Time Allocation Committees (TAC) processes

Overview

Gemini Time Allocation proceeds in two phases, outlined here. Important dates for the upcoming semester are given in each Call for Proposals. This page describes the National and International Time Allocation Committee (TAC) process.

After the Call for Proposals deadline at the end of March or September, each National Gemini Office (NGO) collects the proposals from its community, resolves logistical issues and errors, and evaluates them for technical feasibility. Each Participant then peer reviews its proposals via a National TAC (NTAC), and creates a ranked list based primarily on scientific merit, but also on strategic feasibility. NTAC meetings generally occur 4-6 weeks after the proposal deadline (late Oct/early Nov or late April/early May).

The outcomes of the NTAC meetings are ranked lists of approved programs and recommended time allocations to fill each Participant's time allocation, for consideration by the International TAC (ITAC). These approved programs are forwarded to Gemini via an NTAC package, due ~6 weeks after the proposal deadline each semester (generally 10-13 May or 10-13 November). 

The Gemini International TAC merges the proposals to create a single combined list of programs for execution. The ITAC only sees those proposals that are forwarded by each NTAC. Prior to the ITAC meeting in the third or fourth week of May or November, the ITAC chair merges all of the forwarded programs from each Participant to produce a draft queue. The merging process is a deterministic process that steps through each Participant's ranked program listing and distributes the time according to each Participant's share of the total time. The draft queue is shared with the ITAC members and with the Gemini Heads of Science Operations who construct a preliminary telescope schedule (incorporating engineering and commissioning work with classical time requests and other considerations) and communicate any issues or concerns for operations.

During the ITAC meeting, the ITAC representatives analyze the draft queue in detail; they consider instrument availability, Target of Opportunity (ToO) programs, duplicate observations and the distribution of programs across different observing condition constraints, and adjust the ensemble of proposals to construct the best "queue", both for the individual communities and for the observatory. Duplicate observations are defined as observations of the same target with the same instrument configuration.  When such conflicts arise, the higher ranked program retains the observation and the target is removed from the lower ranked program. The PI of the lower ranked program may replace the target with an alternative. 

The outcome of the ITAC meeting is a recommended queue list (including scientific ranking bands), and classical program list for execution in the coming semester. This recommendation is forwarded to the Gemini Director for detailed consideration. The final allocation is at the Director's discretion and changes can be made, although these are generally minor in nature. The final approved list of programs is shared with the NGOs and ITAC members, and "NGO feedback" is generated which includes relevant comments intended for the Principal Investigators for proposals that were forwarded to ITAC.

Within a few days of sending the NGO feedback, the observatory notifies all successful PIs directly by email with information about their time allocation and how to start the Phase II process. At the same time, all the approved programs are published on the Gemini website and the "Phase II skeletons" are made available in the Observing Database. This generally occurs around 15 June or 15 December.

Technical Assessments of Proposals

This page is intended to be a brief guide for staff at the National Gemini Offices (NGOs) when they are technically assessing proposals at the Phase 1 stage. Additional information is given on each instrument's page (in the "Observation Preparation" area); these are designed to assist the Investigator when writing their proposal. The proposal should contain enough information to allow the assessor to judge whether the amount of time requested, with the instrument configuration and observing constraints requested, will allow the scientific goals to be achieved.

• Target Coordinates

  Is the RA, dec for each science target available this semester and also for the instrument this semester (see the accessibility information in the semester's Call for Proposals.

• Targets of Opportunity

  Target of Opportunity proposals must be identified as such in the Phase 1 Tool (PIT) xml file, and also whether it is rapid or standard. If the schedule for triggers is known please note this, for optimum filling of the queue. All proposals for Rapid Target of Opportunity (RToO) followup must submit a separate proposal for Standard Target of Opportunity followup (SToO) in conditions better than SB/CC/IQ=Any, if such followup is planned. Upgrades to good conditions will not be approved for RToO programs, and the SToO proposal is required if such conditions are necessary for later followup. Starting in 2012B, time in ToO programs that is not triggered will be charged to the partner (not the program) at the 50% level. Therefore time allocations to ToO programs should not be overly optimistic.

• Instrument Configurations

  Is the instrument configuration appropriate for the science? For example:

  • Are the components (e.g. filter, grating) available?
  • Does the grating and slit combination provide the appropriate resolution?
  • Is the camera Field of View appropriate?
  • Has the correct readout mode been chosen?
  • Also, if this is an LGS observation then the Altair component must be selected in the PIT xml.
• Instrument Overheads

  Are there additional overheads that need to be included? Such as:

  • GMOS MOS pre-imaging;
  • Separate skies for infrared observations;
  • Readout overheads e.g. GMOS nod & shuffle, thermal IR short exposures; see the "ITC, Sensitivity and Overheads" section in each of the instrument pages;
  • Non-standard calibrations, see the "Calibrations" section in each of the instrument pages.
• Guide Stars

  Are appropriate guide stars available? This is especially important for AO work, for the small-field on-instrument WFS of the infrared instruments, and at high galactic latitudes or for example the Hubble Deep Field where guide stars will be scarce. Programs that use the P2 or P1 probe arms or the GMOS OI have a large field available and checking these guide stars at Phase 1 is not usually important. For more information, see the "Guiding Options" section in each of the instrument pages.

• Exposure Time and Signal to Noise

  Are target flux levels given? If not, note this and that an assessment cannot be made. If fluxes are given, does the PI state that the exposure times have been confirmed using the Integration Time Calculator (ITC), for the selected observing conditions? If not, check the exposure times look sensible using the sensitivity tables given in the "ITC, Sensitivity and Overheads" section in each of the instrument pages.

• Observing Conditions

  It is essential that the observing constraints are correct for achieving the science and for properly filling the queue.

  • LGS and GeMS programs require IQ70 CC50 or better, and dark or grey skies if the guide stars are faint (some GSAOI+GeMS observations can be carried out in IQ85 with degraded performance);
  • NGS-AO usually requires IQ70.
  • Thermal infrared usually requires WV50/80 and SBany;
  • optical observations usually require WVany and SB50/80.
  If any of these are not selected then check the proposal and correct as necessary. If CC80/any or IQany have been selected check that this is realistic and not an error. Note in your assessment any unusual observing constraints so that the TAC knows these have been confirmed.
• Timing Constraints, Non-Standard Observations

  Please note in your assessment whether the observations are time-critical or have timing constraints. Also note if there is anything unusual about the planned observations, for example an infrequently used instrument mode, or unusual acquisition and guiding.

• Classical Programs

  If this is a classical program that requires good observing conditions is there a backup program for poorer weather? If so is it feasible? If not please note that a backup plan is needed. If there is no backup program then the time will be used for queue observations if the conditions for the classical program are not met.

NTAC Proposal Evaluations

Similar to other time allocation committees, the NTACs must evaluate their communities' proposals for scientific merit and technical feasibility. However, because the overwhelming majority of Gemini observations are queue scheduled, the NTACs should also consider "strategic feasibility" of the forwarded proposals, which is a strong function of the proposals' ranking. Here we list some important issues that the time allocation committees need to consider when determining a proposal's feasibility.

• The Call for Proposals
• Keck or Subaru Exchange Time
• Classical-Mode Gemini Proposals
• Science Ranking of Queue Proposals
• Joint Proposals
• Target of Opportunity Proposals
• RA Distributions
• Observing Constraints
• Poor Weather Programs

• The Call for Proposals

  Consider the guidelines and restrictions provided in the Gemini Call for Proposals. These include restrictions on target accessibility and instrument availability.

• Keck or Subaru Exchange Time

  Subaru or Keck classical proposals must have a total request of integer nights or 10 hours. For Keck and Subaru there are restrictions on sky brightness and possibly on RA that are described in the Call for Proposals ; note that no dark time is available on Keck. Keck proposers are required to complete the Keck cover page which should be emailed by the PI to their NTAC. Please forward all Keck and Subaru proposals, as pdfs, to the Gemini ITAC Chair for technical assessment at these observatories.

• Classical-Mode Gemini Proposals

  The Gemini classical nights recommended by each Participant will be removed from the Participant's queue allocation before the queue merge is carried out, and so these should be highly ranked programs. Proposals requiring unusually good observing conditions (e.g. seeing better than 0.4 arcseconds) or containing targets with widely distributed RAs, should be executed in the queue and not as classical nights. The total time request must be integer nights or 10 hours. A backup program that can be done in poor conditions should be specified; if the conditions specified are rare and no backup is specified it is likely that the allocated time will revert to queue.

• Science Ranking of Queue Proposals

  There are three science bands defined for queue programs, where band 1 is highest priority. Each Participant's queue time will be filled to 80% with band 1, 2 and 3 programs. A cloud cover constraint of CC80 has been defined, corresponding to one magnitude of extinction, or 40% transmission. The remaining worst-conditions time will be used by band 4 poor weather proposals, and this time is not charged to the Participants. The band boundaries will be drawn at 30%, 60% and 80% of the total available queue time. The available queue time is the advertised time on each telescope, less the time required for classical programs, and less the time estimated to execute rollover programs accessible in the semester. The rollover time is typically 5% of the advertised time.

  When estimating where in your NTAC-ranked queue-program list the band 1/2 and 2/3 divisions will fall, do not assume that these will be 30% and 60% of the advertised Participant time. First subtract your requested classical time from your allocation, and then estimate the band 1/2 and 2/3 divisions at the conservative side of the 28% and 57% time boundaries, to allow for rollover time and the vagaries of filling the queue. If you have a program that must fall in band 1 make sure it is in the top ~25% of your available queue time, and if you have a program that cannot be in band 3 make sure it falls in the top ~55%. Rapid Target of Opportunity proposals must fall in band 1, and LGS programs must lie in bands 1 or 2.

  Band 3 programs need special attention and programs that fall below 55% of your allocation may fall in band 3. First check that the PI has not selected "Do not schedule in band 3". If they have but you think they can produce useful science and will accept band 3 time, then edit the proposal to reflect this, otherwise promote or drop the program. Next, examine proposals that have tight observing or timing constraints, those that use uncommon GMOS configurations, or those that have RAs historically in heavy demand. Such programs are very unlikely to get data and should be promoted or dropped. Technical assessments must be done on both the main program as well as the modified program described in the band 3 tab.

• Joint Proposals

  Joint proposals are evaluated by each NTAC to which the proposal was submitted (each "joint component"). The final ranking of the joint proposal in the merged queue is a time-weighted average of the individual rankings (that is, a normalized average of each Participant's rank scaled by the time allocated by that Participant). The final disposition of joint proposals is not decided until the ITAC, when each Participant learns the disposition of each of the other joint components. NTACs should take care to provide instruction to their ITAC representative how the joint proposal should be handled based on what the other NTACs have decided. For example, if a joint proposal is not supported by one or more other Participants, should the ITAC member drop support, or add time to reach a certain minimum level? The "minimum useful time" (indicated in the PIT) can be very useful in making these decisions as the sum of the "minimum useful time" from each component should equal the minimum time required to execute the program. NTACs must forward to ITAC all Joint proposals whether recommended time or not, so that all Joints are recognized as such and properly considered.

• Target of Opportunity Proposals

  ALL proposals that do not have pre-defined targets (i.e., unknown coordinates) MUST have one of the "Target of Opportunity" boxes checked in the PIT XML file. Target of Opportunity (ToO) programs are defined as "Standard" or "Rapid" trigger, depending on the necessary response time. All ToO programs are reviewed during the ITAC process to check for duplication or conflict.

  Rapid ToOs (immediate or less than 24 hour response time) programs must be ranked in band 1 to be recommended at the ITAC and so should be in top 25% of your rankings. The most common type of rapid ToO is Gamma Ray Burst follow-up observations. Only one GRB program can be active at any given time, and only one, or possibly two, such programs per telescope will be recommended by the ITAC. (If it is two, the Head of Science Operations for that telescope will work out an "activation schedule" with the two teams involved.)

  Standard response ToO programs (response time > 24 hours with no guaranteed upper limit) may be in any band, however their ToO status does not afford them any special priority. That is, a band 3 ToO program will not be executed before band 1 or 2 just because it activates a target. For a highly ranked program, a typical standard response time may be 2-3 days, however a program in band 3 could have a much longer response time, or not be observed at all. ToO transients requiring followup on timescales shorter than a month are not appropriate for band 3. Ensure that all band 3 ToO programs are truly viable in this low-priority band.

  All proposals for Rapid Target of Opportunity (RToO) followup must submit a separate proposal for Standard Target of Opportunity followup (SToO) in conditions better than SB/CC/IQ=Any, if such followup is planned. Upgrades to good conditions will not be approved for RToO programs, and the SToO proposal is required if such conditions are necessary for later followup. This change is necessary for accurate filling of the queue, as ToO programs now make up a significant fraction of the Observatory band 1 time. See the Target of Opportunity (ToO) web page for further information.

  Do not overallocate time to ToO programs. Starting in Semester 2012B, time in ToO programs that is not triggered will be charged to the Participant (not the program) at the 50% level. The 50% value was determined from the long-term average completion rate (and hence the time charging rate) for non-ToO programmes. That is, the 50% charge for unused time makes the overall charging rate for ToO programmes match that for non-ToO programmes. These rates will be periodically reviewed. This change has been implemented to ensure fairness across the partnership and avoid inadvertent "time banking". The charging was agreed to by the Operations Working Group in February 2012 (Resolution 22.8). The Observatory does not wish to dissuade TACs from awarding time to ToO programs, but the time allocated must be realistic. Some Participants are accumulating unused time via unrealistic allocations. When Participant time balancing is taken into account at the Call for Proposals, all other Participants then get their time significantly reduced, hence a mechanism such as this is required to reduce the hit to the other Participants. In 2011B more than 70 hours of Band 1 ToO time, and more than 60 hours of Band 2 ToO time, was not used.

• RA Distributions

  The RA distribution of the program set forwarded to ITAC should approximately follow the availability for the semester. Historically in the North in the A semester the RA range 10 - 14 hours has been oversubscribed, and in the B semester RA~4 hours can be a problem. We cannot complete programs if the the RA distribution does not track what is available.

• Observing Constraints

  It is essential that the observing constraints are correct for properly filling the queue. LGS and GeMS programs require IQ70 CC50 or better, and dark or grey skies if the guide stars are faint; NGS-AO usually requires IQ70. Some GSAOI+GeMS observations can be carried out in IQ85 with degraded performance. Thermal IR usually requires WV50/80 and SBany; optical usually requires WVany and SB50/80. If these are not selected then check and correct as necessary. If CC80/any or IQany have been selected check that this is realistic and not an error.

  Forward to ITAC a set of programs that match the occurrence of the various constraints, as far as possible. That is, no more than 50% should require CC50 etc. CC50 is the constraint that is usually filled first so lower ranked programs should require CC70 or worse.

• Poor Weather Programs

  Poor weather programs can be submitted either via the regular 6-monthly Call, or, at any time, directly to the Observatory using the Phase I Tool. If submitted via the regular Call, the TACs must rank the program in the normal way. If the program ends up as a science program in bands 1-3 it will be executed before any Poor Weather queue programs. If submitted directly to the Observatory, and approved, the program only exists in the Poor Weather queue. There are very few programs that can be done in poor weather, and so the chances of getting data from the Poor Weather queue are quite good. If a program qualifies as a poor weather program (and is to be considered as one), the "Poor Weather Candidate" flag must be checked in the "TAC" section of the Phase I Tool (PIT) XML file. If the program is ranked high enough to achieve band 1, 2 or 3, it will be included in the queue as a regular program, otherwise, if it has the "Poor Weather Candidate" flag set, it will be added to the poor weather queue (i.e., band 4). No time is deducted from the Participant's allocation for Poor Weather programs and there is currently no limit to how many poor weather programs can be submitted by the NTACs, regardless of Participant share. This means that poor weather programs can overfill a Participant's time allocation, at the bottom of the Participant's ranked list, but they should still be ranked (relative to each other) in case not all of the poor weather programs are accepted. The "poor weather queue" is reviewed during the ITAC meeting and modified if necessary.

Assembly of the Queue and Classical Schedule 

The draft queue and classical schedule are assembled from the proposals that have been recommended for time by the NTACs. One important concept in bringing together these ranked lists of proposals is that of the merging sequence of which an example is:

The merging sequence, of which a recent example is shown, is a sequential list of all the Gemini participants. The frequency with which each participant occurs in the seqeunce is fixed to the nominal time allocations. The participants's recommended classical time is removed from their total allocation before the queue merge is carried out. Any minor advantage in being the participant with 'first pick' (Canada in this case) is reduced by rotating the starting point amongst the partners in subsequent semesters.

As of 2014B, the merging sequence includes observing time for Large and Long Programs, effectively as an additional participant.

The merging sequence is cycled through as many time as necessary to use every participant's available queue time. The available time in each observing constraint (e.g. dark time, SB50) is also checked and reduced on each allocation as appropriate; when the time in each bin is used no more programs requiring that condition are merged. Currently, we allow the poorer conditions to be overfilled, but the best image quality, dark and dry time, and photometric time, are not allowed to be overfilled. If a program cannot be scheduled it is skipped.

A simplified worked example of the first few steps of an early version of the merging process can be found in Puxley & Boroson (SPIE...1996; Gemini Preprint #13).

ITAC Terms of Reference

The terms of reference for the International Time Allocation Committee (ITAC) are:

• The ITAC consists of one representative from each National Time Allocation Committee (NTAC) and two Gemini staff scientists (the current ITAC membership is shown on the governance page).
• The ITAC will advise the Director on resolving any conflicts that arise during assembly of the draft queue and classical schedule.
• The ITAC will recommend the cut for the overfilled, merged queue and the position of the scientific ranking bands.