QSO Phase1 Megacam

The main concept behind the queue observation scheme with MegaPrime is to perform observation programs only during sky conditions required to meet their science goals, as defined by the investigators. This can only be achieved if the programs are all grouped together in a database and are selected appropriately according to a set of constraints, rules and sky conditions. Programs are then carried out by a well trained, local team of observers in a service mode (i.e. investigators are not present at the observatory).

During 1999, CFHT has started a project to implement the necessary software and to review all the issues for achieving a queue/service observing mode with its CFH12K mosaic camera. This Queued Service Observations (QSO) Project has been developed in parallel to other projects necessary for the data acquisition (NEO), processing and analysis (Elixir), and archiving and distribution (DADS). The necessary software tools for proposal submission (Phase 2), selection of programs, management of the observations and execution of the observations have all been developed within the QSO Project. Most of these software components are for internal use only except for two obvious exceptions: Poopsy, proposals submission tool developed and maintained at CADC, and PH2, a Web based tool developed and maintained by CFHT for the second tier of proposal submission (see below).

The QSO mode has been used for CFH12K for over 200 nights between January 2001 and January 2003. Since the semester 2003A, MegaCam has been operated in the queue mode; for instance, the semester 2006B has been very succesful with 94% of all of the programs completed and validated! Improvements made on some sub-systems of MegaCam have steadily increased our observing efficiency. This tutorial was written as a guide for helping in the preparation of the proposal submission for the QSO mode with MegaPrime.

NOTE: For all technical information on MegaPrime/MegaCam, please refer to this page.

This document presents the information for submitting a QSO proposal with MegaPrime for the semester 2007B. A complete description of the submission process with Poopsy and an outline of what will have to be done following the TAC evaluation for the second phase, planned for December - January, 2007-2008, are included. A few QSO Rules used for the selection of the programs are presented and some other issues related to the QSO programs are also discussed.

IMPORTANT NOTE : At the time of the writing, the non-sidereal guiding option with QSO MegaPrime has not been implemented. Due to other urgent issues at CFHT, it is unlikely that this mode will be commissioned before some time so the mode is not offered for 2007B. Sorry!!

For more information about the submission of your MegaPrime QSO proposals, contact the QSO Team.

The Scientific Advisory Committee (SAC) has recommended that, starting at the semester 2003A, the QSO mode should be the ONLY mode of operation for the MegaPrime mosaic camera. So, the classical mode is NOT offered for MegaPrime for 2007B. There are several reasons behind this decision, including the severe time constraints included in the components of the CFHTLS, and the difficulty in the implementation of different software components to cover the classical mode.

Note: Observation of Solar System targets through the QSO system is possible through ephemeris entries in the Phase 2 Tool. However, as mentioned above, differential tracking with MegaPrime (i.e. non-sidereal guiding) is a complex implementation and is not offered for 2007B. Please consult the QSO Team for further information about the progress done on this issue.

Many types of programs can benefit from the queue observing mode. Programs requesting excellent or exceptional sky conditions, surveys, short- or long-term monitoring, target-of-opportunity programs, are all well suited for the QSO mode. Also, contrary to the classical mode, it is now possible to submit very short programs necessitating only a few hours of observations. Programs scientifically valuable during bad sky conditions are also possible. During the submission phase with Poopsy (Phase 1), you are asked to specify what type of programs you are submitting for the QSO mode. These types of programs are defined below:

Regular: This category includes all of the normal QSO programs, with fixed or Solar System objects. For all these programs, the Phase 2 must be completed before a specific deadline prior to the semester with QSO periods. Only investigators with Solar System targets or other restrictive constraints might be allowed to modify their observations during the observing semester.

Target-of-opportunity (TOO): For this type of programs, the targets are not known before the beginning of the semester. In fact, the TOO programs are divided into two distinct categories: 1) Programs approved by TAC during the evaluation of all the proposals but for which the targets will be identified during the semester (e.g. gamma-ray bursts). For these programs, the investigators will have the possibility to define their targets using the appropriate tools developed for submission of the queue observations during the semester. 2) Programs submitted during a semester to observe unpredictable objects (e.g. supernova). In that case, the program must first be submitted through a special Web form (available on the QSO page) and approved according to the CFHT policy. If accepted, a certain amount of time is allocated and the information, entered during the Web submission, is used by the QSO Team to prepare and carry out the observations.

Snapshot Programs: Since 2002A, we have changed the definition of a QSO "snapshot program". In a queue mode, it is very important to have stored programs requesting bad sky conditions and low completion level. The definition of a snapshot program is the following: 1) A snapshot program describes valuable science to be obtained on targets observed only in the worse sky conditions (seeing larger than 1.2" (in r band) and preferably during non-photometric conditions); 2) A snapshot program preferable request observations with the "standard" ugriz filter set; 3) A snapshot program should preferably include short blocks of observations; 4) The time allocated for such programs is not accounted for in the statistics of time spent for the different Agencies; 5) The proprietary time for the data is 3 months following the end of the semester. Note that the chance of such programs to be executed, even partially, during a semester can be quite high.

For MegaPrime programs requesting the queued service observing mode, two submission phases are necessary. The first phase (Phase 1) is done through CADC Poopsy and consists in a general description of the program used for the evaluation by the Time Allocation Committee (TAC). This is the "normal" submission procedure for all proposals requesting time at CFHT. The second phase (Phase 2) is requested for ALL the MegaPrime queue programs for which telescope time has been allocated by the TAC. As described below, it is done prior to an observing semester (with a few exceptions) through an entirely new Web based tool developed at CFHT. During this phase, all the information necessary for the local staff to perform the observations is entered by the investigators and stored in a database at CFHT.

Phase 1: Poopsy

The first submission process for your queue proposal must be done through CADC's Poopsy. This Web interface has been updated with a special section allowing the submission of QSO programs. As described in the next section, it is very important to include as much information as possible now in your proposal. The Poopsy database, located at CADC, is copied at CFHT and the relevant information for queue programs is parsed to the CFHT queue database. For instance, if you enter all your targets in Poopsy, you will not have to do it again during the Phase 2!

Phase 2: PH2

The second phase aims at gathering all the information necessary for carrying out the observations for queue programs with allocated telescope time. This phase is done using a Web based interface, specially designed and developed for this purpose at CFHT. The tool is populating a relational database which constitutes the "brain" of all of the QSO operational processes. The Phase 2 for semester 2007B should take place in December-January 2007-2008, and will last several weeks. Investigators with programs requesting updated information during a semester (e.g. moving targets, target-of-opportunity) will be allowed access to PH2 during that semester. The Phase 2 for all the other regular programs will have to be completed by the deadline specified by CFHT. Complete information will be made available with the opening of the Phase 2 submission tool.

The PH2 tool is installed on the Web server at CFHT and is also available on a server at CDS for the French community. This architecture is important to avoid the traffic on the intercontinental link as much as possible.

The current version of Poopsy, now available for the proposals for 2007B, allows investigators to send proposals for the queue mode. A section has been created which includes some information necessary for TAC evaluation, and for the QSO Team in the preparation of the queue database. Below, we review the questions related to QSO proposals as introduced in this version of Poopsy.

As explained above, the queue mode is the only mode of observation for MegaPrime for 2007B. Answer "Yes" and complete the Queue Section.

2- Indicate the type of queue observing program: Regular, Target-of-Opportunity, Snapshot (Run Info section)

3- Indicate a global image quality (IQ) constraint (r band) describing your program ? (Run Info section)

The main concern of the queue mode is to observe targets under sky conditions required to meet the science goals defined for each program. During Phase 2, the investigators will have to define precisely observational constraints for their project. However, it is important that a global image quality for each program is defined during Phase 1 as well. An efficient queue can only be achieved if the database contains programs requesting a wide range in constraints, especially on the image quality, and the TAC will strongly consider the choice of image quality indicated here for the overall selection of the queue programs.

It is clear that the probability of a specific program depends strongly on the image quality requested versus the statistics of the site. The chance that your program is performed is less if you request an IQ of 0.55" than if 0.8" is required. Of course, science should be the priority in your selection of IQ but beware that it cannot be the only criterion in your choice. So, do not request a better image quality than what is really needed for your program. For guidance, the table below describe approximately the seeing statistics on Mauna Kea. These data are for the R-band and were taken with FOCAM at CFHT between 1993-1995 (more recent statistics confirm these values). Note that with MegaPrime we globally achieve about the same IQ distribution than with CFH12K, but not much better.

For your information, the table below gives the average weather statistics for Mauna Kea. Unfortunately, we have been hit really hard by weather in the first few semesters of MegaCam, in particular for the A semester when time lost during the first few months of the winter reached level of 50% and even more.

4- How many hours are required for this queue program for this semester ? (Run Info section)

A significant difference between the classical and queue modes is the way the observing time is accounted. In a classical mode, nights are usually requested and the total time asked includes the different operational overheads. In a queue mode, the time requested is in HOURS, and might or might not include overheads. For queue programs for the semester 2007B, you can follow these directives:

i)- You must request HOURS of observations. If the total time of your program is fractional (e.g. 32.4 hr.), please indicate so (.4 hr in a queue mode is possible).

ii)- In your calculation of integration time, ONLY the readout time for the CCD mosaic should be included. This is 40 seconds for the entire mosaic (no binning is offered) for each exposure. Slewing and acquisition of a guide star SHOULD NOT be accounted for in your calculations. Detailed instructions regarding the calibrations are given in the next section.

5- How many additional HOURS would be requested to complete this project ? (General Info section)

If you are planning to submit an additional proposal for the same program for the subsequent semester (and the following, if necessary), please indicate the number of hours that might be required.

Note: The sky brightness options presented in Poopsy (i.e. position of the Moon) are different than the ones be presented in PH2. For the Phase 1, these are used to help us evaluate the best periods for a scheduling a queue observing period covering as many programs as possible.

One of the main advantages of the queue mode scheme is the possibility to share calibrations between a set of programs. More so, since the queue runs are spread over about 15 -20 nights, the quality of the calibrations can also be greatly improved compared over the ones obtained during a short run in a classical mode. To achieve this, a calibration plan has been defined and will be carried out regularly by the queue/service observers. This plan includes the necessary "detrend" frames for removal of the instrument signatures (bias, darks, flat-fields, fringing) and the astronomical calibrations (standard stars, astrometric fields). More details can be found on the Elixir page.

1- No programs under any circumstances are allowed to request "detrend" calibrations during Phase 1 or Phase 2. These calibrations are exclusively handled by the QSO and Elixir Teams.

2- If your program includes the u*, g', r', i', z' filters , the astronomical calibrations will be automatically done during the QSO runs and distributed to you. You do not have to include these calibrations in your proposal justification, that is, the integration time requested should not include the time necessary for these calibrations.

3- If your program include any MegaPrime filters with a broad bandpass (u*g'r'i'z') and you prefer to obtain your own astronomical calibrations, this time must be added during the Phase 1. However, TAC will evaluate this request. Of course, the integration time will be automatically charged to the program for this kind of observations during the Phase 2.

4- Photometric calibrations for the narrow-band filters offered for 2007B (Ha,HaOFF, CN, TiO, OIII) MUST be included in the science program and will not be covered automatically by the QSO/Elixir calibration plan.

Following the submission of your queue proposal, the TAC will review it as in the past. However, the way the queue proposals are classified differ slightly from the ranking used for the proposals for the classical mode of observing. Following the TAC evaluation, your proposal, if accepted, will receive the following grades, ranks and allocated time. This information will be entered by us in the queue database and will be available through PH2 for the second submission phase.

* Most of all the snapshot programs will be accepted so a given percentage does not make sense

Important Note: Telescope time is scheduled for A + B programs. Programs below the normal evaluation cut-off are generally C and S programs and are considered for overfilling the queue database. There is, of course, a good chance for these programs to get some data especially if the conditions requested are realistic and take advantage of poor sky conditions. Final choice of C programs is left to the QSO Team, after revision of the actual constraints imposed by the A+B programs.
TAC Rank: The proposals will be quantitatively ranked with respect to all the proposals included within this group and within this Agency (e.g. 2/5,3/5, etc.).

Integration Time (I-Time): The TAC will allow a certain number of hours, the "integration time", for each approved program. During the Phase 2, this number constitutes an upper limit and is automatically taken into account by the software in the preparation of the observations.

An important component in the preparation of the proposals during Phase 1 and 2 of all the queue mode programs instituted around the world is the availability of a exposure time simulator. Since the investigators are not present during the observations, it is crucial for the QSO observers to know that if the observations are undertaken during sky conditions requested, the science goals should be reached because the proposers have verified that using a robust exposure time simulator. It is not always easy to judge the science merit of an exposure frame and this is better accessed by quantitative evaluation.

In preparation for the QSO observation, the investigators have now access to an elaborate exposure time simulator for the MegaPrime mosaic. Developed first for CFH12K under the supervision of Jean-Charles Cuillandre, the simulator, named DIET for Direct Imaging Exposure Time, allows now the user to evaluate the integration time in different sky conditions for MegaPrime. Moreover, the simulator offers different combinations of exposure time and number of exposures to minimize overheads. A Web based interface for the simulator is also available. For this Phase 1, several exposure time tables produced by the simulator are available here. It is strongly recommended to use the simulator during Phase 1 and Phase 2 of the queue programs with MegaPrime!

Maybe the most difficult task facing the queue observing model is found in the selection process leading to the execution of a science program. This selection algorithm can be based on simple criteria (e.g. mounted filters) but it becomes immensely complicated when other parameters like actual sky conditions, completeness level, science merit, monitoring constraints, or targets visibility are taken into account.

The process, resulting in the choice of a specific program to be undertaken for the queue observations with MegaPrime, will be done in three steps:

1- Selection: This is the first selection of the viable observations stored in the database according to instrumental constraints, sky constraints, actual sky conditions, completeness level, and the position of the targets.

2- Ordering: This second step creates a prioritized list of observations to be sequentially executed regarding their TAC grade, rank, target positions, and user's priorities.

3- Human filtering: The final step consists in the possibility for the QSO observer to modify the queue list according to special constraint like the focus sequences, calibration plan, etc.

Without going into too many details, each of these steps include an algorithm based on a set of observing rules. More information will be available during the release of the Phase 2 tool but some rules might be of interest already for the Phase 1 submission process. The rules given here are not presented in any order of priority. Among them:

As much as possible, images should not be obtained in worse IQ (or sky brightness) conditions than required. The IQ (sky brightness) measured should not exceed the upper limit of the IQ band required by more than 10-15%.
Images can be obtained in conditions better than required, if no other observations actually requesting these conditions are available.
For very bad IQ periods (>1.2"), the snapshot programs requesting bad conditions will be executed, unless other possibilities exist among the regular programs.
The priority of the programs started is automatically increased compared to programs not started.
When possible, the observations will be tentatively done with airmass smaller than 1.5.
During selection and ordering of the queues, the priority goes from grade A to B to C, followed by the snapshots.
Inside these grades, priority is given according to the TAC rank as much as possible.
Balancing Agency time has a very high priority and might exceed the other selection and ranking criteria.
As much as possible, the observer will execute the observations belonging to a given program according to the priority index (high, medium, low) given by the investigators during the Phase 2 period.
A QSO run should never be completed without getting all the necessary calibrations for all the programs fully or partially executed during the run.
All fields requesting photometry and done during non-photometric conditions will be calibrated with short exposures during photometric periods.
No programs will be recycled for completion during the next semester.
When started, a monitoring program receives a higher priority so that the observations to be repeated can be carried out within the specified timeframe period.
The QSO Team will always try to obtain the required number of observations for a given monitoring program. In case of other constraints, the minimum number of observations specified during Phase 2 is the minimum acceptable.
For target-of-opportunity programs submitted during the semester, these program will be subject to the same selection process and prioritization based on a grade and a rank as the regular programs.

The QSO Team has integrated a Email communication system inside PH2 ("HelpDesk") that allows the investigators to send request to the Team during the Phase2. This system is also available during the semester through PH2. It includes several discussion forums, and keeps tracks of all the communication related to the QSO mode for each program identified by its runID. Even if a bit cumbersome, using the Helpdesk is much preferable that using a normal email system. It ensures that everybody related to the observing process is informed,

After each observing night, a report detailing what observations were performed is available on the CFHT Web site. These reports include the observations groups executed and the sky conditions at the time of the observations. This does not mean that your data will be immediately available (see below). The goal of these reports is to inform the community of the progress of the queue and, in particular, the current status of your program.

Current global statistics for a semester are also available on the QSO Web page.

As part of the data quality control assessment, all data taken will be automatically processed and calibrated by the Elixir Team. Data evaluation will be done in two steps: during the observation by the Service Observer ("on-line" evaluation) and, during and after the data processing. This last step is very involving and represents one of the reasons why data cannot be distributed immediately after a QSO run, unless specifically justified during the Phase 2 period. If the observations are judged satisfactory, the queue database is then updated by the Queue coordinator.

Data distribution will be ensured by the DADS Team. Our goal is to be able to distribute the data to the PI of each project (or another member if specified during the Phase 2) and the relevant calibrations at the period specified during the phase 2. Due to the heavy workload, it will not be possible to send the data to the investigators during an observing run. However, for certain types of programs (e.g. TOO) where looking at the data as soon as possible is important, this will be possible under the supervision of the Queue coordinator.

A special procedure is taking place regarding the proprietary period for data obtained with MegaPrime in the QSO mode. By default, the proprietary period of QSO data extends to 1 year + 1 month starting at the end of the QSO semester. For instance, data taken for the 2007B semester (August 1st - January 31st) will have a default release date set to 02/29/2009. (Note: The CFHTLS is different). If an extension is requested in Poopsy and approved by TAC, a new date will be set for this program through the QSO system. The release date for the data is indicated in the fits headers by the keyword REL_DATE. This system applies only to QSO data for MegaPrime (and also CFH12K and WIRCam). For snapshot programs, the proprietary time is 3 months following the end of the semester.

The QSO Team is formed of Pierre Martin (QSO Project Scientist/Manager),Billy Mahoney (Database/System Programmer), Tom Vermeulen (System Programmer), and Mary Beth Laychak, Peter Forshay, Adam draginda, Rachael Zelman, and Todd Burdullis as Service Observers. During a QSO run, supervision is ensured by the QSO Coordinator (one of the CFHT Resident Astronomers) who, among other things, is responsible for managing the queue database, planning of the observations, and maintaining the contact with the investigators, if necessary. Observations will be conducted by the Service Observer and the Resident Astronomers, with a strong involvement by the Observing Assistants. Software support will also be provided during the observing nights. For TOO programs and decisions related to the viability of some programs, the CFHT Executive Director acts as the final authority.

Image Quality (IQ)	Frequency (%)
IQ 0.55"	5
0.55" < IQ 0.65"	25
0.65" < IQ 0.80"	30
0.80" < IQ 1.0"	25
1.0" < IQ 1.2"	15
IQ > 1.2"	5

Sky Conditions	Frequency
Usable Nights	~70%
Lost to Weather	~20-30%
Usable Photometric Nights	~50%

Grade	Grade Name	Global Priority	% of Telescope Time
A	"Must-Do"	Highest	30-35
B	"Prioritized"	Good	35-50
C	"Best Effort"	Medium	20-30
S	"Snapshot"	Lowest	*