Some capabilities to use more than one science instrument simultaneously (in parallel) will be available for JWST Cycle 5. Additional instrument combinations may be offered in future cycles.

See also: JWST Coordinated Parallels Roadmap, JWST Coordinated Parallels Custom Dithers

Parallel observing refers to simultaneously operating more than a single science instrument, each viewing a different area of the JWST focal plane. For JWST, there are 2 basic modes of parallel operations: coordinated parallels and pure parallels. 

Coordinated parallel observations are planned as part of a primary program, and are intended to amplify or supplement the primary science proposed in a given proposal. That is, the primary and parallel observations are planned together in a single program to accomplish the science goals, and hence are under the purview of a given proposal PI and science team. 

Pure parallel observations are from programs that make use of parallel observing slots derived from other accepted proposals. Hence, pure parallel programs are separate proposals, have separate PIs, and are selected independently from any primary programs they may be attached to. Several pure parallel slots have to be used for calibration observations, but there typically are several openings for pure parallel science observations as well.

There are a number of subtleties related to each of these modes of operation. For example, pure parallel observations are not permitted to impact or change the primary observations to which they are attached (with one exception described below), whereas coordinated parallel observations can be tailored to accommodate both the primary and parallel datasets since they are specified within a given proposal. (For example, dither patterns that work well with both instruments can be selected.) These subtleties are described below. 

The goal of enabling parallel observing is to enhance the efficiency and science return of the JWST mission.  However, users need to be aware that there are tradeoffs that need to be balanced. Not every potential parallel observing slot can be used for science.  One example is the limited data downlink capability available to the observatory.  For example, running NIRCam by itself in certain readout modes for several hours can use the entire data volume downlink capacity for a single contact period, leaving no room for storing parallel data taken during that period. Another consideration (for pure parallels) is that many calibration activities need to use pure parallel observing slots, and parallel calibration activities may take precedence over pure parallel science.  

Only certain modes and combinations of instruments will be allowed for cycle 5 proposals (see below).  Additional combinations may be made available in future observing cycles.

Principles in the use of parallel observing with JWST

See also: JWST Science Parallel Observation Policies and Guidelines

A detailed description of the policies related to JWST parallel observing is available. In summary, there are 3 principles that govern the policies for proposing and planning parallel observations:

1. For coordinated parallels, the science goals of the parallel observations must be tightly linked to the science goals of the overall program. In other words, accomplishing the main goals of the proposal must require parallel observations. Coordinated parallels whose goal is to address ancillary topics will generally not be approved.
   
   
2. Pure parallel observations may not impact the primary observations they are attached to^†. This means that pure parallel exposures must fit under the resource footprint (exposure plus overheads) of the primary exposure to which they are attached. It also means that parallel instruments are not allowed to change optical elements while the prime instrument is performing a dithered exposure sequence.
   
   
3. The needs of the JWST calibration program, a large fraction of which takes place in parallel mode, take precedence over pure parallel science programs.

Coordinated parallels 

See also: APT Coordinated Parallel Observations; JWST Position Angles, Ranges, and Offsets, JWST Background Model, JWST Data Volume and Data Excess, JWST General Target Visibility Tool Help

Coordinated parallel observations are crafted within the APT template used for the primary instrument mode. Nine template combinations are supported for Cycle 5, as shown in Table 1.

Table 1. Template combinations supported for cycle 5 coordinated parallel observations

Instrument teams have worked with developers to provide several ways of scheduling joint observations in parallel, for instance, to designate one instrument as the primary, as shown in options 1–3 in Table 1. 

One anticipated use case is to use both instruments at one epoch to observe adjacent areas of the sky, and to return at a second epoch when the instrument fields of view have rotated 180° on the sky. Users should obviously investigate the availability of the relevant position angles using the JWST target visibility tools since not all positions on the sky have the needed flexibility in available position angles to accommodate this strategy. Users should also be aware that background levels due to zodiacal emission and thermal emission from the telescope will change between two observations at different position angles. However, other science cases may not require such a "180° strategy" and can simply obtain observations (e.g., with NIRISS WFSS) of a nearby areal region to accompany NIRCam or MIRI primary imaging observations.

All dither patterns available for a given instrument template will also be available when it is selected as the primary instrument in coordinated parallel mode. However, for almost all prime+parallel instrument mode combinations, you also have the option to select from sets of customized dither patterns that have been specifically designed to produce good results for both instruments simultaneously. The number and type(s) of customized dither patterns that are available depend on the specific prime+parallel mode  combination. When a parallel instrument is selected, these additional dither patterns will become available in the dither pull-down menu in the relevant APT template. Note: customized dither patterns are not available for the NIRSpec IFU + NIRCam Imaging combination.

Since NIRCam observations involve up to 10 detectors (8 short wavelength and 2 long wavelength), proposers of coordinated parallel programs may need to consider selecting readout patterns that are less data intensive than they might otherwise choose. (Some of the most data intensive readout patterns for NIRCam are disallowed by APT in parallel mode.) Alternatively, depending on the science use case, a user may decide that areal coverage is less important than the parallel coverage and may opt to use only one of the two NIRCam Fields of View. APT will compute and display the data volume and data rate for each exposure, and you can adjust the detector readout pattern and/or the number of modules being read out as needed to stay within the allowed limits.

A number of additional coordinated parallel options are being considered for future cycles, including the possibility of using up to 3 instruments simultaneously. 

Pure parallels

See also: APT Pure Parallel Observations

Unlike coordinated parallels, pure parallel observations are proposed as entirely separate programs of investigation. Pure parallels use parallel observing "slots" created by exposures from other accepted programs that do not already have parallel observations specified. Pure parallel observations will not be allowed to influence the dither patterns or other aspects of the observing strategy of the primary observations to which they are attached, since the primary observations will belong to different science proposals.

In Cycle 5, the following 5 observing modes can be requested for pure parallel observations:

1. MIRI imaging
2. NIRCam imaging
3. NIRCam wide field slitless spectroscopy (WFSS)
4. NIRISS imaging
5. NIRISS WFSS

Not all observing templates will be allowed to have pure parallel observations attached. This is for a variety of technical reasons or concerns about how a pure parallel could potentially impact the data quality of the primary observation. Any of these 5 pure parallel instrument modes can be used in combination with any of the 11 models listed next.

Templates that are allowed to have pure parallels attached are the following:

1. MIRI imaging
2. MIRI LRS (only if Subarray = FULL)
3. MIRI MRS
4. MIRI WFSS
5. NIRCam imaging
6. NIRCam WFSS
7. NIRISS imaging
8. NIRISS WFSS
9. NIRSpec fixed slits
10. NIRSpec IFU
11. NIRSpec MOS

Templates that are not allowed to have pure parallels attached are the following:

1. NIRCam time series
2. NIRCam grism time series
3. NIRCam coronagraphy
4. NIRSpec bright object time series
5. NIRISS AMI
6. NIRISS SOSS
7. MIRI coronagraphy
8. MIRI LRS (if Subarray = SLITLESSPRISM)

Pure parallels FAQ 

1. What is a pure parallel observing "slot?" 
   A parallel slot corresponds to a single exposure of a prime program (usually a single dither position). A slot includes instrumental overheads, however mechanism moves are not included in the parallel slot, and the slot is not constrained by them. 
   
   

2. Can a pure parallel slot be assigned to more than one program?
   No; based on the specifications listed in the proposal, STScI will assign the pure parallel program to a prime slot. This is a one-to-one assignment, with no additional program being attached.
   
   

3. How are visits from prime programs used in pure parallel programs?
   The main "scheduling units" of the prime program are visits. Within a visit, each slot corresponds to a single exposure, i.e., a single dither position. If a pure parallel program selects 3-slots, thus creating a slot group, made up of a 3-point dither in one visit prime observation, APT will automatically create a 3-exposure one visit pure parallel observation. 
   
   
4. How do instrument configuration changes (i.e., filters/grisms) work for pure parallels? 
   Pure parallel observations are constrained by the number of mechanisms reconfigurations performed by the primary program. Parallel programs may not affect prime observations. Parallel programs may change mechanisms only if those changes are synchronized with mechanism moves in the primary program. In particular, it is currently not possible to change the mechanisms of the parallel instrument while the observatory is slewing between dithers. STScI is exploring whether it is possible to modify JWST's operating system to enable more flexibility.
   
   
5. Could you give a practical example of how mechanisms changes work?
   If a prime program has a visit with the following characteristics: 3 filters, each of them using a 4-point dither pattern on the same target, with a 500 s exposure per dither, for a pure parallel program, this will translate to:
   1. Twelve parallel slots of about 500 s duration each. Three of these slots correspond to the first dither position per filter.
   2. A pure parallel program could therefore use 3 different optical element setups following the same 4-point dither pattern used by the prime. 
   3. In this example, the pure parallel program could not use 4 or more optical elements setups.
      
      
6. Will the prime instrument "wait" for the pure parallel instrument to change filter/grating?
   Yes, the prime instrument will wait but the additional time will be charged to the parallel program. 
   
   
7. What if a pure parallel program wants to use X filters/gratings, but the available slots only allow for (X-1) or less number of filters/gratings?
   Because reducing the number of filters would be a change with respect to what the TAC approved, PIs are currently required to submit a change request to the Telescope Time Review Board (TTRB) in order to modify the number of spectral configurations. Pure parallels investigator teams are strongly encouraged to include in their proposals contingency plans for such cases, specifying which spectral configuration could be dropped if needed, and what the science impact would be. Approved programs that include contingency plans could fall back to those plans and would not need to submit a TTRB request. 
   
   
8. Overall, how many pure parallel opportunities were available in Cycles 1-4?
   The answer depends on the specific needs of each pure parallel program. For example, averaging over Cycles 1-4, roughly 350 visits per cycle were available for pure parallel observations requiring at least two instrument configurations and at least 1500 seconds of parallel exposure time per configuration, totaling roughly 1,400 hours of observing time. Variations from cycle to cycle are approximately 25%–30%. Of this available time, the pure parallel instrument can be MIRI about 70% of the time, NIRCam about 83% of the time, or NIRISS most of the time, due to the distribution of prime instrument usage in observations that allow pure parallels. If only one instrument configuration is required, the average number of visits increases by almost a factor of 3 and the overall time by 55%. In contrast, increasing the requirement to at least three instrument configurations reduces the number of available visits almost by half, with an equivalent decrease in total available time.
   
   
9. What are the characteristics of pure parallel slots and the prime visits that host them?
   The characteristics of pure parallel slots and the prime visits that host them depend on the suite of prime programs approved in each observing cycle. The previous question and answer gives average characteristics for three specific pure parallel observing scenarios, but it is not possible to simply describe the relevant multi-dimensional parameter space. Instead, proposers are strongly encouraged to use the jwst-pure-parallel Python package [documentation, repository] to assess pure parallel observing ideas. Proposers can use the package to see how many parallel exposure slots would have been available in previous cycles, with the characteristics needed for the proposed observations (e.g., number of required instrument configurations, number of dithers, exposure time).
   
   
10. Can you say anything about over-subscription on parallel slots?
    Cycle 2 approved parallel programs did see a degree of over-subscription: both accepted pure parallel programs competed for the same slots, in terms of galactic latitude and number of slots per prime visit. In such cases, STScI works with the PIs to accommodate, as best as possible, both programs. Table 2 lists previously approved pure parallel programs and provides additional information about hours allocated and hours used.
    
    

    Table 2. Slots currently used in Cycle 1 and Cycle 2 pure parallel programs

    Cycle	Pure parallel program	Hours allocated	Hours used	Category	Title
    1	PID1571	591	521	Galaxies	PASSAGE--Parallel Application of Slitless Spectroscopy to Analyze Galaxy Evolution
    1	PID2211	38	44	Solar system	A pure parallel survey of water in the asteroid belt
    1	PID2514	150	299	Galaxies	PANORAMIC -- A Pure Parallel Wide Area Legacy Imaging Survey at 1-5 Micron
    2	PID3383	205	160	Galaxies	JWST Wide Area 3D Parallel Survey
    2	PID3990	600	512	Galaxies	A NIRCam Pure-Parallel Imaging Survey of Galaxies Across the Universe
    2	PID4681	410	323	Galaxies	JWST Wide Area 3D Parallel Survey
    3	PID5398	400	314	Galaxies	POPPIES: The Public Observation Pure Parallel Infrared Emission-Line Survey
    3	PID6434	600	738	Galaxies	SAPPHIRES: Slitless Areal Pure-Parallel High-Redshift Emission Survey

    
    

    
    

11. Why can’t pure parallels change the filter while the prime visit is dithering?
    Observatory operations and software do not allow the parallel to change a spectral configuration during a dither. This constraint was included to protect prime observations from any impacts due to mechanism movements. Changes to parallels can be made between separate prime dither patterns. Modifying this constraint would require a profound change in the code  that would require extensive resources for implementation and testing.  Allocating such resources would have to be weighed against other  priorities.
     
    
12. Can’t the prime observations be adjusted to allow more flexibility for pure parallel observations?
    Dither patterns use one instrument configuration per instrument for all points in a dither  pattern. This approach reduces execution overheads, conserves limited  onboard memory, and controls testing complexity. In principle, prime observers could break a large dither pattern into multiple smaller dither patterns or even into individual exposures, providing an opportunity for mechanism moves between each pattern. However, this would lead to increases in observing overheads, sometime substantial  increases, and would require each prime program to adjust its observing strategy to accommodate the pure parallel observations. This is contrary to current JWST observing policies. In particular, as stated in this page: “Pure parallel observations will not be allowed to influence the dither patterns or other aspects of the observing strategy of the primary observations to which they are attached, since the primary  observations will belong to different science proposals.”
    
    
13. What is the cost of a pure parallel observation?
    Pure parallel observations are not awarded prime observing hours, yet they do consume significant observatory resources of various kinds. For example, pure parallel observations use extra observatory time to compile onboard scripts and move mechanisms, since these operations do not occur in parallel with the corresponding prime operations. Pure parallel observations consume mechanism lifetime, which is limited. Pure parallel observations use extra onboard recorder space and Deep Space Network downlink capability, which can cause scheduling challenges. Pure parallel programs require substantial user support because they can contain hundreds of visits that must be attached to dozens of prime visits, which evolve over the course of a cycle. Pure parallel programs receive signifcant grant funding. See this JSTUC presentation for more details. Despite these costs, pure parallels can yield great science, so we encourage proposals.