Minutes of the ING Wide Field Survey Workshop ============================================== Place: IoA, Old Observatory Meeting Room, Cambridge, UK Date: 28+29 Oct 1999 Version Jan 2000 Prepared by Nic Walton List of attendees: Guy Rixon GTR (ATC) 28/10 only Jon Davies JD (Cardiff) Judy Hynes JH (Cardiff) Alex Kambas AK (Cardiff) Simon Hodgkin STH (CASU) Jim Lewis JRL (CASU) Paul Groot PG (CFA) Nigel Hambly NH (IfA, Edinburgh) Nic Walton NAW (ING) Aprajita Verma AV (Imperial) Peter Bunclark PSB (IoA) Paul Hewett PH (IoA) 28/10 only Derek Jones DHPJ (IoA) Richard McMahon RGM (IoA) Chris Sabbey CS (IoA) Richard Sharp RS (IoA) Steve Smartt SJS (IoA) Gavin Dalton GD (Oxford) 28/10 only Jane ?? J? (Oxford) 28/10 only Alan Penny AP (RAL) Nick Cross NC (StAnd) Simon Driver SD (StAnd) Keith Horne KH (StANd) 28/10 only ?? Street ?S (StAnd) Nial Tanvir NT (Herts) 29/10 only Thursday 28 October 1999 ------------------------ [0. A combined list of action points arising from the meeting is given in Appendix A at the end of these minutes. ] SESSION 1: STATUS OF THE WFS PROJECT 1. RGM opened the meeting with a welcome. He noted that the first day would primarily concentrate on technical and organisation issues. The second day would focus on science exploitation. 2. Nic Walton (ING) presented a brief overview of the Wide Field Camera (WFC) on the Isaac Newton Telescope (INT). (Overheads available as Appendix B.1) 2.1 Key points were the improvements to the new Data Acquisition System (UltraDAS) commissioned in Aug/Sep 1999. The WFC array is now read out in under 60s (down from 180s), with a further upgrade due Q1 2000 to reduce this further to ~30s. 2.2 NAW noted that a coordinated 'Half-Arcsec' programme was being started at the INT. The aim was to bring the INT image quality to the inherent quality of the ING site, e.g. ~0.7 arcsec (Wilson et al, MNRAS, 309, 379 (1999)). Elements of the INT HAP programme include characterisation of the INT seeing, improved thermal isolation of the INT dome, etc. The 0.33 arcsec/pix scale at the INT/WFC is well matched to the desired 0.7 arcsec image quality. 2.3 KH queried the stability of the detector linearities. MJI noted that these changed when the controllers were changed to use San Diego Generation II ones with the UltraDAS in Aug 1999. The new linearity corrections are available on the survey WWW pages. 3. The status of the Wide Angle Survey (WAS) was presented by Richard McMahon (IoA) (Appendix B.2 for view-graphs). 3.1 The time allocation since the first runs in Sept 1998 was noted. The complexity of the project management was highlighted. 3.2 The filter choice was noted. Early observation were carried out using Harris BVRI filters due to the intermittent and late arrival of the new sloan-gunn g', r' and i' filters. 3.3 RGM noted that there is no evidence for fringing with the sloan-gunn r' filter and the i' fringing is reduced and more stable than was the case with the Harris I. The RGO U filter cuts off below the 4000A break. The WFS Z filter is the gunn Z which uses a RG850 glass, slightly different to the RG830 glass used in the sloan-gunn z' filter. SD noted that Jon Davies would shift to using the sloan-gunn filters for the Virgo work in 2000A. 3.4 RGM noted that the WAS has mapped some publically released Sloan Digital Sky Survey (SDSS) fields. Thus it will be possible to map the WFS onto the SDSS. The WFS is observing for ~10 times longer than the SDSS (in the Northern Hemisphere) and thus going ~1.3 mags deeper than the SDSS. Further, the image quality of the WFS data is often better than that from the SDSS (e.g. median july WFS data had a FWHM of 1.0 arcsec c.f. 1.4 arcsec reported in the publically released SDSS sampler set). 3.5 The grid size for the survey fields is now standardised on 30 arcmin in RA and 20 arcmin in Dec. Thus some 6 exposures are needed per sq degree. 3.6 To mid Oct 1999, some ~70 deg^2 had been observed in various filters. This includes a major survey of the South Galactic Cap (10 deg^2 in u'g'r'i'z', 8 deg^2 in g'r'i'z' and 15 deg^2 in g only) 3.7 For Semester 2000A emphasis will be placed on the 0 Dec strip in the RA range 10-14h, where Gemini and 2dF have fields. 3.8 RGM noted that the web pages giving details of the survey coverage were being updated (e.g. /home/rgs/public_html/wfsj1610.html) 3.9 There was some discussion about review time scales. NAW noted that a further call for proposals would be issued by the end of January 2000 with a deadline of 31 March 2000. This would be for programmes starting in Semester 2000B (Aug 1 2000 on). The form of the review panel was under discussion. 4. Gavin Dalton (Oxford) presented a status report on the deep UBVRI survey (see Appendix B.3 for overheads). 4.1 An overview of the 16h+40 (3x3 deg) deep field was shown. This was observed with the WFC using the rotator at 0 and 180 deg to maximise the efficiency of field coverage and give overlapped contiguous areas 4.2 From the stacked images the following (5sigma) detection limits were reached: Band KPB V R i Z Time(s) /exp 900 1000 1000 1200 600 No of exps 3 3 3 3 1 Det limit 26 25.5 25.25 24.5 22.0 The Z images were all obtained in twilight 4.3 The BVR stack of one field gives an image quality of ~0.9 arcsec on most sub-fields 4.4 GD is making a careful study of the systematics in the data sets. He showed that comparing the i'-Z/V-i' and I-Z and V-I color-color diagrams, that the i' data is more sensitive in selecting quasars in the range 4 6 with a density of 0.05/arcmin^2. 5. The observational and data pipelining aspects of the Faint Sky Variable Survey (FSVS) were presented by Paul Groot (Amsterdam/ CfA) (Appendix B.4). 5.1 The observing strategy seeks to survey 50 deg^2. To date (10/1999) some 9 deg^2 has been observed. The May 1999 run included two fields to be observed with the Chandra X-ray satellite. 5.2 The variability survey is being carried out in the Harris V filter. The survey is sensitive to variability on times scales of ~10 mins to a week. All fields are observed within 30 degrees of zenith in order to reduce atmospheric extinction effects. B and I images for colour information have been obtained for all fields. Additionally the WAS has acquired z' data. 5.3 The fields will be re-observed after one year to search for long-term variability. 5.4 The FSVS reaches limiting magnitudes of 24.5 in B and V, and 22.5 in I for a S/N of 5. 5.5 PG described the analysis procedures employed. The data sets are processed independently by collaborators in Tucson and Amsterdam with Iraf and Sextractor. Light curves and colour information for all objects are generated. It was noted that the FSVS pipeline generating object catalogues and light curves is run in real time at the telescope. Two papers describing the survey and initial results are in their final stages. Preliminary results were presented at the ING Instrumentation conference in Sheffield in April 1999. 5.6 The selection of sub-samples for spectroscopic followup is progressing. First observations of WHT+WYFFOS spectra of the brighter variables have been applied for. PG stressed that the scientific exploitation of the data was now in progress. 5.7 A discussion followed on the limits to the precision of the variability that you can detect. PG stated that variabilities at the milli-mag level are being obtained for V=18 mag stars. PG noted that a stellar classifier has been developed by the FSVS team. MJI suggested that this should be compared to the Sextractor based classifier employed by the EIS (ESO Imaging Survey), and also to the classifier that he had developed. 6. A large programme utilising the WFC in bright time in semester 99B is that of Horne and Penny aimed at detecting planets by searching for transient events in open clusters. This programme is not formally part of the WFS. Alan Penny (RAL) presented an update on progress with this programme, high-lighting areas of interest to the WFS (Appendix B.5). 6.1 The observation strategy involves repeated R band exposures of a small number of open clusters. 6.2 The accuracy of the differential photometry is crucial. Milli-mag precision is required (as being obtained in the FSVS). [... missing areas here? ...] SESSION 2: PIPELINE PROCESSING, DATA PRODUCTS AND THE ARCHIVE 7. The development of the WFS data pipeline was presented by Jim Lewis (IoA) (Appendix B.6). A fuller description is available in the paper submitted at the ING Sheffield Conference (Irwin & Lewis, 2000). 7.1 The design philosophy is to enable a reduction pipeline having the minimum user intervention. Since August 1999 the WFC data has been written in Multi-Extension Fits format (MEF). 7.2 JRL noted that the generation of flat fields is from twilight flats, but in principle super sky flats from dark science data could be generated. SD commented that these would be preferable when searching for low surface brightness science objects. To generate sky flats for the six main survey filters, the two hours of morning and evening twilight is utilised. 7.3 Detail was provided on a number of the pipeline tasks: 7.3.1 A full 2-D bias subtraction is performed, the master bias generated from bias frames obtained at the start and end of a run. 7.3.2 The non-linearity of the individual CCDs has been characterised. For the post Aug 1999 data, CCD#2 and CCD#4 were found to be non linear at the <1% level, fit with a quadratic term. CCD#3 was non linear at the ~3% level but could be fit with a simple linear term. CCD#1 required a higher order term to fit it's non-linearity. The data is correct for non-linearity by means of a Look Up Table (LUT). GTR suggested that the WFC preflash LEDs be used for routine linearity checks. 7.3.3 A first order World Coordinate System (WCS) is generated for each frame by reference to the know geometry of the WFC and the pointing information from the telescope control system in the FITS headers. This is accurate at the 5-10 arcsec level. 7.3.4 Bad pixels are corrected by interpolation with reference to a Bad Pixel Mask - noted in the FIST headers. 7.3.5 The sky levels are adjusted so that all CCDs have the same sky level, this correcting for the individual CCD gains. 7.3.6 Image detection and an accurate WCS generation are the latter stages of the pipeline. 7.4 Peter Bunclark gave a brief mini presentation on issues surrounding the choice of WCS 7.4.1 There is currently no agreed FITS standard with which to define the WCS, see Calabretta and Greisen (A&A, 1999) for a discussion. 7.4.2 Partly because of this there is only a limited availability of software capable of correctly handling the coordinates of large field data. The Oct 1999 ADASS meeting again aimed at a WCS standard, however there remain differences of opinion. For instance, the STScI would wish for the generalisation where RA and Dec is described by a 39 order polynomial. GD noted that GMOS on Gemini will require coordinates with better than 0.1 arcsec accuracy over a 6 arcmin field. 7.4.3 PSB noted that the WCS used in the WFS was based on a cubic term to describe the WFC geometry. This was supported by the IRAF software and was a 'de-facto' standard. 7.5 SD noted that studies of the goodness of flat fielding using various techniques should be investigated. Likewise cross checks between WFS groups should be attempted. This was required to increase confidence in the use of the standard WFS pipeline flat fielding based on twilight flats. KH suggested that the pipeline be adapted to generate PSF variations across the CCDs. 8. The Photometric calibration of the WFS data was addressed by Derek Jones (IoA) (Appendix B.7) 8.1 The transformations between the UBVRI system of Johnson et al, the SDSS system and that of the WFS was discussed. 8.2 Insufficient standards have been observed by the WFS at high airmass to date. Hence the extinction coefficients are taken from the JKT. 8.3 The colour equation for g' is determined as -0.139+/-0.005. For the other filters (u', r', r', z') the colour equations equal ~zero, thus the WFS filters (except g') are on the same system as the SDSS. 8.4 Further blue and red flux standards are required to improve the calibrations. 9. Additional pipeline processing topics were discussed by Mike Irwin (IoA) (see Appendix B.6). 9.1 MJI mentioned that the Carlsberg meridian telescope (CMT) was now operating with a CCD camera in drift scan mode. It is equipped with a sloan-gunn r' filter, and thus provides extinction measures for every WFS survey night. The CMT data will shortly be used to assess the photometric quality of each WFS night. 9.2 De-fringing now affects only the i' and z band data sets. The scale of fringing varies for each chip. Super-sky flats are generated, with stars and objects eliminated to form the fringe map. 9.3 r' band fringing is <0.1% (c.f Harris R ~0.5% of sky). The i' band fringe is stable night to night and through a night. The z band data shows variations varies over a night (caused by sky line intensities fluctuations through the night) 9.4 The fringe map is scaled (with reference to the background) and subtracted from the flat fielded data. The residual fringing is corrected to the 0.2% level for the i' data. The z band data fringe maps vary through the night. From a principle component analysis of multiple z band fringe maps, the components to the fringing can be determined. This technique has enabled reductions in the z band fringing from 6% to 0.5% to be made. 9.5 The de-fringing process is now automated. The sky flats need to be well exposed to ensure that the fringing is <0.5% in these flats. 9.6 MJI discussed the generation of the image catalogues from the WFS data sets. 9.6.1 The current software measures the flux in the detected sources in various ways. A range of core fluxes enclosing different percentages of the flux are given. 9.6.2 A stellar discriminatator has been developed that utilises the curve of growth of the enclosed object flux. Referring to 7.5, MJI noted that it was difficult to determine the PSF for each frame directly. However, with the aid of the stellar discriminator, it was possible to recover the stars in each frame and thus re-form the PSF for the frame. 9.7 Catalogues are available for the WFS data now from the WFS www pages. JD requested that the WFS software be tweaked to enable the optimum detection of low surface brightness galaxies in the data. 9.8 The catalogue data is being employed for quality control purposes. Sky brightnesses and image quality (seeing) is tracked for each exposure. Data sheets summarising these are available on-line for each WFS night. It is envisaged that this information will be incorporated into the sybase data base to enable searches to be made based on a limiting sky brightness. Each data frame will be tagged with a quality indicator (grading sky brightness, photometric or not, and image quality). 9.9 The astrometric fit currently relies on the PMM (2nd edition) and APM catalogues. 10. Data policy and access to the archive was presented by Jim Lewis (IoA) (Appendix B.8) 10.1 The data are copied to DDS3 DAT tapes (12GB native capacity) and sent to the IoA at the end of each WFS observing block. 10.2 The raw data is processed, with the reduced data being held on-line and the raw data deleted from disk. 10.3 The reduced data is available through the WFS query www page to the UK and NL communities (and to the rest of the world after the one year proprietary period). To date there has been little demand for the raw data. This is however available through the normal ING data archive if required. 10.4 In the near future an improved www interface to the data is foreseen. Increased capability to handle sub-queries will be provided. 10.5 The meeting agreed that the end user required ease of access to the data, for which the astrometric and photometric integrity was assured. 11. Additional information on the development of the improved user interface to the WFS data archive was presented by Simon Hodgkin (IoA). 11.1 The main portal to the WFS is at http:// www.ing.iac.es/WFS. Links from here lead to the data download area, log files, status of WFS fields. etc. 11.2 In the near future (by Q2 2000) these pages will include: - on-line source catalogues (including photometry, astrometry, shape) - quality control information for each image (seeing, sky levels, etc) - cross links to other archives' data - the ability to include cross-referencing to user supplied catalogues - improved www navigation capabilities - improved on-line documentation 11.3 A discussion ensued as to desirable areas to be addressed. It was though that weather and data from the Carlsberg Telescope (nightly gray term etc) should be associated to the WFS data. GD and NAW suggested that postage stamps of the data be available for download The first day of the meeting closed at 17.15. Friday 29 October 1999 ---------------------- SESSION 3: SCIENTIFIC RESULTS AND EXPLOITATION OF THE WFS 12. First results from the Pleiades were presented by Simon Hodgkin (IoA) (Appendix B.9). 12.1 This talk addressed the rationale of the project to determine the mass function in the Pleiades. To date 1.7 deg^2 has been observed complete to I=21.7 12.2 87 A grade brown dwarf candidates found, the faintest being at I=22.5 ( < 30 M(Jupiter)) Followup spectroscopy programmes are underway utilising UKIRT. 12.3 Conclusions to date are that the mass function is rising to 0.04 M(solar) (alpha=-0.8). This implies that there are ~500 brown dwarfs to a mass of 0.01 M(solar). The mass of this population is 20 M(solar) or 2.5% of the mass of the Pleiades. 13. The use of the WFS to search for intermediate red-shift Type Ia supernova was presented by Nic Walton (ING) (Appendix B.10). 13.1 The WFS was employed with associated allocated Spanish time to search for new SN in August/September 1999. The campaign successfully found ten new SN in the required red-shift range. These were photometrically and spectroscopically followed up with a number of La Palma and French telescopes. 13.2 NAW noted that the software employed at the telescope to analyse and make subtractions of the reference and search images could have more general use for searches of transient objects (both in brightness and position). 14. A status report on the Millennium Galaxy Catalogue (MGC) was presented by Simon Driver (St Andrews) (Appendix B.11). 14.1 30 deg^2 in Kitt Peak B in a zero deg strip has been observed. Emphasis is placed on coverage in a single band. 14.2 The MGC is now a substantial catalogue, out sizing that from the EIS. 14.3 Key issues concern linking the MGC counts to galaxies at the bright end (photographic surveys) and at the faint end (CCD data). 14.4 The present data set contains 6500 galaxies sufficiently bright (B<19.5) to be observed by the 2dF survey. This will give their red-shift distribution. 15. Virgo and the galaxy populations of nearby clusters was presented by Jon Davies (Cardiff) (Appendix B.12). 15.1 The motivation and objectives were given. 15.2 To date the WFS data is giving ~5000 galaxies/deg^2 to a surface brightness of 23 mag/arcsec^2 in B. However, this result is currently open to revision as refinements to the use of the Sextractor programme are made. 15.3 Colours in U, B and Z have been obtained for 3 fields at the cluster centre. Radio data is being obtained at Jodrell Bank. 16. Results from the search for High red-shift quasars were shown by Robert Sharp (IoA) (Appendix B.13). 16.1 A key aim is to determine the source of the flux responsible for the re-ionisation of the universe at z=5. The project requires r', i', z' data for z>5 objects. 16.2 Modeling predicts that surface densities of quasars at high z will be ~5 4 Nic > > Can you tell me when these are likely to be updated? The information > on the progress of the surveys etc is pretty old and I was hoping to > give an update to the Wide Field Astronomy Panel. > > Thanks > > Colin >