HSpot deals with two fundamental types of target: fixed targets and solar system objects.
A fixed target is any object that does not require a differential tracking rate. This can be a star, a galaxy, an AGN, etc. Herschel works with Equatorial J2000 coordinates and only target entry in Equatorial J2000 will be accepted (this was set to facilitate checks for duplicate pointings [two or more users requesting similar observations of the same object, or region of the sky], which are extremely complicated if many coordinate systems were used for target entry). If the source is known to NED or SIMBAD these coordinates are used, if not, the user must enter a J2000 R.A. and Dec. On some occasions, for nearby stars, the proper motion of the target may become important; this can be entered in HSpot if necessary, once again, the epoch must be in 2000 coordinates. All fields can be edited after name resolution.
Some types of observation could be very unforgiving with the coordinates, particularly PACS spectroscopy and HIFI observations in the high frequency bands (Bands 6 and 7) are sensitive to poor pointing accuracy and may have been affected seriously if the quality of the position of a fixed target was low. Many faint far infrared targets do not have good coordinates and the optical position and far infrared position of a particular target may not necessarily coincide: this should be carefully bourne in mind when comparing Herschel and optical data: a discrepancy may not necessarily indicate poor pointing, it may be simply an indication that the optical and far infrared source do not coincide.
Typical problems found when entering user-defined coordinates for targets were missing "minus" signs in declination and, where several positions are available for a target, picking an inaccurate one.
![[Warning]](../../admonitions/warning.gif) | Warning |
|---|---|
| Caveat Emptor! The PI was (and is) responsible for the accuracy of the submitted coordinates and for ensuring that the coordinates supplied were good enough to obtain the requested data. Observations that failed because of bad coordinates could only be repeated if the time required came out of the proposal's HOTAC allocation, requiring other observations to be cut in duration, or even removed completely from a proposal. |
The pointing performance of Herschel varied considerably through the mission as it was better understood and refined. Detailed information of the pointing performance and its evolution can be found in Section 2.4.5.
A moving target is a solar system object that requires a differential tracking rate to be programmed. On target entry the user needed to select the "Moving" tab and resolve the NAIF ID of the target name. The Herschel Mission Planning System used the NAIF ID to calculate coordinates for the time of observation and to calculate the differential tracking rate required. This needed to be less than 10 arcsec/minute at the date of observation (this limited the capability of Herschel to see objects passing very close to the Earth, although faster rates up to and even slightly exceeding 30 arcsec/min could be permitted, on a case-by-case basis, if scientifically justified). User entry of target coordinates is not permitted, as any solar system object with a reliable enough orbit to have been observable by Herschel will have a NAIF ID.
Around 800 moving targets (satellites, comets, asteroids and TNOs) are in the HSpot database. More than a million have been catalogued, but it is obviously impractical to store all of them in HSpot as most were not observable by Herschel. To observe a solar system object that was not in HSpot, for example, a new discovery, it was necessary to send a Helpdesk ticket requesting that it be added. A minimum of two or three working days were needed normally for it to be included and for the ephemeris to be linked to HSpot.
The NAIF ID is shown for solar system objects in the HSpot "position" column rather than an R.A. and Dec. No position data is given for solar system targets in the observing log as this is calculated 'a posteriori" when the data is processed, based on the ephemeris used in downlink (which may not be the same as the uplink ephemeris used to schedule the observation) and the timing data for the observation as executed.
NAIF is NASA's Navigation and Ancilliary Information Facility. This offers an information system called SPICE for spacecraft navigation. SPICE uses a unique 7 digit identification code for all natural solar system bodies, while spacecraft are identified with a negative integer code. Because of the simplicity for this system of ID codes and given the increasing possibility of confusion of objects (for example, there are both planetary satellites and asteroids named Io, Ganymede and Dione and increasing numbers of asteroids are later found to show cometary activity and may receive multiple designations), it is increasingly used for telescope scheduling. A short summary of information about NAIF IDs is given in the relevant secion of the HSpot Users' Manual on the Standard Ephemeris for moving target entry.
When a Solar System Object has a well-controlled orbit of high accuracy (for a periodic comet this means two returns for which a successful linkage has been made, for a asteroid or minor body it usually means observations at a minimum of 6 or 7 oppositions, apart from Earth-crossing objects for which the criterion is typically 3), it will receive a number from the Minor Planet Center. A numbered comet has a designation such as 190P/Name, while an asteroid receives just a number. An unnumbered asteroid has a NAIF ID starting with a 3. Objects with such a designation have a relatively low accuracy ephemeris that may be considerably in error when extrapolated even a short time into the future. As an example, even an object with three oppositions may have a position that has a 3-sigma error of more than 60 arcseconds when extrapolated 5 years into the future. If the spread of observations is unfavourable, or there are few astrometric observations, it may not even be possible to obtain a good ephemeris extrapolation with a 3-opposition orbit. With 4 oppositions the 3-sigma error in the extrapolated position may still be greater than 20 arcseconds over 4 years. This means that faint objects that have not been observed recently may have been difficult to locate and identify with Herschel and thus were high-risk observations. It also meant that an object may not have been centred on the detector, so the resultant data quality may have been defficient.
For comets, non-gravitational forces can make predictions quite inaccurate for returning objects, even when they are well-studied and the closer that they come to Earth, the worse the situation is in terms of ephemeris accuracy. Experience has shown that close approach comets always needed ground-based astrometric support campaigns and late-time ephemeris updates to make observations possible with Herschel and that even updating the ephemeris with the latest data 4 days before the observations were to be executed (this was the latest possible time for update to be made to allow it to be processed in time for Uplink) and re-planning the observations was not always enough to avoid positional errors of several arceseconds between the expected and the actual position at the time of observation if no high-precision radar observations were available to tie-down the position with exactitude. As an example, the difference between the predicted and observed postion of 45P/Honda-Mrkos-Pajdusakova was 7 arcseconds, despite an ephemeris update four days before the observations were executed, as no radar data were available until after the execution of observations with Herschel.
When re-processing data in bulk re-processing at HSC, the best available ephemeris is used. As in the case of 45P/Honda-Mrkos-Pajdusakova, this may be considerably better than the one that was available at the time of observation. Observations in the Archive thus are processed with the best possible positional information at the time of processing, not the best that was available at the time of observation.
Three problems may be present when there is uncertainty in the ephemeris. In approximate order of increasing importance these are:
Possible errors in the required tracking rate.
In general the tracking errors should be kept below 1 arcsecond during the observation. This could be a problem with long observations on objects moving at high velovity on a strongly curved track for which the interpolation of the position may not be good enough. In the end though, it seems that this has not been an issue for any solar system object observed by Herschel.
Difficulties with photometry
For the observation to have been carried out successfully, the target must have been centred in the array to within a certain level of accuracy. If this was not achieved, it may be difficult, or impossible to obtain good quality photometry.
Problems with target identification
Not all Solar System Objects have suitably accurate ephemerids to have been observed successfully at all; occasionally there may be errors of tens of seconds of arc, minutes or even, for a few objects, degrees in the ephemeris position. In the HSpot Users' Manual a list of solar system objects included HSpot is given in which flags objects with deficient ephemerides at the time of launch. It was always recommended to check though to see if a better orbit was available when requesting time.
In detail, the issues that users may find when ephemeris information is uncertain are:
Tracking
In general this should not be a problem with distant objects, it may become a serious problem with more nearby ones, particularly Near Earth Objects where it may be difficult to keep the target accurately centred.
Photometry issues
For PACS photometry, the source position must be known with high enough precision that it should fall within a bolometer matrix of 52x52 arcseconds. In practical terms this means that the following criteria of positional accuracy should be fulfilled.
-- For aperture photometry: 15 arcseconds.
-- For PSF fitting: < 10 arcseconds
For SPIRE the main consideration is that the FWHM of the detectors is 18 arcseconds and the jiggle amplitude 6 arcseconds: if the positional error is greater than the jiggle amplitude there will be light losses.
For HIFI it should be remembered that the smallest aperture (that of Band 7b) is 13 arcseconds and for PACS spectroscopy the pixel size was of the order of 9 arcseconds, thus necesitating, in both cases, centering at the arcsecond level to avoid light losses.
Target identification problems
For numbered asteroids the ephemeris should be of sufficient precision in almost all cases.
For unnumbered asteroids and minor bodies it has been essential to take astrometry to refine the orbit before observations could be attempted with Herschel.
For numbered and ToO comets, recent astrometry may have been essential, depending on the case. A numbered comet has almost invariably required post-recovery astrometry to refine the orbit before observation could be attempted and any close-approach object will need extensive late-time astrometry. Recently discovered comets with a short orbital arc have invariably required late pre-Herschel observation astrometry to refine their ephemeris.
Standard practice at the HSC was to download the ephemerids for Solar System Objects from the JPL Horizons database every 4 weeks, as our planning was done in cycles of 2 weeks. This means that in an extreme case the ephemeris information for an object may have been as much as 2 months out of date. Normally this did not matter, as the errors will be too small to be significant for Main Belt asteroids. If the orbit is well-defined, the error is usually almost entirely in the direction of motion, so a Solar System Object (asteroid or comet) will reach a given point in its orbit slightly advanced or slightly delayed with respect to the ephemeris prediction. For Main Belt Asteroids (MBAs) advances or delays of 10 or more minutes of time in the position along the orbit are not unknown; but, at a typical distance of observation of 2AU from Earth, this translates into a very small error in the actual observed position on the sky.
For objects that come closer to the Earth and move more rapidly, the error in the ephemeris may be much larger, even though the absolute precision of knowledge the object's position may be an order of magnitude better than for an MBA if it had been intensely observed. In the case of Comet 103P/Hartley 2, at the time of observation it was found to be almost 30 arcseconds away from the ephemeris prediction published only 2 months previously. For Near Earth Asteroids and comets an ephemeris may become effectively completely unusable in a week or less. Observers had to be aware that it was their responsibility to ensure that there is sufficient knowledge of the ephemeris of a target for effective scheduling and to warn the HSC sufficiently in advance to take the necessary measures to schedule with the most up-to-date available information.
Normally observations were planned and sent to MOC for uplink to the satellite a minimum of 2 weeks in advance and more usually at least 3 weeks in advance. If knowledge of the object's position was likely to be insufficient at that time, the observer was expected to request -- in advance -- that the observations be re-planned closer to the date of execution, taking advantage of a better ephemeris. Such cases required careful forward planning, both at HSC and MOC to ensure a late-time re-delivery of the observations could be made and processed successfully. On numerous occasions it was also necessary to coordinate carefully the re-delivery with additional ephemeris calculations and updates at JPL, who frequently made exceptional efforts to support the observation of solar system targets with Herschel, including rapid processing of radar observations from Goldstone to make a highly accurate ephemeris available within a few hours of an observation at Goldstone, without which some observations would have been totally impossible to execute successfully.
It was the responsibility of the observer to warn the HSC, via a Helpdesk ticket, of potential ephemeris problems that might have affected scheduling of a Solar System Object. This needed to be done far enough in advance to be taken into account in the standard planning cycle and for the necessary actions and additional deliveries to be discussed and agreed with MOC. However, no re-planning of observations could be contemplated less than 4 days in advance of their execution; if the ephemeris was potentially not robust for 96 hours in advance of execution, the observations had to be designed to be robust enough to compensate for any positional errors that may have occurred.
It was not sufficient to assume that the HSC would automatically spot all potential conflicts with Solar System Objects in advance, although this was part of technical assessment of proposals and every effort was made to anticipate such conflicts.