Astron. Astrophys. 338, 1102-1108 (1998)

3. Observational data

3.1. CDS observations

The observations were taken of an active region at coordinates X530, Y-200 on the 9th July 1996. In Table 1 a summary of the different observations carried out by the two instruments is given. The relative starting position of the slit for each dataset may be seen in Fig. 1 and are discussed further in Sect. 4.1.

Fig. 1. An image of the Sun taken with EIT (courtesy of the EIT consortium) showing the position of the slit for datasets s3477r01 (top row), 165749 (second row), s3477r01 (third row) and 173109 (bottom row) on the 9th July 1996. This EIT image was taken in Fe XV 284Å at 17:09 UT. The axes are in units of arc sec from disk centre. Note the position of the limb at 800-900 arc sec.

Table 1. A log of the various density datasets observed with the CDS and SUMER instruments.

The CDS observations were taken using the 308-381Å region of the Normal Incidence Spectrometer (NIS). Within this wavelength range there are two potentially useful electron density sensitive line ratios involving lines of Fe XIII . These are Fe XIII 359.64/348.18 and Fe XIII 318.12/320.80. Despite there being a small amount of blending of the Fe XIII 359.64Å line with a weak Ne V line at 359.37Å, the Fe XIII 359/348 ratio was considered to be the more reliable density diagnostic. The Fe XIII 318/320 ratio was found to be very unreliable due to a significant blending problem of Fe XIII 320.80Å with Ni XVIII 320.55Å. It should be pointed out at this stage that the Fe XIII line observed at 359.7Å is in fact a blend of two Fe XIII lines, one at 359.64Å and a weaker one at 359.83Å. For the purposes of this work these blended lines shall be referred to only as the stronger line in the blend, i.e. Fe XIII 359.64Å.

In order to obtain a fast time sequence of images, the windowing capabilities of NIS was used, with a window of 40 wavelength pixels (3Å) and 71 spatial pixels (119 arc sec). Slit number 5 of CDS was used with a size of 4240 arc sec². The integration time was 10 seconds with a cycle time of 13.5 seconds between each set of images. A total of 200 images were obtained in 45 minutes. In the end it was decided to rebin the resulting files by a factor of two in time to improve the signal-to-noise at the expense of time resolution. This resulted in 100 images with a time resolution of 27 seconds. For these observations, the solar rotation compensation mode was switched off, and thus the image was generated by the rotation of the Sun.

Standard routines from within the CDS software tree were used in the data reduction, see O'Shea (1997) for details. An automated procedure was used to fit each of the images in the 100 time frames, with the integrated flux used to produce observational line ratios for the particular region of interest along the slit. Individual profiles were selected at random in order to check the accuracy of the automated line fitting.

3.2. SUMER observations

For our density diagnostic study the O IV 1407.39/1401.16 density sensitive line ratio was chosen. Actually, as pointed out in the paper by Wikstol et al. (1997), there are a number of reasons to suggest that density determination using the O IV 1399.7, 1401.2 and 1407.4Å lines may be the only method to infer electron densities in the transition region using SUMER data. In their paper, the above authors pointed out that due to factors such as blending and the weakness of certain lines, there are surprisingly few good line pairs to be found.

The SUMER datasets that were used in this analysis are shown in Table 1 and these will henceforth be referred to only as the last number in the filename shown. The two datasets 165749 and 173109 were observed using slit number 6 of size 0.3 120 arc sec². For both datasets a window of 50 wavelength pixels (2.2Å) and 120 spatial pixels (117 arc sec) was used. Each dataset consists of 200 images that were obtained in 33 minutes. Again the solar rotation compensation mode was switched off. Each image was taken with an exposure time of 8.95 seconds and a cycle time of 10 seconds. As with the CDS data, it was decided to bin in time by a factor of two to improve the signal-to-noise. Again, standard procedures within the SUMER software tree were used to do the data reduction (see O'Shea, 1997).

© European Southern Observatory (ESO) 1998

Online publication: September 17, 1998
helpdesk.link@springer.de