The data for the 615.1 and 617.3 lines were taken simultaneously. Those of the 614.9 line stem from a different observing date, thus from a different scenery on the Sun. For comparison, the average I profiles obtained from the flat field spectrograms from disc centre are shown as well. Typical observed profiles of these lines are also shown in Kneer et al. (1996).

Fig. 1. Stokes I and V profiles observed with high spatial resolution in plages near disc centre, normalized to the local continuum intensity. The V zero crossings have been chosen as reference wavelengths. The data of the Fe I lines are from simultaneous observations. The dotted profiles are from disc centre of the quiet Sun.

From the I profiles we determined the following line parameters: continuum intensity ; line centre residual intensity ; velocity from the position of the line centre; and the full width at half line depression FWHD. The values of will be given relative to the running mean velocity taken over derived from the position of the minimum intensity . In this way small-scale, mainly granular velocities are averaged out and is measured relative to the background of the 5 min oscillations. The V profiles yield: the maximum amplitude = max() with and the unsigned blue and red extrema of the V profile, respectively; the amplitude asymmetry ; the wavelength separation of and ; and the velocity deduced from the position of the zero crossing of the V profile. Again, is measured relative to the average line centre position, i.e. its running mean over . , , , , and are obtained from parabolic fits about the extrema of the profiles, from a third order polynomial fit about the zero crossing of V. The Stokes V data are determined only at spatial positions where the V signal is sufficiently strong. We chose as a lower limit = 0.02 , which is well above noise and possible spurious, instrumental V signals.

© European Southern Observatory (ESO) 1997

Online publication: July 8, 1998
helpdesk.link@springer.de