About this Article

Published October 8, 2025

The Hubble and Webb annual proposal cycles follow well-defined cadences for members of the astronomical community:

• The Call for Proposals is issued;
• Principal Investigators submit their proposals by the deadline;
• The Telescope Allocation Committee (TAC) reviews the proposals and submits recommendations to the STScI Director;
• The Director approves the final program selection for each cycle;
• The approved teams produce flight-ready versions of their programs, converting Hubble Phase I proposals to Phase II programs and finalizing single-stream Webb programs.

The ingest of these completed programs triggers another annual sequence of events: STScI staff integrate the new cycles of observations into the overall observing plan for the coming year for each mission.

Getting observations onto the telescopes is a two-stage process. First, planners construct the Long-Range Plan (LRP), which maps observing windows over the next 12 to 18 months for most observations, indicating when they could be scheduled. Next, the actual observing schedule is built on a weekly basis, drawing observations from the LRP.

Constructing the Long-Range Plan

Both Hubble and Webb can observe objects over the full celestial sphere at some point during the year, but both have a limited field of regard at any specific time. For Webb, the constraints are set by the requirement that solar radiation should not strike its optics. Thus, Webb can observe targets between 85^o and 135^o from the solar line-of-sight, covering ~40% of the sky. Webb observations are skewed to minimize pointing in the direction of (solar) orbital motion, with the aim of limiting mirror degradation through micrometeoroid impacts. Observations within the Micrometeoroid Avoidance Zone will be scheduled if such is required to achieve the science goals.

Hubble’s move to Reduced Gyro Mode in 2024 reduced its average coverage to ~50% of the sky, with more complex evolution of the specific (RA, Dec) access throughout the year. In low Earth orbit, Hubble’s observing schedule is interrupted when the orbit passes near the South Atlantic Anomaly. Superimposed on these global considerations are more detailed constraints set by orientations and/or timing requirements. All of these constraints need to be taken into account when identifying appropriate scheduling windows for each observation.

Each annual TAC recommends a science program that covers close to one year’s worth of programmatic science. The individual targets, however, are not distributed uniformly over the sky (see Figure 1) – there are obvious concentrations, primarily around high-latitude extragalactic deep fields, that tend to attract multiple major programs. These concentrations lead to scheduling overlap that make it impossible to execute all observations in the same calendar year. In contrast, scarcity of targets in particular regions can lead to under-subscription at other times of the year.

The key take-home message is that the science program recommended by an annual TAC, either Hubble or Webb, will not result in observations that pack neatly into 365 days of continuous scheduling. It is inevitable that some programs will be delayed by a year or more. Programs with tight constraints are particularly vulnerable to such scheduling challenges. It is essential that the Principal Investigator ensures that those constraints really are necessary to achieve the science goals.

By the same token, proposers should not add constraints after a program is accepted, since those constraints directly impact the ability to schedule those observations. Such additions will generally only be allowed if the constraints could not have been foreseen when the proposal was submitted.

Figure 1: Hubble Cycle 33 GO Sky Distribution

Planners start identifying scheduling windows once observing programs are finalized, soon after acceptance for many Webb programs and following Phase II submission for Hubble programs. Some observations require multiple iterations to manage issues, such as intensive data volume, before they become schedulable, while a few programs, notably Target of Opportunity (ToO) programs, cannot be scheduled until more information is available later in the cycle. With around 10,000 observations per cycle, constructing the individual plan windows and then combining those windows through multiple iterations to build an optimized, efficient LRP takes time, typically requiring at least 10 weeks from the PI notifications.

Figure 2 shows the initial LRP from Webb Cycle 4 planning, mapping the observations on a daily basis in terms of the total duration (exposure time plus overheads) and the corresponding data volume. In general, planners work to match 80% capacity in both cases, keeping the data volume in hand; the extra capacity in duration is contingency to allow for additional in-cycle observations. The narrow spikes above the 80% lines generally reflect time-critical observations, notably exoplanet transits or eclipses, or data-intense programs.

Figure 2: Webb Cycle 4 Observations Initial Long-Range Plan

The overriding priority in creating the LRP is maximizing observing efficiency while maintaining all necessary observing constraints. Observations don’t exist in isolation; they live in an ecosystem defined by all the other accepted programs. Figure 2 shows that the full Cycle 4 program of observations extends well beyond the Cycle 5 boundary, with smattering of observations with plan windows in late 2026. Principal Investigators with observations in the tail often request that their programs be pulled forward in the schedule. This is actually less than a zero-sum game, since pulling those observations forward not only displace already scheduled programs, but also lowers the overall scheduling efficiency. Consequently, those requests are only considered if there is a compelling scientific rationale.

Constructing the LRP is not a once-and-done activity. The plan evolves weekly. New observations, including ToOs, Director’s Discretionary programs and repetition of failed observations, enter the scene during the cycle, and future observations will be pulled forward to fill out the near-term calendar if suitable opportunities arise.

Building the Short-Term Schedule

Both Hubble and Webb are scheduled on a weekly basis. The schedule, the Science Mission Specification (SMS) for HST or Short-Term Schedule (STS) for Webb, draws from the pool of observations with appropriate plan windows in the LRP. Each weekly schedule lays out observations that maximize the observing efficiency, while taking account of time-critical observations and giving preference to observations where the plan window is drawing to a close, particularly those from previous cycles. This step includes identifying appropriate guide stars for each observation. If necessary, observations are pulled forward from future opportunities. Finally, Survey (Webb) or SNAPSHOT (Hubble) observations are added to fill out the week’s observing to the maximum extent possible.

Constructing an SMS/STS takes time. For Webb, STScI schedulers begin compiling and matching observations on Monday. Hubble schedulers start the previous Thursday. Once the schedule is complete, schedulers generate flight products. Those products are transferred to the respective Flight Operations Systems, generally on Thursdays prior to the start of execution. Typically, the final schedules are transmitted to the observatories on Sunday via the Tracking and Data Relay Satellite System (TDRSS) for Hubble and the Deep Space Network (DSN) for Webb. The first observation for both missions is generally scheduled for that evening, 10 days after work started for Hubble and a week for Webb.

Scheduling on the two observatories follows different operational strategies. For Hubble, observations are quantized in orbits, with Earth occultations forming natural boundaries. If an observation fails due to, for example, lock on a guide star, Hubble continues with the planned schedule, picking up the next target for the subsequent orbit. Webb, at L2, follows a different approach with event-driven scheduling. If an observation fails, then the schedule moves onto the next target that can be observed. This can lead to Webb sitting idle if there are upcoming observations that need to execute at a specific time, notably exoplanet transits and eclipses. In those cases, the scheduling team may develop a revised STS, eliminating dead time, that can be uploaded when the DSN presents a suitable opportunity. Around 30% of Webb’s time is devoted to exoplanet observations, and most of those are time critical. Consequently, the short-term schedule requires frequent revisions to maintain observing efficiency, despite the innovation of event-driven scheduling.

Targets of Opportunity and Director’s Discretionary Time Programs

Both Hubble and Webb accommodate observations of transient phenomena discovered during observing cycles. Those programs are submitted either as Target of Opportunity (ToO) proposals, adjudicated by the annual TAC and anticipating in-cycle events such as supernovae or gamma-ray bursts, or as Director’s Discretionary Time (DDT) proposals, responding directly to new discoveries and submitted during the cycle. In practice, DDT programs may have a longer path to implementation since they must be sent for external community review. This usually adds at least 5 days to the timeline. For that reason, targets that require rapid (< few days) turnaround observations (e.g., kilonovae, young supernovae) are much better suited to ToO proposals, submitted via the annual TAC process, especially if there is a reasonable expectation for such phenomena in any observing cycle.

The challenges involved in building DDT and ToO observations into the schedule depend on the required response time. If the turnaround time is a few weeks or more, then these observations can be incorporated through the standard processes. Faster turnarounds are more challenging. With pre-planned, uploaded weekly observing schedules, Hubble and Webb do not have the flexibility of ground-based observatories to swap in new observations. Allowing for the SMS/STS preparation time and the subsequent week of actual observing, any DDT/ToO observations within 2 to 3 weeks are disruptive and require adjustments to the schedule currently under preparation or already executing onboard. Rapid turnarounds of less than a week generally require an interrupt in the ongoing observations to upload a revised set of flight products.

The intercept schedule is built to minimize the impact on other time-critical programs and maintain efficiency but still accommodate the new observations. These re-builds are major efforts, often requiring around-the-clock work from schedulers and instrument teams, which is why the number of these opportunities is limited. Ultra-disruptive (< 2 day) response times are possible — Hubble recently achieved a turnaround time of just under 24 hours (from ToO trigger to start of observation) for follow-up observations of the kilonova candidate, AT2025ulz. This kind of response is only possible if everything aligns, including a near-flight-ready program submission and appropriately placed TDRRS/DSN contacts.

Rescheduling Failed Observations

As noted above, despite all the best-laid plans, a few percent of observations can fail for a variety of reasons. Generally, those observations will be repeated. Proposers are informed if an observation does not go to plan, and the Principal Investigator may request a repeat by submitting a Hubble or Webb Observation Problem Report. These HOPRs/WOPRS are reviewed by the Telescope Time Review Board (TTRB), which will generally recommends the repeat unless the mishap resulted from an error on the part of the observing team (e.g., incorrect target coordinates) or the program is > 90% complete.

Re-inserting the failed observation into the current or next week’s schedule requires the same level of effort as scheduling a DDT or ToO. Consequently, this work will only be undertaken if there is compelling scientific rationale for doing so. Unfortunately, this can lead to significant delays in executing the repeat, particularly for Webb given its narrower scheduling windows. Loosening constraints generally helps accelerate the repeat.

Key Takeaways for Proposers

1. Planning and scheduling Hubble and Webb science programs is a highly complicated, time-intensive process that requires continuous adjustment throughout the year.
2. It is inevitable that some observations allocated in Cycle N will need to be scheduled in the calendar year corresponding to Cycle N+1. Tightly constrained programs may drift even later.
3. The fewer the constraints on a given observation, the easier it will be to schedule, and the higher the prospects of early scheduling. If there are scientifically justified special requirements, those must be specified in the original submission unless they were unknowable at that time.
4. ToO programs, rather than DDT submissions, are the better option for fast turnaround, highly disruptive observations. Find further information for Hubble and Webb.
5. Rapid rescheduling of failed observations will only be attempted if there is a compelling scientific justification.

The planning and scheduling team at STScI is led by Bill Workman.

The dedicated staff that support STScI’s three missions include:

• Long-Range Planning Group: Brigette Hesman (Technical Lead), Vanshree Bhalotia (Deputy Technical Lead), Melissa Hoffman (Deputy Technical Lead/Roman Planning and Scheduling Lead), Fawn Brewer, Ian Jordan, Shelly Meyett, and Alison Vick
• Short-Term Scheduling Group: Kristen Wymer (JWST Technical Lead), Jim Caplinger (HST Technical Lead), Melissa Hoffman, Gary Bower, Anil Dosaj, RJ Lampenfield, Alyssa Werynski, Deena Mickelson, and Ethan Polster
• Planning and Scheduling Operations Support Staff: Danny Jones, Ryan Logue, Don Chance, George Chapman, Mike DeRito, Mike Harr, Merle Reinhart, and Aidan McCormick

Acknowledgements: Thanks to Tom Brown, Andrew Fox, Brigette Hesman, Julia Roman-Duval, Laura Watkins, and Bill Workman for comments and suggestions on the original manuscript. Generative AI was not used in writing this article.