Report of the 71st meeting of the CFHT SAC

Cassis (France), 12-13 May 2007

NOTE: After deliberation by the CFHT Board of Directors regarding SAC's
recommendations, the Board has approved the following recommendations:

     7 - CFHTLS Re-observations of fields
     8 - CFHTLS Wide extension

In addition the Board endorsed in principle the following recommendations

     12 - Large Programs
     13 - Feasibility studies for future instruments

The remaining recommendations will be further discussed at the Board's
next regular meeting, December 2007.

Recommendation #1 WIRCam crosstalk
Recommendation #2 Non-linearity correction of WIRCam data
Recommendation #3 Reporting reduced data-quality issues
Recommendation #4 ESPaDOnS crosstalk
Recommendation #5 ESPaDOnS - Libre-Esprit software
Recommendation #6 Telescope automation
Recommendation #7 CFHTLS Re-observations of fields
Recommendation #8 CFHTLS Wide extension
Recommendation #9 CFHTLS Very Wide extension
Recommendation #10 CFHTLS completion
Recommendation #11 Small multi-year/long-term programs
Recommendation #12 Large Programs
Recommendation #13 Feasibility studies for future instruments

The meeting was attended by SAC members Pierre-Alain Duc (vice-chair), L. Ferrarese, C. Ftaclas, J. F. Gonzalez, C. Gry, H. Hoekstra, R. Jedicke, Dae-Sik Moon, Denis Mourard and N. St-Louis (chair). Jeremy Lim and Shiang-Yu Wang from the Academia Sinica Institute of Astronomy and Astrophysics in Taiwan also attended the meeting. The CFHT director C. Veillet, D. Salmon, J.-C. Cuillandre, Pierre Martin and Hervé Haussel (for the LS steering group) participated in presentations and discussions.

1. Technical Activities Report

SAC would like to see a more detailed breakdown of the development priorities, with at least 3 categories. SAC suggests the following development priorities:

A Elixir WIRCam/WirCam*

A MegaPrime*


A Elixir – ESPaDOnS / Queue ESPaDOnS



C Observatory Automation

C ESPaDOnS from Prime

Derrick Salmon reported on the technical activities of this semester. Work has been carried out on the dome bogies that were displaced after the October 15, 2006 earthquake. The underground conduit for OHANA is now in place and work on the grounding grid repair as well as on the installation of protective barrier posts is in progress.

Back to top

2. Instrument and Project Reports


Since the last SAC meeting, there have been very few problems with Megaprime resulting in only a small fraction of time lost to overheads. Also, a significant source of the slowly degrading sensitivity of the camera has been identified (this was particularly bad in the lower right corner of the mosaic); it was very likely caused by fine aluminium powder on a camera window and ISU plate from rubbing in the filter holder. After cleaning, an 8% improvement in the sensitivity was measured (15% in the lower right corner). A thorough maintenance teardown of the MegaPrime camera is planned for next September and CEA has been invited to participate.

SAC is pleased about these improvements and that a maintenance exercise is taking place in September. We look forward to hearing about the outcome during our next meeting.

We heard that an order was placed with CEA for the retrofitting of the rotating valves and for a spare pulse tube. As for the flange problem, CFHT staff is working on printed circuit boards, which will later be sent to CEA for inclusion in the flanges. SAC supports and encourages these positive exchanges between the two agencies.

J. -C. Cuillandre presented the SNDICE (SuperNova Direct Illumination Calibration Experiment) project aimed at improving the photometric accuracy of MegaPrime data. The concept involves using LEDs covering the spectral range of the camera to illuminate the primary mirror. Efforts are mainly being carried out by the DICE team in France and therefore this project does not represent a significant workload on CFHT staff.

Back to top


Except for the crosstalk problem, WIRCam has been performing well in the past six months. The main efforts have been directed towards the data reduction pipeline, now renamed ‘I’iwi. Given the worries about the pipeline expressed during its previous meeting, SAC views this as very positive improvements. SAC was pleased to hear that there is now more staff involved with WIRCam and its data reduction pipeline.

The main problem that remains is crosstalk between the various amplifiers of the array. The effect, which contributes positive or negative signals, is difficult to see on individual images, and had thus so far been underestimated. It is however clear when a large set of images are stacked together and critical for deep cosmological observations. Some encouraging improvements have been made, particularly regarding the negative crosstalk, and staff is consulting with people responsible for other IR cameras to seek advice.

SAC Recommendation #1 -- WIRCam crosstalk SAC takes note that the negative crosstalk for WIRCam has been reduced. SAC recommends that significant efforts be made to further reduce any crosstalk. Hardware solutions should be given priority over software corrections.

Back to top

We were made aware that another aspect of the WIRCam data reduction that is still not included in the ‘I’iwi pipeline is the correction for non-linearity. This should not represent a major time investment by the staff since all the necessary data are in already in hand.

SAC Recommendation #2 – Non-linearity correction of WIRCam data SAC recommends that a high priority be given to including the non-linearity correction into the WIRCam ‘I ‘iwi pipeline.

Back to top

During discussions concerning the crosstalk in the WIRCam data, it became clear that there are some communication problems between the various parties involved when problems with data quality assessment and data reduction arise. To ensure the best quality of WIRCam as well as MegaPrime reduced data, it is important that all parties be informed when a problem is identified.

SAC Recommendation #3 -- Reporting reduced data-quality issues SAC recommends that CFHT clarify the reporting channel for quality issues for reduced data. This channel should insure that all relevant parties (CFHT, Terapix and CADC) are informed in a timely manner.

Back to top


The operation of ESPaDOnS has continued to be very stable in the past 6 months. Work has been carried out to try to improve the throughput by a better alignment of the fibres using the exposure meter. This seems to work well but the required adjustments are extremely delicate to make. An attempt has been made to implement a temperature-based focus model for ESPaDOnS but did not succeed. Work will continue on this issue during the June engineering run using data recently acquired.

We were also informed about the results of the poll carried out by Nadine Manset regarding the possible reduction of the crosstalk in ESPaDOnS. Although a majority of users who responded to the survey stated they could carry out their present research projects with the current level of crosstalk, there was a strong desire to see the crosstalk reduced to the 1% level.

SAC Recommendation #4: ESPaDOnS crosstalk

Based on the response from the community and on the fact that the crosstalk is significantly lower in NARVAL, SAC recommends that efforts be made to reduce the crosstalk between linear and circular polarisation in ESPaDOnS to the 1% level.

Back to top

SAC was also presented with a progress report on running ESPaDOnS in queue mode. In addition to the obvious benefits of carrying out observations in queue in general, it is clear that many ESPaDOnS programs would benefit from running in this mode since they involve monitoring and time-constrained observations. We were informed that significant steps forward have been made towards incorporating ESPaDOnS in QSO. In particular, phase 2 tools have been designed and are well advanced. Work on designing additional QSO tools has also been carried out. The most challenging task will very likely be to make ESPaDOnS work under the New Environment for Observing (NEO) framework, i.e. to make communications between the instrument and the TCS possible. If all goes well, tests should be carried out in semester 2007B and ESPaDOnS should be offered in QSO for 2008A. The first ESPaDOnS QSO run will most likely be carried out in the presence of the PIs to avoid time lost to reliability issues.

SAC is pleased with the rapid progress in preparing ESPaDOnS to operate in queue mode and encourages the staff to pursue their efforts in this direction. One crucial aspect of queue observing with ESPaDOnS is that it requires an automated data reduction pipeline that can run under the New Observing Process (NOP). SAC is worried that this might be a problem within the current agreement for the Libre-Esprit software package since CFHT staff does not entirely control the source code.

SAC Recommendation #5 -- ESPaDOnS -- Libre-Esprit software

There is an expectation that ESPaDOnS will be run in queue mode and that it will likely be used for Large Programs in the near future. SAC recommends that a mandatory condition for this to occur is that CFHT gain more control over the Libre-Esprit software for purposes of running the queue.

With ESPaDOnS operated in queue mode, there will be even less opportunities for the staff to interact with the CFHT community. To alleviate this problem, CFHT encourages astronomers to come visit for a few weeks.

Back to top


Since the last SAC meeting, the SDSU controller code was completed and changes have been made to the controller itself to correct timing problems in AO mode. An interaction matrix generated with FlyEyes was shown to be nearly identical to one generated with Avalanche Photodiodes (APDs). Finally, we were informed that the AO loop has been closed early in 2007 using the AOB’s artificial star. Poor weather has prevented further progress on this aspect.

SAC congratulates Kevin Ho and his group for this achievement.

Back to top


Work has been carried out on two aspects of the VASAO project. First, a seismometer consisting in an inertial pendulum system was fabricated and has been functioning on a lab bench in Waimea. The aim is to provide an inertial reference frame for VASAO control. The next step is to install it on the telescope and to acquire data.

The second part is the development of a two-colour experiment. A pair of CCD cameras, one looking through a blue filter and the other through a red filter, will be used to acquire images of a bright star at a rapid frame rate. The differential motion will be used to see if atmospheric tip/tilt corrections can be obtained. Results may be obtained before the next SAC meeting.

The CFHT director also described an option being considered in which a laser originally for the ELP--OA project in France but that is not going to be used would be purchased and used for testing the modeless laser concept of VASAO.

SAC approves the limited level of activity on this project by CFHT staff and recognises that the experiments taking place are interesting and represent crucial steps towards demonstrating the feasibility of the VASAO concept. SAC wishes to continue to be informed of the future development of the project.

Back to top


The OHANA activities are continuing although no formal report was presented during this SAC meeting. The CFHT director is taking a pro-active role in discussions with other big telescope directors on the mountain to encourage them to include this project in their future development plans. The parties concerned seem generally responsive but need to be convinced that some selective experiments that might not seem immediately interesting scientifically represent crucial steps towards the future success of this endeavour.

Back to top


Progress has been made on the DIMM telescope attached to the rotating dome. After several attempts using five different modem systems, communication problems between the moving and stable parts of the dome were solved. Also, after problems with the Meade telescope controller, mainly because the software is proprietary and cannot be accessed, it was replaced by a Galil motion controller. The goal is to get the system ready for the end of April and perform field tests. SAC is pleased with these advances and looks forward to see the results of the test at its next meeting.

SAC agrees with this approach and encourage the continuing development of the OHANA project.

Back to top

Telescope automation

The automation of the telescope is an in-house project that is an intrinsic part of the Golden Age plan. Steps towards reaching automation include making the operations more robust and reliable. The initial goal is to achieve remote observing by one observer from Waimea. If there is a problem during the night, someone has to go to the mountain, of course, but even that is rare these days as the staff is normally able to solve problems remotely. Weather monitoring is also an issue: e.g. rain or snow detection.

Activities have been concentrated on determining the requirements to reach this goal. Thirty systems that will be impacted have been identified. Operating scenarios have also been identified with the aim of trying to identify different situations that could occur during a night and how they can be handled. The next step will be to determine how to implement the necessary changes in terms of manpower, materials and time. Initially, testing of remote observing would be carried out from the summit.

SAC Recommendation #6

Telescope automation SAC requests for the next meeting a report on the telescope automation activities being carried out by the staff. This should include a cost-benefit analysis and a description of the benefits to be gained by implementing the telescope automation.

Back to top

3. CFHT -- LS

Progress report

The steering group chair, Hervé Aussel, presented the status report of the Legacy Survey for the past 6 months. J. -C. Cuillandre has been elected to replace him for next year, starting after the May Users/SAC meeting.

The report included science highlights for each component of the survey. It is clear that the number of publications being produced from survey data is increasing rapidly and some papers have a very high impact. For example, for the SNLS, the Astier et al. (2006) paper has received 352 citations, for the Wide component, the Hoekstra et al. (2006) paper received 62 citations and for the Deep component the Ilbert et al. (2006) paper received 29 citations so far. For the Very Wide, interesting results were presented at the Users meeting and are described in a forthcoming paper (Kavelaars et al. 2007). Also, there are indications that new research projects are being carried out based on the CFHT-LS data, confirming the legacy value of the survey. The scientific value of the survey is widely acknowledged worldwide.

Back to top


Concerning the status of the observations, the SNLS and Deep components are on schedule to complete the observations by the end of 2008A, as planned. The SNLS team predicts that by the end of the survey, they will have spectroscopically confirmed 500 new type Ia SNe. However, for the Wide component the planned observations will not be entirely completed. The Wide component will meet its goal (set by SAC at the mid-term review) of a full coverage of the W1, W2, W3 and W4 fields in the g, r/2 and i bands by the end of semester 2007B. The planned LS observations in semesters 07A, 07B and 08A should allow the team to complete the u/2 and z/2 observations of fields W2, W3 and W4 but they will need an extension to complete the W1 field in 2008B in these bands. The SG is requesting 50 hours during this semester to do this. The need for such an extension had already been mentioned in the previous report from the SG.

SAC was also made aware that even after the initial validation by CFHT, a small fraction of observations were subsequently rejected by Terapix as non-compatible with the data quality standards set originally for the survey. This generates some missing data, which need to be re-observed. The steering group estimates that they will need approximately 15 hours in semester 08A (which is part of their normal allocation) and 15 hours in 08B (which they are asking in addition to their request for an additional 50 hours) to re-observe these fields. CFHT staff told SAC that they are not informed of the reasons of these data rejections.

SAC Recommendation #7

CFHTLS Re-observations of fields SAC is surprised that some LS images have been rejected by Terapix after being validated by CFHT without any feedback to CFHT. We recommend that Terapix inform the steering group and CFHT of the reasons for rejection. The steering group should justify to SAC why the data cannot be used for their science goals. Without such a justification, the fields should not be re-observed.

Back to top

The justification of the request of an extra 50 hours of observations in 2008B is mainly to preserve the legacy value of the Wide component of the LS data. It is important that the dataset remains homogenous in quality and in depth to ensure its usefulness for future science projects.

The SG also makes the case that if the full CFHT—LS time allocation were granted in 2008B, it would be possible to reach the full band coverage in z in the W1, W2 and W4 fields. To do this for the W3 field would require and additional extension of 30-40 hours in 2009A. However, the scientific gain of completing the full band z observations are not clear. Only very marginal improvements would be gained on the accuracy of photometric redshifts.

SAC Recommendation #8 -- CFHTLS Wide extension

In order to preserve the legacy value of the Wide component of the Legacy Survey, SAC recommends that the 50-hour extension into 08B be granted to complete the coverage of the W1 field. We do not recommend that the observations of the second epoch of z be carried out. Up to an additional 15 hours in 08B should be granted to re-observe the fields not validated by Terapix if a sufficiently strong case is made to the SAC that the existing data are unsuitable for the LS science goals. This recommendation also applies to the 15 hours that are planned to be used in 08A to re-observe fields not validated by Terapix.

Back to top

The steering group is also requesting an extension for the Very Wide component of the survey to allow proper motion measurements to be made. Only one square degree of the LS—VW is appropriate for such 3rd epoch measurements. Such proper motion measurements would allow the detection of ~50 new halo white dwarfs. The goal would be to assess the contribution of compact stellar objects to the dark matter halo of galaxies. More specifically, they are requesting 16 hours in 08A and 36 hours in 08B. The SG argues that this would increase the Legacy value of the VW data.

SAC Recommendation #9 -- CFHTLS Very Wide extension

Because the science being proposed was not in the original allocation of the LS and was not supported by the mid-term review, SAC does not support the proposed Very Wide observations within the Legacy Survey.

SAC would like to take this opportunity to remind the SG that their mandate should extend to the end of the final data release. More specifically, this corresponds roughly to one year after the end of the survey (either 08A or 08B).

Back to top

Terapix activities

Concerning the T0004 data release that should have been made in February 2007, Terapix has informed SAC that it has been delayed to May 20, 2007 due mainly to the loss of 1.4 FTE during the last year.

SAC was also informed that due to recent improvements in the Elixir flat fielding and fringe corrections, all Megacam data will be reprocessed starting in June 2007. This implies that the data will need to be re-processed and re-graded by Terapix. Terapix is studying a solution in which the Quality Fits for step1 would be run at CFHT and the images graded by CFHT.

Back to top


A report of CADC activities associated with CFHT was submitted to SAC. In the era of an increasingly large fraction of queue observing, CADC is playing a very important role of intermediary between the users and the facility. Nearly 100 % of CFHT data are archived at CADC. For the LS, CADC is the sole distributor of CFHT—LS imaging data, including Terapix stacks. The CFHT—LS imaging collection is quite popular. The ultimate goal is to fully automate data transfer, handling and distribution, which would minimize personnel costs. The main emphasis in the past 18 months was therefore aimed at improving the Terapix-CADC data flow.

Back to top

CFHT / CADC / Terapix communications

In view of recurrent communication problems between CFHT, Terapix and CADC that could impede on the proper delivery of data releases, steps have been taken to alleviate this problem. An oversight group has been formed composed of the CFHT director, C. Pritchet and H. Aussel. The mandate of this group is to provide advice and if necessary arbitration between the different groups regarding CFHT—LS data distribution. Videoconferences have been taking place every two weeks and seem to have greatly improved the situation.

SAC is pleased that these measures have been taken and encourages the three entities to continue in this vein until the final data products are released.

Back to top

4. Telescope Use and Scheduling

Joint Canada-France proposals

The members of the SAC-TAC from Canada and France met at the end of the first day of the meeting to discuss joint Canada-France proposals. The informal discussions are viewed as a very positive and extremely useful exercise and it is planned that these meeting will continue in future semesters.


Pierre Martin presented the QSO report. The second semester for which two instruments are being fully offered in queue mode (06B) has been the most successful run for Megacam since QSO has started.

The observing efficiency of Megacam is considered to be fully optimized. In spite of some bad, unpredictable weather and several time-critical observations, the A+B+C programs were completed at 94%. The fraction of time lost to weather and technical problems was as expected at 27% in 06B and the global validation rate is excellent at 92%.

The 06B run was less successful for WIRCam with one run lost to a technical problem with the cryogenic head and several nights lost to bad weather. Some engineering time was used to improve the guiding and to establish a focus model. Nevertheless, the observing efficiency of WIRCam is constantly improving and now reaches 80-85%. The fraction of time lost to weather and technical problems was extremely high at 46% and the global validation rate was 94%. The completion of A+B programs was 79%, which is good considering the large number of nights lost to technical problems.

Operational overheads for Megacam have reached a level below which it will be difficult to make further improvements. They are mainly dominated by filter changes. Dome rotation and guide-star acquisition still contribute but at a much smaller level. Focus sequences have been removed almost completely from these statistics. For WIRCam, the overheads are due to a two-step focus sequence, guiding acquisition and pointing corrections and finally to dithering patterns; they still account for about 25-30% of a night. A focus model has now been implemented saving ~ 30 minutes a night.

QSO staffing is now satisfactory and stable with the recent renewal of two experienced observers. The new Canadian resident astronomer, Daniel Devost, has arrived and will start being trained for queue coordination very soon.

Back to top

Balance of Agency time

A report presented to SAC by P. Martin examined the possibility of balancing agency time on a yearly timescale instead of a semester timescale. Presently, the balance is done during each semester as much as possible; if this is not possible, the situation is corrected during the next semester so in reality the balance is achieved on a yearly timescale. In fact, the experience of the staff is that if a strong unbalance is generated, it could be difficult and even impossible to correct due to unpredictable weather, so it is best avoided. For A programs, there is generally no problem because they are similar in number between the C and F agencies; B programs can suffer due to uneven pressure on a given RA.

The conclusion of the report is that it is likely that, for general PI observing, removing the semester-by-semester balancing requirement would create more problems than it would solve.

However the conclusion reached by the QSO team for the CFHT – LS is different. In order to ensure that the survey is finished by the end of its assigned date, it is suggested that priority be given within QSO to the CFHT – LS with respect to C and F PI programs, starting in 2008A.

SAC Recommendation #10 -- CFHTLS completion

To ensure the completion of the survey, SAC recommends that priority be given to LS over Canadian and French PI programs in 08A and 08B.

Back to top

5. CFHT'S Golden Age

No updated metrics and goals were presented at this meeting; this will be done during the next fall meeting of SAC.

6. CFHT 8th User's Meeting - May 2007

Long-term options 2008-2012

During the 8th CFHT user’s meeting, several possible Large Programs using the three main instruments in usage at CFHT were presented. There were many interesting ideas covering a large range of scientific subjects such as Kuiper belt objects, brown dwarf and planet transits, stellar magnetism, baryon oscillation and the nature of dark energy. During the meeting, a discussion also took place on the near and long-term future of CFHT. Those present clearly expressed their interest for the existence of Large Programs in the near future (2008-2012). However, there is also within the community a will to reserve a large fraction of observing time for PI projects.

Interest was also expressed for programs that are not necessarily large but can span over more than one semester. Based on this and on input from the TACs, SAC believes that some research project can be best served by this type of program but that it needs to remain a small fraction of the observing time that is allocated in this manner so that future TACs do not have too little time to attribute.

SAC Recommendation #11 – Small multi-year/long-term programs

SAC believes that the implementation of small multi-year/long-term programs would be beneficial to the science mission of CFHT. We recommend that each agency allow its TAC to commit telescope time over more than one semester. We suggest that in any given semester, the TAC allocates at most one night per semester, up to 12B, with a total cumulative allocation not exceeding 4 nights per agency per semester. Only programs receiving top evaluations (A queue) should be granted multi-year/long-term status. Multi-year/long-term programs approved in the B or C queue will be granted time only for the current semester - allocations over future semesters must be approved by that semester's TAC. Multi-year/long-term programs do not need to ask for allocation each semester to qualify.

Back to top

SAC Recommendation #12 – Large Programs

SAC recommends that participating agencies reserve 40% of the observing time for Large Programs (of the order of 600 hours a year) to be carried out over several semesters with any of the three main instruments. We suggest the following:

· Definition:

A Large program is a project that requires a minimum of 300 hours in total and does not extend beyond 12B.

· Frequency of calls:

The call for proposals should be made once a year until the fraction of time reserved for Large Programs has been allocated.

· Proposal details:

· The call for proposal should encourage Large Programs to be collaborative efforts between all participating agencies.

· The proposal should include a rough description of tasks to be carried out by the members of the team.

· An observing strategy and data reduction and management plans should be provided and the proposers should consult CFHT and CADC staff if necessary to help in the preparation of the project.

· Immediate access to pre-processed data by the participating communities should be mandatory.

· Process:

A) A call for letters of intent with a deadline of 15th of October 2007 should be issued requesting a full list of collaborators and up to 1 page of scientific description.

B) At the fall meeting, SAC will select and recommend to the Board experts to be members of the large-program TAC (LPTAC). The selection will be driven by the proposed research as described in the letters of intent.

C) A first call for proposals should be issued with a deadline of 1st February 2008.

D) Soon after, external referees (at least 3 per proposals) should be selected by the LPTAC.

E) Proposed projects are discussed and critiqued by participating TACs during their regular spring meeting. TACs should not rank the proposals but should provide feedback to the LPTAC.

F) The LPTAC meets and selects the projects for a given call. A technical evaluation carried out by CFHT staff should be part of this process. The final selection should include a reasonable balance of agency time. If less than the pre-allocated 40% of time is allocated to Large Programs, the remaining observing time will be used to execute regular PI programs, as ranked by the national TACs.

· Constitution of the LPTAC

The LPTAC should consist of: ITAC + 2 SAC members (ideally Chair and Vice Chair) + CFHT Director (as a non-voting member) + experts from outside based on letters of intent (selected by Board with advice from SAC).

Back to top

Beyond 2012

During the 8th Users Meeting, there were several presentations on possible instruments for the longer-term future of CFHT even though there is no formal process yet for proposing future instruments. SAC views this as very positive. There were several interesting suggestions such as using new IR detectors, an Ultracam-type instrument, an IR spectropolarimeter, a tachometer (network of 4m telescopes) and of course OHANA and VASAO.

Concerning the long-term future of the CFHT after 2012, the SAC has the following views:

· To be realistic we think that it would not be effective to have an observatory that is run like a service facility with a large suite of instruments that can serve a wide range of needs.

· In the context in which C, F and H remain the main shareholders of the corporation we also believe that a single instrument telescope would be unlikely to serve the needs of the communities.

· We believe that concentrating on a small number (between 2 and 4) of unique instruments on the telescope is the best way of maximizing the scientific return of the telescope. These instruments should ideally take advantage of CFHT’s unique site.

· We believe that there should be at least one new instrument.

SAC Recommendation #13 – Feasibility studies for future instruments

SAC recommends that a call for proposals for feasibility studies of new instruments be issued with a deadline of 1st October 2007. The call should specify the amount of money that CFHT is willing to contribute towards feasibility studies and an estimate of the maximum amount of money that the agencies are prepared to contribute to these instruments. The proposal should include a complete science case. At its November meeting SAC will select a small number of proposals for feasibility studies.

Back to top

Next meeting

The next SAC meeting will be held in Waimea either on 6, 7 and 8 November 2007.

Back to top