CFHTLS & MegaPrime News

News from January 2004 and on

[previous 2001-2003 news have been integrated to the new CFHTLS site in early 2004 and aren't considered updates anymore].

Feb. 21, 2008 0038 Cosmologists unfold the dark cosmic web Science
Aug. 06, 2007 0037 Notes on 06B&07A semesters - Missing i' filter General information
Jul. 16, 2007 0036 Details of the T0004 release Data release
Jul. 04, 2007 0035 T0004 Official Release available to LS members Data release
Feb. 16, 2007 0034 World wide release: first stacks + catalogs by Terapix Data release
Jan. 02, 2007 0033 CFHTLS and the CFHT at 2007 Users' Meeting Science
Oct. 23, 2006 0032 The CFHTLS continues after strong earthquake General information
Jun. 30, 2006 0031 World wide release: year 1 of CFHTLS public Data release
Apr. 18, 2006 0030 Wide survey: opening of the W4 field General information
Apr. 07, 2006 0029 Statistics: 3 years of scientific operations Instrument performance
Mar. 10, 2006 0028 XMM-LSS/CFHTLS cluster catalogue available Science
Feb. 18, 2006 0027 T0003 Wide+Deep release candidate available Data release
Dec. 13, 2005 0026 Discovery of an exotic Kuiper Belt Object Science
Nov. 22, 2005 0025 SNLS First Scientific Results Announced Science
Sep. 12, 2005 0024 Note on the half-mosaic night in 05BQ03 Instrument performance
Aug. 25, 2005 0023 Semester 05A completed & Semester achievements General information
Aug. 12, 2005 0022 MegaPrime image quality further improved Instrument performance
Jul. 05, 2005 0021 CFHTLS mid-term review: SAC and Board reports General information
Apr. 05, 2005 0020 Dates of CFHTLS releases to the world Data release
Feb. 22, 2005 0019 Semester 2004B completed, 2005A started General information
Jan. 21, 2005 0018 CFHTLS Workshop in Paris - May 2005 General information
Nov. 24, 2004 0017 04BQ04 brings promising IQ improvement Instrument performance
Nov. 23, 2004 0016 Terapix Data Release T0001 available at CADC Data release
Nov. 04, 2004 0015 04BQ01 light contaminated data corrected Data release
Oct. 06, 2004 0014 04BQ02 & 04BQ03 CCDs 27->35 problem General information
Sep. 28, 2004 0013 The CFHTLS and the CFHT 25th anniversary Public outreach
Aug. 25, 2004 0012 04BQ01 technical misfortunes General information
Aug. 23, 2004 0011 All 2003 Deep data reprocessed Data release
Aug. 04, 2004 0010 Semester 2004A completed General information
Jun. 29, 2004 0009 CFHTLS Elixir data flow Data release
Jun. 13, 2004 0008 CFHTLS 04A (thus far) Elixir data ready Data release
Jun. 02, 2004 0007 Image quality investigation Instrument performance
Mar. 18, 2004 0006 CFHTLS 03B Elixir data ready Data release
Mar. 15, 2004 0005 Terapix CFHTLS performance pages Data information
Mar. 03, 2004 0004 Survey Depth Goals page updated General information
Feb. 11, 2004 0003 CFHTLS Primary Site at CFHT reshaped General information
Feb. 05, 2004 0002 CFHT's CFHTLS Observing Status site opening Data information
Jan. 15, 2004 0001 CFHTLS Elixir data first release Data release

Feb. 21, 2008: Cosmologists unfold the dark cosmic web

Based on the Wide survey, the new result probes the dark cosmic web to unprecedented scales. The press release can be found here.

Aug. 06, 2007: Notes on 06B&07A semesters - Missing i' filter

The end of the 06B was marred by two serious mechanical issues on the instrument: first the shutter failed and had to be replaced (run 07Am04), then the filter juke-box failed, destroying the i' band filter in the process (run 07Am05). The consequence of both of these failures was shortened observing runs, translating in less data collected for the CFHTLS (and PIs alike). The nights were however not lost as WIRCAM was immediately mounted on the telescope.

MegaCam was promptly repaired in both cases by the CFHT staff and a special team from CEA (who built MegaCam). Note that it was not just a repair: in depth studies and understanding of the origins of the accidents led to implementation of mechanical, control, and monitoring improvements. Teams worked around the clock at the summit for four weeks, and MegaPrime found its way back on the sky, successfully completing a short run under bright Moon conditions at the end of July.

A new i' band filter is on order from Barr Associates and should be delivered to CFHT mid to late October. An intense calibration program of the filter will ensue on the sky to quickly reference this new filter to the previous one (specifications are similar for both filters of course).

Note that both of these failures had no impact at all on the data, hence the "Normal Instrumental Setup" flag set for the observing runs.

Jul. 16, 2007: Details on the T0004 release

The Synoptic T0004 Table is now completed and opened to the public without restriction at TERAPIX. This table provides useful information about the release and all quality assessment meta-data built during the production of the release. The table is available here and the summary of the T0004 can be found here.

The T0004 release has important new features as compared to T0003:

  • 2 sets of Deep data :
    - 25% best seeing images stack (D25: ugriz with seeing down to 0.60" for some fields and with exposure time up to 15hrs)
    - 85% best seeing images stack : the deepest wide field ever produced with a 4-meter telescope (D85: up to 50 hrs in some fields)

  • New Wide field: addition of the W4 wide field with 11 deg2 complete in ugriz.

  • New astrometry: done with 2MASS instead of USNO. It provides more accurate and more stable astrometric calibration than T0003 for ALL fields (a gain in rms internal and external error position by a factor of 2). It also provides a more stable an accurate field-to-field photometric calibration.

  • Number of stacks:
    - Deep: 40 [20 D25 (4*ugriz) + 20 D85 (4*ugriz)]
    - Wide: 524 (35 u + 133 g + 173 r + 137 i + 46 z)

    The following image is a 1% section of one of the Deep85 field gri color image. This image produced by Terapix comes from the PNG file made available in the Synoptic T0004 Table.

  • Jul. 04, 2007: T0004 Official Release available to LS members

    The CFHTLS consortium is pleased to announce that Release T0004 of the Terapix LS Stack Products is now available for download from the CADC.

    Retrieval instructions are available at:

    Information about the content of the T0004 release, including an explanation of the stacking process, stack naming conventions and catalog production details, can be found on the Terapix Summary page at:

    These image/catalog stacks are available only to registered CFHTLS members and all regular CFHTLS data access rules apply to the use of these data products. To review the rules regarding data use and access please see the Data Access Policy pages at:

    Feb. 16, 2007: World wide release: first stacks + catalogs by Terapix become public

    As announed on the the Worldwide CFHTLS Data Releases page, the first worlwide release of CFHTLS stacks and catalogs created by Terapix has become public and is available at CADC. It consist of Terapix' release T0003 which was offered to the French and Canadian communities on Feb, 16 2006.

    Jan. 02, 2007: CFHTLS and the CFHT at 2007 Users' Meeting

    The CFHT Users' Meeting will be held in Marseille (France), May 9-11, 2007. The meeting is now open for registration at the following address.

    CFHT Users' Meetings are held every three years. They are a unique opportunity for the members of the CFHT communities to share their scientific results and discuss their ideas on the evolution of the Observatory and on the kind of instrumentation and services they would like the CFHT Corporation to offer in the coming years.

    This meeting will be of paramount importance to get a chance to build any ambitious observing program covering 2008B to the early 2010s, as well as for any new instrumentation for 2013 and beyond. The CFHTLS is strongly encouraged to come to share science results, and present its interest in other future surveys with the CFHT.

    Oct 24, 2006: The CFHTLS continues after an strong earthquake paralizes the CFHT for 4 nights

    On Sunday the 15th of October at 7:07AM, a 6.7 magnitude earthquake lasting up to 30 seconds hit the Big Island. The epicenter was located only 20 miles away from Waimea and gave a good shake to the CFHT headquarters, as well as the telescope and the dome. Fortunately, no one really got hurt on the island. It was the strongest earthquake recorded on Mauna Kea since telescopes have been erected there, but the damages on the CFHT have remained fortunately limited (the entire telescope jumped up and down by more than an inch!). Thanks to the heroic efforts by the CFHT staff, the telescope was back on the sky only four days after the earthquake. This page documents the repair effort on a daily basis, along with photographs of some of the damages.

    Quality assessment on the first sky data show the telescope pointing was off by only one arcsecond, and MegaPrime's image quality has not been altered by the quake. The CFHTLS data collection has resumed since.

    The main earthquake was followed a few minutes later by a short (5 sec.) independent (not an aftershock, more of a consequence of the main quake) magnitude 6 earthquake. Since then, more than a hundred small aftershocks have been felt in the area. The following map from the Hawaiian Volcano Observatory points the location of the various events:

    June 30, 2006: The first year of CFHTLS data becomes public

    As planned in the original survey schedule, and following the agreed calendar, the first year of CFHTLS data (individual frames, the Elixir ones) is now public and available for download at CADC. Please follow the link from the CFHTLS world wide release page.

    The first worlwide release of Terapix data with happen in February 2007 and will be the T0003 release.

    Apr. 18, 2006: Wide survey: opening of the W4 field

    To balance the effect of the bad weather during the past winters on the W2 field, following a recommendation by the SAC, and following a consultation with the Wide community, it has been decided to open a CFHTLS-Wide W4 field observable during the spring and summer. The field is located in the region of SA 22, and fully covers the VVDS 22h field and the UKIDSS DXS field. The W4 field is centered on RA=22:13:18/+01:19:00/2000. The first 16 pointings are already entered in QSO's PH2 for the semester 06A. The pointings have been chosen such that linear scales of at least 4 degree are available, and also to avoid as much as possible high extinction areas where no other ancillary data are available.

    The field is outlined on the following plot: the first set of fields are plotted with solid green squares (the dashed line represent the possible new pointings should the data gathering on W2 get even worse over the next year). The blue square represents the VVDS area, and the red square represents the UKIDSS DXS coverage.

    Apr. 7, 2006: Statistics over three years of scientific operations

    The following page presents statistics on 1) Observing efficiency, 2) Image quality and seeing, 3) Sky transparency and brightness, and 4) Data rate, over the first three years of operations of the MegaPrime/MegaCam imager.

    These statistics aim at showing 1) the dramatic effects from the bad winters Mauna Kea has been suffering from for 3 years in a row now, 2) the consistent image quality & seeing performance of MegaPrime mounted on CFHT atop Mauna Kea, 3) the high rate (75%) of photometric conditions (when the dome is open...) during observations, and 4) the regular increase of number of images acquired per night through the gradual elimination of various observing overheads over the first 2 years of operation.

    The page, which will be updated regularly, can be found in the "Statistics" sections of both the MegaPrime/MegaCam and CFHTLS Data sites. The direct link can be found here.

    Mar. 10, 2006: XMM-LSS/CFHTLS cluster catalogue available

    The CEA group led by Marguerite Pierre just announced that the first part of the XMM-LSS cluster catalogue, which includes the W1 region, is now available on line at the following location:

    For external users, there are two levels of data access:
    - loging as guest: browse the gallery of identified clusters.
    - registering as a visitor: retrieve the catalogue of public clusters (both optical and X-ray data).

    Feb. 18, 2006: T0003 Wide+Deep release candidate available

    At the request of CFHT's Science Advisory Committee, Terapix and CADC agreed to release "candidate" (pre-release) data products that are not yet in their final form. This release candidate consists of partial datasets that concern the Deep and the Wide data obtained between June 1, 2003 and September 5, 2005. The Pre-Wide and the Very Wide components of the survey will be made available in the coming weeks. The stacks and catalogues were produced for the Deep and Wide fields in u,r,g,i,z at Terapix and are available from the Canadian Astronomy Data Centre (CADC) exclusively to authorized CFHTLS users.

    Location: You can find Terapix stacks and catalogues for the four Deep fields (all in u,g,r,i,z) and the three Wide fields (some in u,g,r,i,z, some in g,r,i and some in i only, depending on the progress of the survey in each filter) at this location.

    In particular the tree*.html files provide user friendly description and access to the data.

    Information: individual u,g,r,i,z stacks and catalogues, chi2 images and catalogues as well as quality assessment data on the Deep and Wide are available. Various useful information about the Deep and Wide data sets can be found also.

    Validation: These data have been fully validated by Terapix and are suitable for scientific exploitation. It is expected that the final full release dataset will be basically identical to these stacks and catalogues. In case of modifications prior to the full release, Terapix will make a new announcement. More details about the Deep and Wide data will be soon available on the terapix pages.

    Dec. 13, 2005: Discovery of a Large Kuiper belt object with an Unusual Orbit

    A team of astronomers working in Canada, France and the United States have discovered an unusual small body orbiting the Sun beyond Neptune, in the region astronomers call the Kuiper belt. This new object is twice as far from the Sun as Neptune and is roughly half the size of Pluto. The body's highly unusual orbit is difficult to explain using previous theories of the formation of the outer Solar System.

    Currently 58 astronomical units from the Sun (1 astronomical unit, or AU, is the distance between the Earth and the Sun), the new object never approaches closer than 50 AU, because its orbit is close to circular. Almost all Kuiper belt objects discovered beyond Neptune are between 30 AU and 50 AU away. Beyond 50 AU, the main Kuiper belt appears to end, and what few objects have been discovered beyond this distance have all been on very high eccentricity (non-circular) orbits. Most of these high-eccentricity orbits are the result of Neptune "flinging" the object outward by a gravitational slingshot. However, because this new object does not approach closer than 50 AU, a different theory is needed to explain its orbit. Complicating the problem, the object's orbit also has an extreme tilt, being inclined (tilted) at 47 degrees to the rest of the Solar System.

    See the Official Press Release.

    Nov. 22, 2005: SNLS First Scientific Results Announced

    The CFHTLS is a major undertaking from Canada and France, who have joined a significant fraction (about 50%) of their dark and grey telescope time for a large survey using the wide field imager MegaPrime equipped with MegaCam. MegaCam is a mosaic of 36 CCDs, covering a 1 degree x 1 degree field of view. With a PI-less structure and a data access policy granting equal and immediate access to any member of the two communities to both processed images and catalogs, the CFHTLS is both exciting and challenging. The CFHTLS is made of 3 separate surveys, one of which is a deep synoptic survey, the "Deep".

    A subset of this Deep survey is used by the SuperNova Legacy Survey collaboration (SNLS) to measure the distance to far supernovae and the equation of state of Dark Energy.

    The first results of the SNLS survey, to be published in Astronomy & Astrophysics, place strong constraints on cosmological models. Read the full AAS Press Release.

    See also the full Astronomy & Astrophysics Press Release, and the press release of the University of Toronto.

    The CNRS in France has also issued a press release in collaboration with the CEA. CNRS Press Release, CEA Press Release. [In French]

    The SNLS relies on spectroscopic follow-up of the SNe candidates detected on the CFHTLS imaging data, each major telescope involved

  • Gemini press release
  • PParc press release
  • Keck press release

  • For more information: The SuperNova Legacy Survey

    Sep. 12, 2005: Note on the half-mosaic night in 05BQ03

    The failure of the CCD controller south (slave) power supply on August 30th 2005 forced the telescope to shut down for the night of the 31st, and forced QSO to operate the instrument in a half-mosaic mode for the night of the 1st to the 2nd of September (CCDs 18->35 not active).
    Note that CFHTLS data were obtained that night but are not included in the CFHTLS data set of this web site as they do not have a legacy value since they are much altered compared to the normal data (they were taken only for the SNLS and the Very-Wide real-time needs). The odometer numbers for these CFHTLS exposures are:

    Deep/SNLS (sequential): 809995 to 810004
    Very Wide: (sequential): 809963 to 809994

    After heroic work by the CFHT on the 2nd of September, the origin of the problem was identified (a weak +5V line in the power supply) and MegaCam was back on the sky on the evening of the 2nd in full field-of-view mode.

    Aug. 25, 2005: Semester 05A completed & Semester achievements

    The 05A semester has been remarkably trouble free on the instrument side, but badly affected by mediocre weather on Mauna Kea for the second year in a row. The document prepared by CFHT for the CFHTLS mid-term review presented clear statistics on the reasons why the CFHTLS did not get data as fast as initially planned. The main culprits over the first 3 semesters (03B-04A-04B) were the weather, technical problems (including the image quality, an issue mainly solved last December), and the observing overheads.

    While nothing can done about the weather, CFHT was happy to report a few months ago, that the instrument failures have dramatically decreased in 05A as did the observing overheads, specially the focusing which gobbled about 50mn a night. The acquisition time of guide stars was also greatly improved in May thanks to the speeding of the guide probes (a factor of 10 faster, allowing exposures to start earlier).

    The autofocus currently implemented, a thermal model of the telescope, has been extensively tested over the past runs to compare it with the time consuming focus sequences (deriving the focus from the PSF of stars on images taken at different focus positions). Our confidence level in the autofocus rose quickly, and the number of focus sequences needed through the night has dramatically decreased, with many nights free of focus sequences (sometimes the seeing is so unstable that it confuses the QSO observers, who feel naturally compelled to run some focus sequences - one again, blame it on the current Mauna Kea poor conditions). After a few observing runs in this autofocus mode, we estimate the time saved to approximately 50 minutes per night. There is however still a substantial on-going effort to optimize the thermal model of the telescope to ensure the best focusing: the current model still exhibits a 50 microns scatter (the depth of field under 0.5 arcsec. seeing) and we want that scatter to go down to 25 microns at least.

    In addition to saving the time previously used for focus sequences, autofocus is increasing the QSO validation rate! Changes in focus are activated between each exposure and, as a result, less exposures are lost. For example, the run 05AQ05 has a QSO validation rate of 93%, which is the highest rate ever achieved in the history of QSO (CFH12K included). Even better, the validation rate on all CFHTLS in 05AQ06 data is 98%!

    The weather, technical problems and engineering time over the runs since autofocus has become active, amounted to 20 to 27% (mostly weather and bits of engineering to work on the wide-field corrector tuning), a range we consider higher than normal for this time of the year. Indeed, 75% of the time during the first three runs (winter/spring period) of 05A was lost to bad weather. On the recent spring/summer runs we however achieved 5.2 hours of validated data (QSO definition) on average for nights of only 8.2 hours. When scaled to the average night length over the year of 9.5 hours, we end up with 6.0 hours per night of validation rate. With hopefully better weather conditions in the near future, expect this number to improve this coming semester.

    What does it mean for the CFHTLS and for MegaPrime QSO for the immediate future? The CFHTLS re-design done after the recent review process developed a strategy based on an average of 4.5 hours per nights of validated observations. If 5.5 to 6.0 hours per nights were indeed to be the average for the coming semesters, the CFHTLS could actually speed up its planned pace, catching up on some on the time lost over the past two years, and decrease its load on the Canadian and French time allocation.

    What is coming on the instrument side for this semester? After a fruitful visit at CFHT by the CEA team who built MegaCam, a preventive maintenance plan is being put in action. This plan should save us from failures related to the aging of the instrument. After a year of intense work on the image quality issue, we are very close to the final configuration that will put the instrument within the design specifications (see news 0022). And since every second counts, CFHT will pursue its efforts at hunting down and decreasing each and single observing overhead.

    And now back to the CFHTLS status as of Aug. 15, 2005 (including the first observing run of 2005B = 05AQ01), the CFHTLS distribution of validated exposure is (including the 40 seconds overhead per exposure):

     Survey Component :                   Deep      Wide     Very Wide  Pre-Survey
     Total integration [validated exp.]:  451.1hr   214.0hr  158.7hr    23.3hr
     Number of validated exposures:       3840      1515     3528       381
     Current fraction of CFHTLS :         54.8%     26.0%    19.3%      n/a
     Target(*) fraction of CFHTLS :       44.0%     34.0%    22.0%      n/a
     (*) These numbers and future figures will be adjusted to reflect the recent
     decision of halting the Very Wide component of the CFHTLS as of 2005A.

    For a full report on the semester seen from the QSO side, the report has just been posted here.

    Aug. 12, 2005: MegaPrime image quality further improved

    With the knowledge accumulated over the past two years and a half, since CFHT started tweaking MegaPrime's wide-field corrector, we predicted an improvement of the image quality over what was produced by the fortunate flipping of the lens L3 (News 0017) by applying a change in height of the entire corrector versus the primary mirror. Shortly before the start of 05AQ07, the corrector was pushed further up (away) from the mirror by reducing the spacer from 5.5mm to 2.1mm.

    The predictions were correct: the upper left corner and the lower right corner have been recovered and the analysis of the new data show that we are now only left with a slight tilt of the focal plane, which we plan to adjust within this coming 05B semester.

    The improvement is significant in terms of homogeneity of the image quality over the entire field of view. It is stable over several nights, and stable in shape in various telescope configurations in all tested filters. Most important for the most demanding scientific project in terms of image quality, gravitational lensing, the optical distortion is now almost gone!

    The following plots (produced by Terapix' QualityFITS pipeline) show the image quality histogram, as well as the orientation and ellipticity map on the entire image for the three main configurations encountered so far: prior the L3 flip, the L3 flip configuration, and the new configuration (these maps were derived from 30 seconds long, guided exposures on dense stellar fields). Note that the scale of the histogram is different from one configuration to the next (the Y scale represents the number of CCDs having a given FWHM: the highest, the most uniform the image quality is across the field of view). Note how the histogram skewness has now significantly decreased compared to the L3 flip configuration. The individual postscript files can be retrieved here. New (Spacer) - Jul.05:, Initial - May 03:, L3 flip - Dec. 04:,

    In terms of quantitative measurements, here is the IQ map for the exposure 807216o:

     IQ MAP & CLASS TO MEDIAN 1-9 | *0 |+/-1|+/-2|+/-3|+/-4|+/-5|+/-6|+/-7|+/-8|+/-9|
     IQ difference from median:    0.02<0.06<0.11<0.17<0.22<0.28<0.34<0.39<0.41<...
     |0.48|0.45|0.43|0.44|0.43|0.43|0.45|0.48|0.49|    |+1|*0|*0|*0|*0|*0|*0|+1|+2|
     |0.47|0.45|0.43|0.44|0.45|0.45|0.44|0.44|0.44|    |+1|*0|*0|*0|*0|+1|*0|*0|*0|
     |0.49|0.46|0.44|0.42|0.43|0.42|0.42|0.41|0.43|    |+1|+1|*0|*0|*0|*0|*0|-1|*0|
     |0.50|0.47|0.46|0.43|0.42|0.41|0.42|0.43|0.44|    |+2|+1|+1|*0|*0|-1|*0|*0|*0| 
    This is truly spectacular: the image quality degradation from center to the edge does not exceed 0.08 arcsec. on this 0.42 arcsec. seeing exposure! However, when we look at frames taken slightly in and out of focus from the optimal position, we clearly see some areas of the focal plane getting worse. This is an indication of a slight tilt of the focal plane (which can actually be derived using these exposures). The following diagram proposes the orientation and ellipticity map for four images: 807216o.fits and 807221o.fits taken at the optimal focus (to prove it is repeatable), 807218o.fits taken at +90 microns from the optimal focus, and 807224o.fits taken at -90 microns from the optimal focus. The individual postscript files can be retrieved here:,,,,

    807218o exhibits a degradation of the image quality in the lower left part, while 807224o exhibits a degradation of the image quality in the upper right. With no tilt present on the focal plane, a defocus would bring a uniform degradation of the image quality across the entire field of view. The current optimal focus shows signs of image degradation towards the left, but this is minor compared to what we've been facing until now.

    The MPIQ project web site goes in much detail on this latest evolution.

    Jul. 15, 2005: CFHTLS mid-term review - SAC and Board reports

    Following the reviewing process conducted by the SAC in May 2005, and based on documents produced by the CFHTLS Steering Group (soon available on this site), the SAC recommendations and subsequent decisions by the CFHT Board of Directors are presented on the following page:

    Apr. 05, 2005: Dates of CFHTLS releases to the world

    Following the CFHTLS data policy defined by the CFHT Board of Directors, the CFHT Legacy Survey, which officially started on May 30th 2003, will be open to world access at CADC for the raw FITS and the Elixir data products (Processed FITS and ancillary data) according to the following time table:

    Jun. 30, 2006: All data from May 30, 2003 to May 29, 2004
    Dec. 30, 2006: All data data from May 30, 2004 to May 29, 2005
    All data taken after May 30, 2005 will become available 13 months later.

    The rules for the Terapix data releases to the world (stacked ressampled FITS and catalogs) will not follow those for the raw and Elixir data, or the ones for the Terapix releases to the Canadian & French CFHTLS community (releases every 6 months).

    Terapix data products will be made available to the world through data releases yet to be defined, happening in various steps: the first one not later than three years after the beginning of the survey (Jun. 30, 2006), and the last one not later than one year after the end of the survey.

    Feb. 22, 2005: Semester 2004B completed, 2005A started

    The 04B semester started with some great weather but ended bad with the predictable bad winter weather battering the Big Island. This is clearly reflected in the statistics page for the 3 surveys where it appears that the winter field for the wide (W2) got affected due to the natural priority that goes to the deep when the weather affects the operations. Note also that since the finding of the L3 flipping effect, the wide had been put on hold in November, then the weather really affected the QSO runs: in the end the wide got short of its allocated time.

    Note that on the top of that statistics page that the blue framed table now reflects the time as allocated and validated by the CFHT observing process: a 40 seconds overhead is included per exposure. This changes only slightly the stats based on pure integration time (the following tables remain in that form), and it matches exactly the statistics published daily by the QSO team here.

    As of Feb. 2005 (including the first observing run of 2005A), the CFHTLS distribution of validated exposure is (including the 40 seconds overhead per exposure):

     Survey Component :                   Deep      Wide     Very Wide  Pre-Survey
     Total integration [validated exp.]:  343.8hr   145.6hr  109.9hr    9.2hr
     Number of validated exposures:       2898      959      2484       151
     Current fraction of CFHTLS :         57.4%     24.3%    18.3%      n/a
     Target fraction of CFHTLS :          44.0%     34.0%    22.0%      n/a

    For more information on the past semester, feel free to browse the 04B QSO report.

    The whole instrument MegaPrime/MegaCam has behaved very well this past semester after the first run which got affected with several problems (cf News 0012). The most important breakthrough was the finding of the L3 lens flipping (November News 0017) in the wide field corrector which brings a spectacular improvement of the image quality. This has been accepted as the default configuration and won't be changed for another semester (or more) until CFHT finds a mean to improve even further the image quality (in the top right and bottom left extreme corners in particular). With this major step behind, the CFHT staff can now focus on the new top priority for MegaPrime: the reduction of the observation overheads to gain more open shutter time per night.

    As per the data flow, we have now had for several months a steady release of Elixir products to CADC within 10 days after the end of an observing run (and then to the community within minutes thanks to the highly automated scheme in place at CADC).

    Jan. 21, 2005: CFHTLS Workshop in Paris - May 2005

    With the first Terapix-processed data released a few months ago and the second one (T0002) looming, 2005 should see promising preliminary scientific results based on CFHTLS data, marking the first of many productive years of CFHTLS-related science.

    Spring 2005 will be the time of the first external review of the CFHTLS. The comments of the reviewers will be evaluated by the CFHT Scientific Advisory Council (SAC) just before its May 2005 meeting in Paris. To offer to the CFHTLS users an opportunity to share their experience with the CFHTLS data and their first scientific analysis, and to provide SAC with an opportunity to meet CFHTLS users, CFHT will organize a CFHTLS workshop in Paris (IAP) just before the SAC meeting, on May 26 and 27.

    The CFHTLS, with nearly 500 nights allocated over five years, is a major commitment of the Canadian and French Agencies as well as a major program for CFHT. HIA and INSU have both agreed to provide a limited support to those who want to present their CFHTLS experience at this workshop.

    Information and a registration form can be found on this dedicated site:

    The sooner you register, the better for the organizers! If you have any question, feel free to ask the organizers at:

    Nov. 24, 2004: 04BQ04 brings promising IQ improvement

    As part of the ongoing effort for understanding the issue with the MegaPrime image quality (IQ), the wide-field corrector was taken apart one more time between the 04BQ03 and 04BQ04 observing runs and the lens L3 (third from the bottom) was accidentally mounted back up side down in the corrector (the mechanical design allows for this mistake to easily happen, plus this weak lens is almost flat, making it difficult to say on which side the curvature is). The mounting error was quickly discovered within the first moments of the first night of the 04BQ04 run because the focus level (vertical motion) appeared to be shifted down by 5 mm, causing the system to hit some limit switches, and making it impossible to focus in the g' and z' filters. However, to our surprise, the image quality in that configuration in the u*, r' and i' filters (which could be well focused) appeared to be the best ever achieved on MegaPrime!

    The Q observing run could however not be completed in such circumstances (no access to the g' and z' filter), hence the lens was flipped back on the 4th of November and observations started again on the 5th in the optical configuration adopted since June 2003.

    Investigations show that the L3 flipped mode is not the way the design calls for that lens to be mounted, and the reasons for the improved image quality are not currently understood but have provided useful clues on the behavior of the wide-field corrector.

    However, the improvement is so significant (and stable over several nights, and stable in shape in various telescope configurations) in all tested filters (u*, r', and i') that CFHT, in consultation with the CFHT Astronomy Group, the SAC, and the CFHTLS Steering Group, has decided to mount the lens L3 flipped and shift down the limit switch (within safe limits) such that it'll be possible to focus in all filters. The image quality will be thoroughly investigated in that new configuration during the first hours of the coming run (04BQ05) to confirm that we have indeed gotten back to the best IQ configuration.

    Note however that this is still not what the specifications called for, in particular the top left and bottom right corner exhibit a sharp degradation of the image quality. Yet, not only is the the PSF better across the field, it is also a lot more uniform in shape, as shown by the orientation and ellipticity maps. In that respect, the Wide Survey coordinator has preferred to put on hold the observations on the Wide for the duration of this past 04BQ04 run (except for the Pre-Survey which is about astrometric calibration and is not impacted by the IQ issue), in order to benefit from the planned significant IQ improvement coming this December run. Hence no surprise seeing the fraction of the Deep increase versus the Wide and Very Wide in this month's global statistics.

    This new optical configuration will be adopted as the new standard until further investigations (the MPIQ project) show we can move even further in terms of image quality improvement.

    The following plots (courtesy of Terapix) show it all about the before and the after of the L3 flip. The two images are low galactic latitude fields with plenty of stars. 761016o.fits was taken with the June 2003 adopted configuration, and 767523o.fits was taken during the first night of 04AQ04. Elixir reports the same image quality at the center of the field, and they're both taken in the i' filter, and are fairly long exposures (more than 60 seconds). With L3 flipped, the seeing histogram over the field of view gets nicely tightened by a factor of two at least (note that the Y scale on the histogram is different for these two diagrams) and the orientation and ellipticity map show only strong trends in two extreme corners (note that the sticks size has no absolute length value, it's relative within each image, these plots aim at showing the fact that the PSF is a LOT more homegeneous with the new configuration). High resolution postscripts files can be obtained here:,,,

    Nov. 23, 2004: Terapix Data Release T0001 available at CADC

    From Yannick Mellier (Terapix) & David Schade (CADC):

    We are pleased to announce the first CFHTLS release, T0001, now available to all CFHTLS registered users at CADC. The release includes the CFHTLS deep data D1, D2, D3 and D4 in u,g,r,i and z bands.

    You can retrieve these data from CADC using your CFHTLS login and password at the following URL:

    Details on the data content and data quality are available at the terapix web site : under the section "CFHTLS Release T0001". These public pages contain a simple summary table and an explanatory table.

    Comments/suggestions regarding the detail of the data processing can be addressed to Terapix at the following adress : terapix at

    We remind you that the CFHTLS release must not be propagated outside the CFHTLS community, and is only for scientific usage by its registered users. If in doubt please consult again the CFHT Legacy Survey Data Access Policy on the official CFHTLS web site at CFHT:

    The next release scheduled for late December will include the W1,2,3 and VW data sets.

    Nov. 04, 2004: 04BQ01 light contaminated data corrected

    The CFHTLS data taken from the 6th (PM) to the 10th (AM) of August covering the odometer range 756829 to 757179, have been corrected using a dark map correction, much like a dark current correction since the contamination was linear with time. Note that these data were plagged by other very serious effects as documented on the news 0012 from this page.

    The complete list of CFHTLS data suffering from that contamination, and now cleaned from it, and delivered to CADC can be found here.

    Oct. 06, 2004: 04BQ02 & 04BQ03 CCDs 27->35 problem

    The failure of a video board in the South controller on the 5th of September 04 forced us to shut down half the mosaic for two nights (CCDs 18 to 35). A new board was installed and everything went back to normal on the evening of the 7th of September. However that spare board failed somehow, though not completely, on the 8th of September and since then the right amplifier of CCDs 27 to 35 have a saturation value of 32K instead of 65K (this is due to the 16th bit being stuck at 0). The processed data (not RAW) will have the MAXLINB (Level at 1% linearity departure) set to 32K on the Elixir processed data to reflect this problem.

    The file number for the images taken in half mosaic mode are 760630 to 760880. The right amplifier saturation change effect started for the exposure 760956.

    Sep. 28, 2004: The CFHTLS and the 25th anniversary of CFHT

    The spotlight goes on the CFHTLS as CFHT celebrates its 25th year of operation. A stunning image of the CFHTLS deep D1 field prepared by Terapix is presented. Read all about it in the CFHT Press Release. Other ressources include the CNRS Press Release and the INSU Press Release.

    Aug. 25, 2004: 04BQ01 technical misfortunes

    The first run of the 04B semester was unfortunately rich in technical problems that altered the quality of the data acquired on the sky. Fortunately, the problems were not severe enough to the point of cancelling the observing run, and data were taken at the usual rate for all three components of the survey.

    CCD Failures:
    Due to the failure of a vaccum connector flange, some CCD pins were pulled down (shortage to the ground) and this resulted with many amplifiers ceasing to function normally. Most of them started recovering right away though and by the middle of the run, only two amplifiers were still non functional (A amplifier of CCD07 and CCD11, that is the left amplifier). Some amplifiers were noisier than others (A amplifier of CCD17, B amplifier of CCD16 and CCD14) in the first days of the observing run but this did not damage the data enough for having these sections masked by the Elixir processing. Two masks have been built for the Elixir processing: one for the period Aug. 6 to Aug. 14 (A amplifier of CCD00 and CCD11 off, and both amplifiers of CCD07, CCD09, CCD10 off - left image below) and one for the Aug. 14 to Aug. 24. period (A amplifier of CCD07 and CCD11 off - right image below). The Elixir masks for these two periods are 2004B.mask.0.36.00.fits and 2004B.mask.0.36.01.fits.

    Internal light contamination:
    A vacuum gauge of a new type was installed just prior the observing run because of the repeated failures of the original model. Unfortunately, the internal filament emits visible photons and this resulted in a light contamination inside the cryostat. With all the focus going on the CCD problem, it took us a few days to react on that one issue since it was not very obvious (0.5 ADU per second in the brightest areas). The raw data are salvageable by simply subtracting a master dark frame. The gauge was shut down on the morning of the 10th of August.

    Ice on the cryostat window:
    The air flow venting the outside surface of the cryostat window got contaminated with water twice during the run, which resulted in an ice plug forming on the window. The data taken in these conditions are totally useless in the areas occulted by the ice. The CFHTLS data taken on the night it first happened (from the 6th to the 7th) are mostly good with only a few patches of ice left (the ice melted throughout the night fortunately thanks to the ambient dry air of Mauna Kea). No CFHTLS data were taken on the second night it happened (16th of August from dusk until 1 AM of the 17th) while there was ice on the window (the QSO team waited patiently for it to melt after clean dry air was flowing again on the window and dissipated the ice plug).

    Aug. 23, 2004: All 2003 Deep data reprocessed

    As announced on this page last June, new photometric superflats have been derived from a data grid obtained in all 5 bands (u*, g', r', i', z') in October 2003, and are now available at CADC. Following a request from the SNLS groups, all the Deep survey data gathered in 2003 have been reprocessed with a flat-field that includes that new map. The improvement in photometric quality is significant with a zero point uniformity across the mosaic of +-0.01 mag level for g', r', i'i, z', however the u* band still shows residual up to +-0.05 mag offsets that need to be understood.

    Chris Pritchet has studied the residuals from these grids processed by the photometric superflat derived from them. His results can be found here.

    All 2004 data available at CADC already include the improved photometric superflat which was derived after the light baffle was installed on MegaPrime at the beginning of the year.

    Aug. 04, 2004: Semester 2004A completed

    The 2004A semester has just finished and the first MegaPrime run of 2003B is only a few days away. More than a year after starting the survey, the general statistics are:

     Number of exposures (a very rough estimator to check on progress):
                     Deep    Wide    Very Wide
             2003A   457     137     399
             2003B   889     383     773
             2004A   779     147     685
    We see here, as announced earlier by the Steering Group, that the Wide survey has taken the back seat during this past semester and has accumulated little data. The Wide survey will be getting higher priority for the coming semester in order to keep the three components balanced. In terms of balance, here is the current ratio today at the end of 2004A:
     Survey Component :                   Deep      Wide     Very Wide  Pre-Survey
     Total integration [validated exp.]:  166.4hr   55.8 hr  43.5 hr    2.7 hr
     Current fraction of CFHTLS :         62.6 %    21.0 %   16.4 %     n/a
     Target fraction of CFHTLS :          44.0 %    34.0 %   22.0 %     n/a
    Starting 04Am06, CFHT has done its best to stick to its commitment of providing the Elixir processed data to CADC within 7 to 10 days after the end of the observing run. So far, this timescale seems reasonable as CFHT has been able to deliver on time.

    Jun. 29, 2004: CFHTLS Elixir data flow

    Following a recommendation from the CFHTLS Steering Group, the CFHT and CADC have commited to have the Elixir data made available to the community within 3 weeks after the end of an observing run. We are testing our capabilities with the 04AQ06 run, expect these data to appear at CADC mid-July as they will start transferring from CFHT early July.

    Jun. 13, 2004: CFHTLS 04A (thus far) Elixir data ready

    After the light baffle was installed on MegaPrime at the beginning of the semester 04A, new photometric superflats had to be built. We have now obtained these frames at a much higher quality compared to the equivalent frames obtained early 2003 for the semesters 03A & 03B. This was the primary reason for the delay in releasing the CFHTLS Elixir 04A data (that's from January to May included).

    Note that new photometric grids were obtained just before the light baffle was installed, and the new improved photometric superflats for 03A & 03B will be used in the global reprocessing of the CFHTLS data (since June 2003) which should take place later this fall.

    On the technical side, note that from now on the Elixir data are transfered to CADC via network instead of SDLT tapes.

    Jun. 02, 2004: MegaPrime image quality investigation

    CFHT launches a project focused on investigating the image quality issue in MegaPrime, which turned out to be not as good as expected as the past year of operation has shown. All relevant information and documents are presented here.

    Mar. 18, 2004: CFHTLS 03B Elixir data ready

    The last tapes including the remaining CFHTLS Elixir data from the semester 03B were shipped this morning to CADC. The complete 03B semester will be in consequence fully available on line at CADC within a few days.

    Mar. 15, 2004: Terapix CFHTLS performance pages

    Terapix has released a collection of sophisticated pages providing in depth analyses of the data quality of the CFHTLS Terapix data products (products which will be very soon available at CADC). These Terapix pages can be consulted here.

    Mar. 03, 2004: Survey Depth Goals page updated

    Based on actual sky data and comparison with DIET estimations, the survey magnitude depth goals have been updated. Please keep in mind that the images can be taken within a fairly large range of airmass and sky brightness values for some filters and this explains why some of the these numbers differ slightly from the DIET output. Average values between various conditions were taken to build up these numbers. For information on exact depth based on the actual data, please refer to the individual survey web sites.

    Feb. 11, 2004: CFHTLS Primary Site at CFHT reshaped

    To ease the cruising across the many pages focused on the CFHTLS at CFHT, the primary CFHTLS web site based at CFHT ("CFHT Legacy Survey Home") has been made "look&feel" compliant with the CFHTLS Data ("CFHTLS Observing Status Home") and the MegaPrime/MegaCam pages. This web site aims at proposing the large picture on the survey with many pointers in particular to the sites developed by the CFHTLS community, especially the web sites dedicated to each survey component and run by the individual coordinators.

    Feb. 05, 2004: CFHT's CFHTLS Observing Status site opening

    CFHT is proposing this new web site ("CFHTLS Observing Status Home") focused on providing the scientists interested in the CFHT Legacy Survey a large set of information to follow the progress of the survey from the point of view of the data acquired at CFHT.

    Jan. 15, 2004: CFHTLS Elixir data first release

    The CFHT Elixir data released under the Elixir version 2.0 are now available to download from the CADC for all registered CFHTLS users. The recipies aren't perfect yet, please consult this page to find out about the limitations of that release.

    All these data will be re-released later in 2004 using the Elixir version 3.0 with improved recipes for the fringes in i' and z', the sky background modes correction (all filters), the photometric superflats (all filters) to name the three main areas left to be optimized.