Basic JWST operations concept
From the user point of view, JWST will be operated very similarly to Hubble. There will be a phase I call for proposals reviewed by a TAC. We are still discussing the balance between single phase approach like Spitzer and a phase I/phase II approach like Hubble. Our software should be able to support both if needed. There are advantages to single phase submission having mostly to do with shorter time elapsed between proposal submission and proposal ready for the execution. The two phases approach lowers the bar for proposing allowing people to focus on the science rather than implementation. Particularly for large programs the full phase II is a lot of work and it seems hard to commit to it with the high oversubscription that we expect for JWST proposals.
All overhead will be charged to observers, General Observers (GOs) and Guaranteed Time Observers (GTOs) alike. Some overheads are deterministic, i.e. known at the phase II level, others - such as the time to slew from the previous position - are statistical. The overheads will be revised yearly but old proposals will not be reassessed. We are counting on an accurate Science Operations Design Reference Mission (currently being revised) to determine the overheads as accurately as possible before launch.
We expect to have some balance of small, medium and large proposals, managed similarly to Hubble. The JWST Advisory Committee (public JSTAC pages) recommended that the default proprietary time for large proposals be set to zero with the option of requesting a proprietary time for scientific reasons (e.g. time monitoring of a source). Hubble allows Treasury programs. They are by norm large programs with zero proprietary time and addressing broad science issues with large legacy value. A non-proprietary large programs has a lot of overlap with a Treasury and there has been no decision regarding Treasury programs for JWST.
The GTO time is constrained within a minimum and maximum envelope. All GTO time will be concentrated in the first three years of the mission, with most of it front loaded during the first and second year. The time available for General Observers (GOs) will in any case exceed 50 per cent - or at least 4400 hours - even during the first year of operations. With our expected observing efficiency the time on the sky should be around 3000 hours. The available time will increase in the following years.
We are considering adopting similar duplication definitions as Hubble. The definition of duplication will be assessed on the basis of the proposal and not on its outcome. Two observations are duplicate if they observe the same target with similar spectral elements and similar signal to noise. Science goals are not used to define duplications. JWST instruments have largely non overlapping capabilities so we should not have significant duplications across instruments. We are thinking of defining similar signal to noise as that resulting from exposure times within a factor of - say - four. But the precise value has to be defined. Clearly, the definition of "the same target" can be a bit blurred for the case of NIRSpec MSA observations. Our plan is that the software will flag them as duplicate whenever the field overlap is significant (to be defined) and human intervention will be required to identify true duplications based on the slit positions. As it is the case for Hubble, duplications may be allowed if there is a scientific reason to do so, a typical example being monitoring of time variable phenomena.
Programs contain one or more Observations and Observations are composed of one or more Visits. Visits will not be interrupted but a multi-Visit Observation can be interrupted by Visits from other science or calibration programs or by maintenance activities as determined by the optimizations in the scheduling process. Maintenance activities include Wavefront sensing and control Visits (planned every two days but with mirror adjustments less frequent than that), momentum unloads and orbit maintenance.
Once the data are received they will be pipeline calibrated and archived. The users will access the archive and retrieve their data. We are still discussing what pipeline products and data analysis tools are needed.