Docstoc

ST - ECF

Document Sample
ST - ECF Powered By Docstoc
					HST Post Operational Archives
FOS “Final” Deliverables
Anastasia Alexov prepared for STScI visit August 2001 STPOA V1.2 “Final FOS” Release

Calibration Refurbishment of the Faint Object Spectrograph (FOS)
• FOS is the 1st trial instrument for the POA project • Comprehensive review of 24,000 data sets (cal + sci)
– Advantage of „global view‟ of entire instrument archive

• The solution of major known calibration problems:
– Wavelength Zero Points (STPOA V1.0 Aug 2000) – Dispersion Solutions (STPOA V1.1 May 2001) – Flatfields (post-COSTAR) (STPOA V1.1 May 2001)

– Flatfields (pre-COSTAR) (STPOA V1.2 Aug 2001)
– Darks (STPOA V1.2 Aug 2001)
Aug 2001, POA V1.2 Release A. Alexov, IPMG
2

STPOA V1.2 Aug 2001
• STPOA V1.2 has now been installed at STScI

• List of tools in STPOA.POA_FOS package:
poa_calfos - Post Operational Archive Pipeline calibration for FOS. poa_preproc_fos - Task for preprocessing FOS data, for poa_calfos.

poa_procfos_all - Task for running the poa_preproc_fos & poa_calfos.
pfos_pix2wav - Task to calculate l positions from FOS pixel positions.

Aug 2001, POA V1.2 Release

A. Alexov, IPMG

3

POA FOS Processing Criteria
• Solution available for the following FOS datasets:
– Detector: BLUE
– Grndmode: SPECTROSCOPY, RAPID-READOUT, IMAGE – FGWA_id: H13, H19, H27, H40, H57, L15, L65 – Aper_id: A-1, A-2, A-3, A-4, B-1, B-2, B-3, B-4

• POA data “all” processor (poa_profos_all) checks whether data fall within POA processing criteria:
– If yes, runs poa_preproc_fos to update header reference file information and then runs the pipeline poa_calfos – If no, executes calfos instead of running POA correction
Aug 2001, POA V1.2 Release A. Alexov, IPMG
4

Calibration Reference Files (CDBS)
• STPOA has updated & new calibration ref files
– All POA ref files have been made „standard‟, in order to be CDBS compliant
– Current software has the functionality of running on POA reference files located internal to the STPOA package, as well as those same files located in CDBS (param internal_ref=no) – Two sets of raw and processed FOS data have been delivered: A) with internal ref files, B) with CDBS ref files

• STScI can choose to update CDBS in order to match the POA FOS data ingest (external ref version)
Aug 2001, POA V1.2 Release A. Alexov, IPMG
5

Overview of POA FOS Corrections

FOS/BLUE Wavelength Zero Point Issue
• Improvements reduce the systematic deviation in the wavelength zero points from as much as 5 pix to < 1 pix

• Parameters entering poa_calfos correction:
– Orbital position and attitude (pointing) for geomagnetic model – Setup of detector (Ybase, HV, X/Ypitch) – Ambient temperature near Filter/Grating wheel

• Correction is limited if data are kept at original sampling

FOS Science Example using Correction
• Original calibration is offset by ~4 pixels (black curve)

• New calibration only offset by ~0.5 pixels (red curve)
• Improves science results by ~200 km/sec • Can improve the results even more with oversampling

FOS/BL Dispersion Relation - calfos
• Third order polynomial to describe each grating separately

• Incomplete, inaccurate list of lines (PtCr-Ne lamp)
• G130H never properly calibrated, scattered light

FOS/BL Dispersion Relation - poa_calfos
• Physical model of optical path aperture-diode array • All gratings described by same set of parameters

• Improved line list for Pt-Ne lamp & Cr data (NIST)

New Flatfields
• Pre-COSTAR flats follow the same processing as post-COSTAR flats, except with less automation • Diode/Cathode components
– No clear diode component could be extracted

– Therefore the existing method was not altered

• Shifted flats needed to be consistent with POA processed data.
– For most flats the spectrum is simply shifted (and rebinned)
Aug 2001, POA V1.2 Release A. Alexov, IPMG
11

Flats (continued)
• Some modes still “share” flats (as before)
– Effect on science data small whilst the “detective” work required is substantial.

• Port to UNIX and automation
– Existing VMS IDL procedures converted to work under UNIX and automated. This allows any further tweaks in processing up to the .c5h stage to be carried through to all the flats simply by executing one script. – Full automation could only be achieved with post-COSTAR flatfield generation; rest were done interactively
Aug 2001, POA V1.2 Release A. Alexov, IPMG
12

New POA Dark/Background Correction
• BLUE detector diode dark map remains the same

• New scaling - Dark calibration data covering FOS‟s lifetime was used to fit the expected flux as a function of L-shell, this new scaling replaces the previous geomagnetic longitude/latitude model • L-shell reflects the dark flux better than the previously used parameters, moreover it fits the calibration data better
• However, considerable scatter remains even when the dark data is fitted with this new model. This scatter can be reduced when the dependency upon (the true) SUN_ANGLE and SUN_ALT are introduced, but a general model has proven difficult.
Aug 2001, POA V1.2 Release A. Alexov, IPMG
13

Scientific Impact

POA Dark Correction (cont)
• Bug fix:
– While implementing the L-shell scaling we noticed an existing bug in the application of the dark correction to ACCUM‟d exposures
– The dark correction was being applied directly to the dataset which were ACCUM‟d – We fixed this by “ACCUMing” the dark correction appropriately so that when the dataset was deACCUM’d the dark correction was also deACCUM’d to produce the correct end result. – We then ensured that the darks output in the .c7h file were the deACCUM’d darks, appropriate to the deACCUM’d subexposures

Aug 2001, POA V1.2 Release

A. Alexov, IPMG

15

POA FOS/BLUE - STScI Open Options
• Remaining POA FOS/BLUE considerations:
– STScI should decide if they want STPOA to be CDBS based
– CDBS & data ingest decisions effect the ECF and CADC archives

• Will there be ONE FOS calibration pipeline?
– The task poa_procfos_all makes the decision to run the POA pipeline or the old calfos pipeline, depending on the input data; both pipelines are visible and available to the user; is this too confusing?

• Need more science verification & feedback from the users while calfos and poa_calfos can be run on the same data
– If pipelines are merged, there will be no way to make comparisons
– One cannot run calfos on data that fit the POA criteria
Aug 2001, POA V1.2 Release A. Alexov, IPMG
16

The Icing on the Cake - Oversampling
• Obvious complexity and controversy of oversampling any data (not to mention all analysis tools which may be affected) • However - we have a very useful POA FOS sub-pixel correction which cannot be used to its full potential now • We plan to expand the functionality of the pfos_pix2wav tool in order to incorporate the POA sub-pixel correction
– poa_calfos pipeline is run; POA correction is made as a rounded value of the calculated correction (sub-pixel correction is lost in the rounding) – The user then measures their „line list‟ for analysis of line positions

– pfos_pix2wav runs on this „line list‟; the lost sub-pixel correction can be added to line positions before wavelengths are calculated for these positions ==> Oversampling can be “achieved”, w/o the negative effects!
Aug 2001, POA V1.2 Release A. Alexov, IPMG
17

Science Verification
• Using Milky Way absorption in QSO data
– 50 FOS datasets all from one proposal
– Data are from end of FOS lifetime - large YBASE correction – Only H13 grating used in observations – 12 different QSO objects observed

– All 12 were extensively re-analyzed with PI using POA correction
– 5-10 individual absorption measurements made per exposure – Total of ~300 line measurements were used for POA analysis – All 12 QSO‟s showed consistent improvement over calfos results

– The data had previously not been understood; until now...

Aug 2001, POA V1.2 Release

A. Alexov, IPMG

18

MW Halo absorption in QSO spectrum using standard calfos
• Two sets of long exposures (black,red) show repeat error

• A non-physical dependency of velocity on wavelength
• Only one absorption at 1403 A seems to fit expectation

MW Halo absorption in QSO spectrum improved with poa_calfos
• • • • Result of POA GIMP correction in combination with physical dispersion model POA GIMP correction minimizes repeat error from 2 long exposures (black, red) All but one line match central value of expectation (21 cm line) The outlier is QSO Lyman alpha redshifted (= NEW science)

Science Verification (Cont)
• Using LMC data
– 100 FOS datasets
– Data came from the entire span of the FOS lifetime – All BLUE gratings were used in observations
• H13-H40 gratings (10 objects) [entire wavelength span] • H13, H19 gratings (20 objects) [partial wavelength span]

– All 30 were extensively re-analyzed using POA correction – 15-50 individual absorption measurements made per exposure; highly dependant on high/low nebular excitation – Total of ~450 line measurements were used for POA analysis

Aug 2001, POA V1.2 Release

A. Alexov, IPMG

21

Scientific Impact

QuickTime™ and a GIF decompressor are neede d to see this picture.

POA “Final” FOS Deliverables
• Improved HST science data archive
– 5977 FOS BLUE datasets POA re-calibrated (on DVD‟s, for re-ingest)

– On-the-Fly (OTF) re-calibration at ST-ECF and CADC HST archives

• Upgrade calibration pipeline (appropriate s/w environment)
– STPOA V1.2 contains poa_calfos pipeline and other tools (IRAF env.)

• Supporting reference files
– STPOA V1.2 contains internal standard reference files (upgrades & new) – POA re-calibration works on POA internal and POA external CDBS files – STScI may wish to decide whether or not to ingest POA reference files into CDBS
Aug 2001, POA V1.2 Release A. Alexov, IPMG
23

POA FOS Deliverables (Cont)
• Documentation
– All STPOA tools/tasks have standard help files
– POA website provides the most up to date documentation, technical information and software for the FOS – FOS chapter of the HST Handbook - updated and delivered in the fall

• Long term user support (world wide community)
– POA FOS User Support Helpdesk (ecf-poa@eso.org) - since Aug 2000

• Advertising/PR
– All FOS PI‟s have been notified of POA activities and final release – ST-ECF Newsletter #29 contains article on the POA FOS re-calibration – POA FOS News Release using ST-ECF PR (and OPO) on Sept 11, 2001
Aug 2001, POA V1.2 Release A. Alexov, IPMG
24

What’s next
• Software (and Data) Patches before the end of the year
– Can include user feedback

• Upgrade of pfos_pix2wav tool
– To compensate for lack of oversampling of the data, this tool will add the sub-pixel POA FOS correction to line lists in pixel space (before wevelengths are calculated for these positions)

• POA STIS (side 1 data)
– The head of the POA project, Michael Rosa, will be at STScI next week (Aug 28-30) to discuss POA FOS issues and prepare for POA STIS work

Aug 2001, POA V1.2 Release

A. Alexov, IPMG

25