Next: 8.6 Version history of Up: 8 The Pipeline and Previous: 8.4 The Error Propagation

8.5 Glitch removal effects in fast speed AOT 1 observations

8.5.1 Introduction

The purpose of the study reported here was to check if the glitch removal process included in Derive-SPD has an impact on spectral lines detected in fast speed AOT 1 observations. The spectra of three Planetary Nebulae with strong emission lines were selected for this test.

8.5.2 Processing

The following steps were followed:

1.: Check the AAR derived with the OLP version 5.2 to see if there are points belonging to the spectral line which are flagged as glitched.
2.: Derive a new SPD and AAR in which glitch detection is disabled by setting the parameter alpha in CAL06 to a very high value.
3.: Compare those lines in which some points are identified as glitched in the standard AAR, to the corresponding ones in the non-deglitched AAR.
4.: Flatfield/Rebin both AARs and fit a Gaussian to the lines of interest to calculate the line fluxes.

8.5.3 Results

The attached figures show several examples of the comparison between the standard AAR product (crosses) and the non-deglitched AAR (asterisks).

Figure 8.3: Good AOT 1 line fitting

Figure 8.4: Bad AOT 1 line fitting

Table 8.1 displays the results of the line flux calculation after a Gaussian fit, for all those lines in the three observations in which points flagged as glitched were found.

Source Line Line flux Line flux %

( ) standard AAR non-deglitched AAR discrepancy

(W cm ) (W cm )

1 24.31 4.70 5.31 11.5

25.88 7.79 8.78 11.3

2 15.55 4.81 6.11 21.3

3 10.51 3.01 5.48 45.1

15.55 3.88 1.10 64.7

25.88 3.38 4.77 29.1

Table 8.1: Comparison between different processing methods

**Table 8.1:** Comparison between different processing methods
Source	Line	Line flux	Line flux	%
	( )	standard AAR	non-deglitched AAR	discrepancy
		(W cm )	(W cm )
1	24.31	4.70	5.31	11.5
	25.88	7.79	8.78	11.3
2	15.55	4.81	6.11	21.3
3	10.51	3.01	5.48	45.1
	15.55	3.88	1.10	64.7
	25.88	3.38	4.77	29.1

The results of the analysis are:

1.: Most spectral lines are not affected by glitch removal.
2.: In few lines, some points at the wings of the line are flagged as glitched, causing the line profile in this region to deviate from the non-deglitched one.
3.: In few lines, many points belonging to the line are flagged as glitched. Glitch removal applied to these points introduces a substantial error in the line profile and in the line flux.
4.: For those cases described above, the non-deglitched line profiles can be fitted by a Gaussian function much better than the deglitched ones.

8.5.4 Conclusions

Generally the spectrum resulting from a fast speed AOT 1 observation is improved after glitch removal. However, when points belonging to spectral lines are flagged as glitched, one must be cautious about the line profiles and fluxes (except for band 4) because the line profile may be significantly distorted by glitch removal.

If a user wants to derive line fluxes from fast AOT 1's it will be necessary to check the datapoints around a line to see if they are flagged as glitched. If they are it may be necessary to visit one of the data centers so that additional data processing can occur.

Next: 8.6 Version history of Up: 8 The Pipeline and Previous: 8.4 The Error Propagation

K. Leech with contributions from
the SWS Instrument Dedicated Team (SIDT)
and the SWS Instrument Support Team (SIST)