JWST Time-Series Observations Known Issues
A number of known issues for time-series observations in the proposal preparation process is discussed in this article.
Limitations on the total number of frames
A single exposure on JWST can contain a maximum of 196,608 frames due to on-board storage limitations. The constraint is of particular concern for TSOs, as these are the only types of JWST observations that are generally executed in a single, lengthy exposure for optimal science.
Words in bold are GUI menus/
panels or data software packages;
bold italics are buttons in GUI
tools or package parameters.
For example, consider a TSO with the MIRI low resolution spectrometer (LRS) in slitless mode. This mode exclusively supports MIRI's FASTR1 readout mode, in which each group contains a single frame (i.e., Nframe = 1). A single read of the MIRI subarray used for this mode takes 0.159 s. If the target requires Ngroups = 35 to achieve the desired signal-to-noise ratio (S/N), a single integration will have a duration of (35 + 1) * 0.159 s, or 5.724 s. The maximum number of integrations this exposure can hold is 196608/35 ≈ 5617 integrations, or an exposure time of 5.724 s * 5617 ≈ 8.9 hours. Note that the additional reset between integrations that is part of the FASTR1 read mode takes one group time, accounting for the "+1" in the above arithmetic, but does NOT generate any data. It therefore does not contribute to the overall frame count.
Different workarounds are possible for this issue. The best strategy for reducing the total number of frames in an exposure depends on the target, the instrument, and the mode of observation:
- Larger subarrays have longer read times and will therefore produce fewer frames for a given observing time. If your observation is being executed in a subarray, and a larger array configuration is available that can accommodate the target, this will reduce the number of frames.
- The observation can be executed in multiple exposures, with each exposure containing < 196,608 frames. For some TSO modes, it is possible to set Exposures/Dith in APT to a value > 1. In this case, the exposures will be executed back-to-back with identical Ngroups and read mode configuration. No new guide star acquisition or target acquisition sequences are required, and the break between the exposures will be very short (of order a few minutes). When changing the configuration using this strategy, you are advised to check where the exposure breaks will fall in your observation, and adjust the phase start window to ensure the time series will not be interrupted at a critical time (e.g., during ingress or egress).
- For instrument modes that do not have the capability of specifying multiple exposures for TSOs, the observation has to be divided into multiple observations, each with a single exposure. The observations should be linked as a non-interruptible sequence using APT special requirements. In this case, only the first observation should have phase constraints specified in the special requirements; the "non-interruptible sequence" requirement will ensure that the subsequent exposures immediately follow the first. In this case, the break between exposures will be longer, as new observations require new guide star and target acquisition sequences.
If a workaround was implemented successfully, the APT error should disappear.
Scheduling of phase curve observations
See also: APT Timing Special Requirements
Time-series observations of periodic events, such as transiting exoplanets or eclipsing binaries, require phase constraints to allow APT to specify the observations such that they cover the scientifically interesting part of the target's orbit. These are specified in the special requirements tab in APT. For very lengthy observations, where the duration of the visit is longer than the gap between the scheduling windows, APT is unable to schedule the observations. This impacts all phase curve observations of transiting exoplanets. In these cases, APT will return the following error:
Expressed mathematically, this software limitation occurs when:
Visit scheduling duration > PERIOD x (1 - (<number 2> - <number 1>))
where PERIOD is defined as the gap between the start phase windows, not the period of the exoplanet; and <number 2> and <number 1> are the 2nd and 1st entries in the phase range, respectively.
The workaround for this issue is to double the length of the object's period and halve the numbers in the phase window. This effectively halves the opportunities for observation, but allows the observation to be scheduled as designed. For example, for a target with period P = 0.8 days and starting phase window = (0.3, 0.4), the workaround would require changing P to 1.6 days, and the phase window to (0.15, 0.2).