It is important that you be aware of and keep up to date with new or updated SPICE data being released. In most cases you'll want to use these data in your local processes (tools). But only YOU, the end user, can know for sure which SPICE data should be used for whatever analysis you are undertaking.
Having the ability to select which ancillary data are used in your own observation geometry calculations was seen as a "plus" to the originators of SPICE, this being one facet of providing scientists and engineers the chance to truly control the calculations being made. But having this flexibility obviously comes with a cost—the need for each user to pay close attention to the SPICE kernels he or she is using.
First Some Notes on Terminology
In the text below pay particular attention to the use of the terms "reconstructed" and "predicted." And take note that some people use the term "definitive" or "actual" instead of "reconstructed."
When speaking about a SPICE kernel containing time-dependent data, such as SPK, CK, SCLK and binary PCK, it is often not clear whether to use the term "new" or "updated." As a general rule, NAIF very strongly discourages changing the contents of a kernel containing time-dependent data without changing the name; at the very least the file name should contain a version number that gets updated. Changing just the version number portion of a file name most often happens if a previous issue of the file was found to have an error, and the new version simply corrects that error.
The more common reason for issuing an SPK, CK, SCLK or binary PCK is to extend—push forward—the time coverage of the file. Sometimes the time coverage of this next file will overlap the coverage of the previous release. In such a case does one call this next file "new" or "updated?" The file is new, but its contents may partially overlap those of the previous file, in which case the notion of "update" also makes sense.
Operational Mission Kernels
New operational SPICE data can come from several sources, and at widely varying frequencies. Probably the most obvious kind of new data is newly released spacecraft attitude (orientation) data (CK). This is true if a mission produces reconstructed spacecraft attitude, based on downlinked attitude quaternions, as the primary source of pointing information for end users. New reconstructed CKs might be released as often as several times per day, or once per week, or at yet some other frequency.
Some missions, particularly those operated by ESA, provide infrequent long duration predicted spacecraft attitude kernels as the primary deliverable to end users, supplementing this with telemetry-based reconstructed attitude only on special occasions.
New or updated reconstructed spacecraft trajectory data (SPK) may also be released rather frequently, although generally not as often as for CKs. Projects occasionally release new/updated predicted spacecraft trajectory data in SPK format for use in various mission planning activities. Sometimes "predict" and "reconstructed" data are combined together in a single SPK file.
An updated Spacecraft Clock Kernel (SCLK) might be released on a daily basis, or weekly, or less often.
Each project determines its own schedule(s) for releasing kernels, and often such plans change, or special releases are also made—so attention on your part is needed!
A project's Frames Kernel (FK) and individual Instrument Kernels (IKs) are generally implemented pre-launch with the best available data at that time. This process often involves some iteration between the FK and IK producer(s) and the spacecraft and instrument builders who have provided the data used in those kernels. Sometimes there are post-launch updates made, based on some sort of in-flight calibration activity. SPICE users need to keep an eye out for such updates!
Some "generic kernels" are also used in typical project activities. The Leap Seconds Kernel (LSK) and the Planetary Constants Kernel (PCK) are the obvious examples. Keeping up with new versions of these kernels can be problematic in that new versions are released quite infrequently and not according to any kind of schedule.
Some projects have assigned a person to construct and maintain a SPICE meta-kernel (reference the SPICE tutorial named intro_to_kernels) that "points" to the full suite of the very latest kernels needed for typical science operations. This can be a very nice service for science team members. But even in this situation you should pay attention to the SPICE data being provided to ensure this "one size fits all" offering of the project-provided meta-kernel is the right solution for you.
Some projects offer a File Notification Service that will send an email to a service subscriber every time a new kernel is released; the user can then download that new kernel using a script or a manual process.
Some SPICE users on active projects use a scripted "cron" job running the wget tool, or something similar, to query the project's SPICE server every day and download any newly found kernels.
Much of the discussion above speaks to practices of the NAIF Group at NASA/JPL when supporting JPL-managed projects; other NASA centers and operations centers in other countries that are using SPICE have their own procedures. (But NAIF does not currently produce meta-kernels supporting project operations; NAIF makes meta-kernels only for archive collections.)
New or Updated Archived Kernels
The above discussion speaks primarily to SPICE users involved with "current" data analysis on an on-going mission. But SPICE users making use of archived SPICE data, whether from an on-going mission or from a completed mission, must also pay attention to the SPICE kernels being used, sometimes paying attention to specific geometry parameters included therein. Why?
Ideally such post-archive improvements would find their way into the SPICE archive, and such is often the case. In these circumstances the SPICE user must decide whether to use such updated (post-archive-production) data, perhaps to obtain the very best possible answers, or to use earlier versions, perhaps to make computations consistent with those already made by colleagues, or those used in a publication.
- Improved spacecraft trajectory or improved spacecraft orientation might be produced AFTER an archive is made.
- Changes to target body orientation, size or shape data might be determined AFTER an archive is made.
- Improvements might be made to instrument mounting alignments found in a project's frames kernel (FK), or to instrument field-of-view geometry found in an IK, AFTER the archive has been made.
New or Updated Generic Kernels
NAIF makes an announcement each time an updated leap seconds kernel (LSK) or planetary constants kernel (PCK) is released; this is done using the "spice_announce" Mailman system noted at the bottom of the NAIF home page.
NAIF often announces the availability of a new, generic planet ephemeris kernel (SPK), such as "de421.bsp." However, NAIF does not announce new satellite ephemeris kernels (SPK), since these arrive rather frequently.
NAIF also does not generally announce assorted other new generic kernels such as high precision binary earth pck, high precision binary moon pck, tracking stations spk and fk, and digital shape kernels (DSK).
Which SPICE Kernel(s) to Use?
Learning that a new kernel is available is one thing, but deciding whether or not to use it versus another one can require some effort. See
kernel selection for a discussion on this topic.
Improving the Situation
There is certainly room for improvement in the way that availability of "new" or "updated" SPICE data are announced; without making any promises to take action, the NAIF Team solicits your suggestions.