>REPORT ON THE MEETING OF THE OPTICON 3D SPECTROSCOPY WORKING GROUP

REPORT ON THE MEETING OF THE OPTICON 3D SPECTROSCOPY WORKING GROUP


Held: 19-20 February 2001. Astrophysikalisches Institut Potsdam

In attendance:
 Petra Boehm             pboehm AT aip.de
 Thomas Becker           tbecker AT aip.de
 Veronique Cayatte       veronique.cayatte AT obspm.fr 
 Yannick Copin           copin AT strw.leidenuniv.nl 
 Begona Garcia           bgarcia AT ing.iac.es
 Bianca Garilli          bianca AT ifctr.mi.cnr.it 
 Rachel Johnson          raj AT ast.cam.ac.uk
 Andreas Kelz            akelz AT aip.de
 Markus Kissler-Patig    mkissler AT eso.org 
 Martin Roth             mmroth AT aip.de
 Juergen Schmoll         jurgen.schmoll AT durham.ac.uk 
 Lowell Tacconi-Garman   lowell AT mpe-garching.mpg.de 
 Jeremy Walsh            jwalsh AT eso.org (Chair) 

Jeremy Walsh opened the meeting. The WG has been expanded to include a new node - the Institute of Astronomy Cambridge. Ian Parry and Rachel Johnson are the two representatives and Rachel was at the meeting. Unfortunately Pierre Ferruit had trouble with travel arrangements and could not attend at short notice.

Klaus Strassmeier, on behalf of the AIP, welcomed the group. He mentioned that the AIP already has an active RTN (Star Formation) and other applications have been made, so there is plenty of local experience.

Euro3D proposal

Martin Roth opened the section of the meeting devoted to Euro3D. A lot of enthusiasm had been generated at the kick-off meeting in Garching and there was clearly a solid basis for a network on 3D Spectroscopy. Already as a result of the first meeting there had been a fruitful collaboration between Yannick and Martin on detection and extraction of a spectrum of a planetary nebulae in the core of M32 from Sauron data taken for another purpose.

The latest version of the proposal had been sent and Martin distributed copies of the Guide for Proposers and the Guide for Evaluators. He stressed that it is very important to obey the formal rules for a successful proposal. The meeting at ESO with Claus Madsen, the ESO representative at the EU, was reported. He had got feedback from the EU about the failed Euro3D proposal. The science was sound but there had been no improved access to infrastructures demonstrated. No RTD has been granted in astronomy. The advice to resubmit as an RTN should be followed. It should be made clear on the proposal how the evolution to an RTN and coordination with Opticon had happened. Claus also suggested to contact the EU before and to give them a copy of the proposal for comments. This can help in checking the formal requirements. The success rate for RTN's is around 15%. The other message he brought is that the EU loves Opticon so showing that we have Opticon blessing (such as a resolution at the OPTICON Board meeting in April encouraging the 3D Spectroscopy RTN) is a good thing!

Martin then outlined the scope of an RTN. The major objective is TRAINING OF YOUNG RESEARCHERS. Maximum age is 35 years. Maximum possible funding is 1.5MEuros with max 200KEuro per partner and no more than 40% in one country. Maximum period is 4 years. 60% of funds must be for personpower. Researchers must be hired from another country and should not have previously spent >12 months in the host country. There was no perceived advantage in terms of the EU in favouring pre-docs over post docs or vice versa.

Aspects of the proposal were discussed such as:
need to have a good workplan with achievable goals;
emphasize a connection with industry, so that something should be presented at an industrial fair once in the RTN;
hardware provision is almost excluded from an RTN;
important to explain intention of project in terms of funding, travel exchange and meetings.

From each partner, the requirements are:
1. Common understanding of the Euro3D goals
2. Commitment of own resources to the Network - the training resources from the RTN should be matched by the number trained
3. Description of team
4. Commitment to work plan
5. Collective training
6. Description of complementary work
7. The official cost of a postdoc and a Ph.D. student at the host institute must be provided. Having accurate figures will then allow a careful assessment of how many researchers we can afford for the RTN.

A vote was taken and it was unanimously decided to go for an RTN.

Martin outlined the evaluation criteria for an RTN (see Appendix E of the Guide for Evaluators). It was not clear how important was the emphasis on connections to other disciplines and Martin was to ask Claus about this aspect. One important aspect is that whilst we must emphasize the complementary expertise of the different teams the proposal must not be interpreted as a collection of teams with different plans. Common goals that could not be attained by individual groups are vital. Since we will be at least 8 participating groups, good coordination is required. A mid-term review of the RTN has to be conducted in order for the EU to monitor the progress of the contract. An internet site for common documents will be set up. Bianca made the point that some of the groups are already involved with large projects so that there is demonstrated organizational expertise.

In the training area, aspects to consider are the career prospects for young researchers and getting women as team leaders.

Martin outlined the procedure that takes place after the RTN is accepted by the EU. The project description becomes an official contract with the EU. Obviously this may require some changes which must be negotiated. There would be a workshop for the coordinator to attend. The Network coordinator liases with the EU and submits documents and reports and oversees the financial contributions according to the contract. There must be an agreement between the node coordinators and the Network coordinator.

Further points:
All teams to submit a figure for the RTN summarizing their science and instrument experience.
Training through the science was seen as a strong point.
On training Yannick was worried since Oasis on CFHT had not widely used on account of the preceived difficulties of the data. Even after a hands-on workshop.
There was discussion about students from other countries. In Italy foreign students are required to pass an exam in written Italian. For other countries (D and DK) this may not be a strict requirement.
Filling out of the B form Participant Profile/Information should be done soon as it requires entry of information from the administration of the participating institute.

A vote was taken on the period of the RTN and 3 YEARS was decided.

Martin emphasized that it was very important to keep to deadlines. If a group could not provide the required information on time it would have to be dropped since it would jeopardize the whole submission. The aim is to have a proposal almost ready by 20 April for comment by the EU, while still allowing enough time for last minute changes.

3D Data Format

Markus reported on the presentation of 3D Data Format plans to the ESO Data Interface Control Board (DICB). This Board aims to homogenize all ESO formats, headers, keywords. Markus attended a DICB meeting and stated that the WG intended to present a 3D Data Format. It was acknowledged that individual instruments have their own pipeline to remove the instrumental signature but that they should then produce a common output format. Scripts would need to be available to convert from other format to the 3D data format. DICB agreed to participate in discussions about a 3D format and would be willing to agree to a proposed standard. DICB showed its willingness to comment on the format.

It was stated that the IRAF project are looking at 3D formats and it was agreed that we should keep in contact.

Martin showed data from SAURON of the core of M32 which he had translated into the PMAS format consisting of a FITS file with associated tables. In this case the Sauron microlenses are square so translation to and from a cube was simple. The problem of dealing with microlens arrays which are not square was highlighted. Care must be taken that we do not settle on a format that looses data in the exchange. For example with circular microlenses and image(s) giving the spectra for each lens this cannot be uniquely mapped onto a 3D FITS image. Ancilliary information on the sky coverage of each individual spectrum must be carried in order not to loose (spatial) information. Yannick emphasized that the format should be lossless. If interpolation is needed to go from one format to another then there is some loss. Yannick wondered if MOS spectra could also be accommodated on the same 3D format to make it more general.

Tiger Format

Yannick Copin
Tiger consists of circular lenses arranged in a hexagonal pattern. The spectra are truncated on the detector. The data format is a collection of spectra in a platform independent proprietary binary format (.tig). Noise information is also held for each spectrum. The spectra are identified by the lens number. A FITS table links the lens number to the lens related quantities. For example the wavelength calibration is stored by lens number.
Advantages
Two files so different Tiger frames can share same table through the reduction process
Table files are not duplicated
Optimal storage: all stored pixels are meaningful (e.g. 1600 lenses with 455 pixels for each spectrum)
Good support for tables (FITS) and good libraries available
Disadvantages
Two files - Tiger frame plus assocaited table
Spatial distribution disregarded - have to build to get sky distribution.

Integral Format

Begona Garcia
Frames stored as 2D FITS images with each spectrum as 1 pixel in the spatial direction. Called ms format (but not to be confused with IRAF .ms multispec format). Since the Integral grid is not regular on the sky (and has offset sky sub-arrays) the correspondence with the sky position is given by an ASCII file which lists the X,Y position (offset) of each fibre.
An IDL routine for quick analysis is available which interpolates the spectra onto an alpha-delta image for a given wavelength range of the spectrum. There is no software to combine slightly shifted image sets.

VIMOS Format

Bianca Garilli
The internal format is a packed spectrum format consisting of one FITS file with a CD matrix specifying the pixel to wavelength transformation. Each row has a single spectrum uniformely sampled and with the same spectral extent. The FITS file table extension gives the fibre ID of each spectrum. A separate FITS table gives the fibre position on the sky allowing translation of the spectrum of a fibre to the sky plane. The final output format is FITS 3D with RA, Dec and Lambda axes. The microlenses of VIMOS are square so the mapping from fibres to sky pixels is direct.
Advantages
The internal format is compact and is the same as for the MOS mode of VIMOS.
For the 3D cube this is standard FITS and the spatial pixel position is reconstructed.
Disadvantages
The internal format gives spectra uncorrelated with position on sky and is not a common format, but was never meant to be.
The 3D cube works well only for data on a regular grid. It may not be applicable to all instruments especially those with rectangular (or other) shaped sky elements.

SPIFFI Format

Lowell Tacconi-Garman
SPIFFI uses 3D cubes. These are standard FITS with NAXIS1=RA, NAXIS2=Dec and NAXIS3=lambda. In future data quality and error will be included.
Advantages
The format is well defined, and there is much existing software which can readily be used.
Disadvantages
It only works on data that is regularly sampled in the sky plane.

TEIFU, SMIRFS and GMOS Format

Juergen Schmoll
TEIFU has data cubes 59x110x981 (X,Y,lambda) with hexagonal lenslets. The spectra are also curved and so are interpolated (in X, Y and lambda) at a very early stage. The data are stored as (real) 3D FITS cubes.
SMIRFS has 72 hexagonal lenslets and each lambda plane is bidimensional interpolated onto a uniform pixel grid. Output is FITS 3D cubes.
GMOS-IFU delivers 2D image of many single spectra with an associated table, since the single spectra can be easily distinguished on the detector.

Giraffe Format

Veronique Cayatte
The internal format is 2D FITS images with a separate table to specify the sky positions of the fibre spectra. There is also a mask image for flagged pixels. The output is a FITS 3D cube and this is the format for analysis (the Ancilliary Data Analysis System).
Advantages
Compatibility with other packages e.g. AIPS.
Possibility to add error and bad pixel masks as FITS extensions.

PMAS Format

Martin Roth
This is a stacked 2D FITS image. A table to give the sky position of each spectrum will be added using a FITS table extension

CIRPASS Format

Rachel Johnson
CIRPASS is a 499 element IFU with hexagonal lenslets for Gemini North. The lenslet array can have a scale range of 0.36, 0.25, 0.12, 0.05'' per lenslet. The instrument is aimed at 0.85 to 1.8 micron and has a spectral resolution of 3000 at 1.35 micron. The detector is a HAWAII 2kx2k.
The raw data consists of a Multi-Extension FITS (MEF) file where each extension holds one read of the detector (non-destructive reads can be made). The output files are also Multi-Extension FITS (MEF) files with stacked extracted spectra, variance and data quality. The information about the relative positions of each lens will be stored in a FITS table extension to the final data.

There was little time for discussion and no agreement was reached on a common format to adopt. It was suggested that Jeremy compose a voting form and distribute it.

3D Opticon Board meeting

There is a meeting of OPTICON on 5 April and there was discussion whether the RTN proposal could be presented. Probably it will not be ready. However at least a draft could be sent. It was not clear whether someone from the Working Group will be invited. The aim is to get a line in the OPTICON meeting minutes blessing the RTN proposal.

ESO Future Instrumentation Conference

Jeremy will aim to get a 5minute slot in the programme to advertise the group and say what we are doing.

Web pages

Petra Boehm reported on the Web page. Markus commented that Frames is not always accepted as a standard. Petra welcomes input for things to include. When the one page presentation of science + instrument for each group, requested for the RTN proposal, is ready they could be added to the Web pages.
Yannick suggested that we try to get a standard keyword in the Journal indexes of keywords which specifically refers to 3D spectroscopy.

NEXT MEETING:

The next meeting of the OPTICON 3D Spectroscopy WG will be in Lyon on 25-26 June.
Last updated 07 March 2001