Minutes of the Herschel spectral mapping telecon of 18th July 2011

On the call: Thangasamy Velusamy, David Teyssier, Kevin Croxal, Peter Imhof,
Chris Pearson, Ed Polehampton, Ros Hopwood, Ivan Valtchanov, Pat Morris,
Vianney Lebouteiller, Steve Lord, Elena Puga, Bruno MerĂ­n

2.- Observer feedback from the HIPE Forum 2011 meeting (BM, Habart,
Polehampton, Exter, McCoey)

- Current feedback on spectral mapping we have was reviewed by
looking at the presentation by Ed Polehampton, Katrina Exter and
Carolyn McCoey at the HIPE Forum meeting. The outcomes are:

- few publications yet on Herschel spectral mapping observations
- in general it was believed that analyzing spectral mapping
in HIPE is doable but still requires a good amount of
scripting and knowledge of hidden information in the software
which is not available to all users. Most Key Program observers
are reprocessing in HIPE and writing to FITS after that because
they found non-working applications. Work is in progress to
assess which of that functionality should stay in HIPE or be
kept in external tools. Users should expect big progresses
on this area in HIPE 8 and 9.

3.- Things that observers can currently do

- There are user pipelines in HIPE to produce fully calibrated
spectral cubes for all three instruments.
- In HIPE 8 there are also working tools to make some analysis
on them like e.g. the velocity maps, which works better than
IDL once the fits are restricted to the proper dynamical range
of the line.
- For HIFI there are two very nice tools to remove fringes
(so far n-times sinosoidal functions) and baselines that in
principle could work for all three instruments. That allows
then combining them after the defringing and baselining into
very clean and science ready cubes. Katrina and
Ivan will test the tools with PACS and SPIRE/FTS data,

4.- Things that observers cannot yet do

- On the on- and off- comparisons:
- For HIFI the off-position check scripts start to be well
documented and there are scripts available at the ICC
that can be run interactively to do that.
- For PACS, the ipipe in the HIPE 7 user release
has the possibility to extract separately the on and off
spectra for comparison.
- For SPIRE/FTS, it was not discussed whether this is
really needed or not.

- Pointing delays and zigzag in spectral maps
- For HIFI OnTheFly maps the effect is maybe more present
than in the other instruments but it has to do with a
low-level program in the pre-processing of the pointing
product which is already identified and is being fixed.
A hand-made correction was developed at the ICC which
removes the zigzag pattern by adding a 0.3 seconds delay
in the timelines which then translates to a 1.4 arcsec
delay in the maps. This is however a very user-intensive
method which is supposed to be solved by the correction
to the production of the pointing products for HIFI.
- For PACS, a similar effect but smaller has been reported on
very few cases and mostly in the edges of the spectral
IFU cubes.
- For SPIRE/FTS this is not an issue since the spectral
mapping with that instrument is not done moving the
spacecraft but rather with the beam steering mirror.

- Extracting point-source and extended source fluxes from an
- For SPIRE/FTS, it was reported than the absolute
continuum fluxes extracted from the FTS mapping data
as compared with the photometry measurements produce
good results on bright sources with the point-source
calibration (built in int he pointed observations
pipeline) and the low-flux regions are well calibrated
with the extended source calibration (default for
the spectral mapping pipeline). The SPIRE is working
on the long-term to improve the situation for intermediate
extended sources but that is yet not available in any
- For PACS, the same happens, but so far it has not
been a source of concern among the users. The spectrometer
beams have been just posted on the PACS twiki page
and should help users assessing whether their sources
are or not extended and convolving from one wavelength
to another to make line ratio maps, etc. Some key
programs have IDL codes to do convolution and comparison
of cubes from different instruments.
- For HIFI, it has neither been discussed as a major

- HIFI-specific problems:
- In order to use an off position to help in the definging
of a spectrum, the users need to make sure that the
off position is not contaminated with other sources. The user
needs to assess whether the current matching technique
worked. There are a couple of scripts at the HIFI ICC to
do that, particularly for the HEB band 7. Ronin distributed
one to Jorge Pineda for that. The script could be easily
adapted to also work for e.g. raster maps.
- There are two scripts at the HIFI ICC: one for correcting
from contaminated off spectra and another one for removing
the standing waves with a self-calibrating method.
- On the fringing issue, the fringes in HEB bands are from
electronic standing waves and therefore they are not sinusoidal
so the hifiFitFringe task will do a good job but not perfect,
since that task just uses sinusoidal functions so far.

- PACS specific problems:
- The lack of spectrometer PSFs has prevented some observers
to quantify the extent in the case of extended sources.
The PSFs are now in the PACS twiki page at the HSC. A complete
3D drizzling algorithm is available also at the ICC and is being
currently tested. If some observer would like to test it under
the condition that it is in development and therefore should be
used at the users' own risk, they should contact the helpdesk
to ask for it. Ideally, we would like to hear back from your
findings too.
- On unchopped spectral mapping, there is some pattern noise
in the cubes which seems to come from detector transients at
the timeline level. Some cubes have bright spots due to overlapping
raster positions. This has only be seen so far in unchopped
mode. The PACS ICC is still working in understanding these

5.- Questions, comments, doubts

- The deliveries of PACS beams or useful information on the instrument
twiki pages are extremely useful for observers but should come if
possible acompained with some README and basic information so that
observers can use them easily.

6.- Next telecon?

- A new telecon will be announced in September 2011 via the
relevant DP Interest mailing lists.