An AOT is an "Astronomical Observation Template". This will be familiar to users of ISO and Spitzer. An AOT is a standard observing mode with an instrument that can be translated into instructions for the spacecraft to carry out the observations autonomously. Herschel observed autonomously between DTCPs, so each observation had to be carried out in a standard way that the spacecraft could understand. Thus, for each of the instruments only pre-defined types of observations could be carried out. The astronomer produced an AOR (Astronomical Observing Request) by taking an AOT and customising it for the required observations.

Following the experience of ISO, the number of AOTs was deliberately restricted to allow observers as many options as possible, without requiring an unwieldy number of observing modes to be validated and calibrated.

The first stage in AOR entry was to define the target. If it is a known object, its name can be resolved with SIMBAD or with NED or, for a solar system target, as a NAIF ID. For unknown names (e.g. start points for scans), J2000 coordinates must be supplied by the observer. After defining the object, the observer needed to check that it was observable by Herschel by calculating its visibility windows. It was also necessary to bear in mind that when defining an observation further, the observer could end up limiting its visibility to part of the target's unrestricted observing window or, worse, making it completely unobservable by putting constraints that were too strong on the observation.

Once the target is defined the observer must then select the required instrument and AOT to be used. Nine basic observing modes were supported: for HIFI, single point (point source spectrophotometry), mapping and spectral scans; for PACS, photometry, line spectroscopy and range spectroscopy; for SPIRE, SPIRE photometer and spectrometer; and the SPIRE PACS Parallel Mode (this last mode was only available for fixed targets and could not be used for Solar System Objects). Each of these modes is further subdivided, HIFI, for example, offered a choice of fourteen different mixer bands. PACS photometry allowed three variants including the mini-scan maps which replaced point-source photometry early in the mission and chopped raster maps. SPIRE Spectrometer offered point source and raster maps, three choices of image sampling, and four choices of spectral resolution, etc. HSpot will guide you through this process of definition with a series of pull-down menus and pop-up windows. All these options still work correctly in post-Operations and can be used to investigate the effects of taking data in slightly different ways to understand your observations.

Warning

Although Point Source photometry may be selected through HSpot, it was not offered as a standard science mode in routine operations and remained only for specialist calibration applications, save for a very few science observations executed in this mode early the mission, which served to show that its sensitivity was lower than had been hoped. Point source photometry for science data was carried out using mini-scan maps only, which are far more sensitive for the same integration time, although calibration observations taken in Point Source mode are fully calibrated and can be used for science.

Figure 6.2. Beating the confusion limit for Solar System Objects. This SPIRE image of Makemake uses the technique of subtracting the background using the shift in position between two epochs and subtracting one frame from the other. The trans-Neptunian Object (TNO) Makemake has an estimated flux of 15mJy at 250 microns, with a confusion noise sigma of 6mJy. A weak, but clear detection is obtained in this image with 15 repetitions (NB: the confusion noise is effectively reached in 2 repetitions). This is thought to be the faintest target to be detected with SPIRE.

For each observation there was a basic minimum unit of observing time required; the observer needed only specify how many repetitions of this unit time are required -- obviously greater sensitivity is obtained through more repetitions (four integrations will give twice the sensitivity of a single one -- although for SPIRE, once the confusion limit was reached, you could not attain better sensitivity however long you integrate, although it was possible to obtain better signal to noise), but the observation took longer for a relatively small gain. At any time the "Observation Est..." (Observation Estimate) button can be pressed and HSpot will give an estimate of the total time that the observation would have taken, including the overheads involved, with a break-down of information about the observation. If the total length of the observation exceeds the maximum permitted, HSpot will give a warning that the observation duration is out of limits.

The one exception that allowed the confusion noise to be beaten is for moving targets -- Solar System Objects (SSOs). By using an off position we can subtract out the background and thus eliminate almost completely the confusion noise (see: Figure 6.2). For a sufficiently long exposure of an SSO, the target will move sufficiently during the exposure that the end of the exposure can act as the "off" for the start; this has proved very effective for detecting extremely faint SSOs down to the limiting sensitivity of the telescope.

The observer could vary the parameters of the observation (more or fewer repetitions, nodding on or off, larger or smaller chopper throw, a wider or narrower range of wavelengths or length of scan, etc.) and see how the time estimate varied. Once an acceptable combination of parameters had been found the observer accepted the parameters that were defined to fix the AOR; this AOR can however be modified later, if necessary.

When a proposal was submitted, HSpot took the currently defined list of AORs and linked them to the proposal. It was thus essential to ensure that the correct AOTs and AORs were defined and that the source visibility and observing time were correct for each target.

The star tracker (the telescope's autoguider) was pointing in the opposite direction in the sky to the telescope. So, when an object lay in the hemisphere away from the Sun the star tracker was pointed into the sunward hemisphere. It was found that when the solar elongation of a target is greater than 105 degrees, sunlight illuminated the bottom of the S/C platform, leading to heating and thermal distortion of its structure. This distortion caused progressive guiding errors that affected not just the current observation, which would see progressive drift in the tracking, but also all the following observations until the the S/C structure had stabilised. On return to cold attitudes, the process was reversed, with the pointing slowly returning to nominal, again causing progressive pointing drift, thus a long observation at hot attitudes would affect not only the observation in question with a progressive pointing drift, but also later observations.

After in-flight study it became obvious that data quality would be compromised by long exposures in the region of solar elongation from 110 to 119.2 degrees (the "hot zone"). HSpot showed observations in this range as having "limited visibility". In practice this meant that according to the degree of incursion into the hot zone increasing strict limits would be put on scheduling. As a rule of thumb, observations longer than 1 hour would not be scheduled in the hot zone save in exceptional cases.

Observations in the hot zone were scheduled at the end of the observing day, unless that were very short incursions that went only slightly into the hot zone of it could be demonstrated after careful assessment by the Spacecraft Environment Scientist and by the Herschel Mission Planners that they would have no significant impact on the quality of observations that followed: this allowed the star tracker base plate to cool during the Daily Telecommunications Period so, only, in normal circumstances, only short and extremely urgent observations will be scheduled in this hot zone. Time-critical observations longer than an hour that only impinge slightly into the hot zone (e.g. SSOs) were treated on a case-by-case basis.

With the improvements in Herschel's pointing performance, it became obvious that degradation of pointing occurred at smaller solar aspect angles than previously thought, being particularly important for PACS spectroscopy taken after observing in a "hot" region of the sky, which could be seriously compromised by even small errors in pointing. This obliged us to reduce the area of the sky considered to be "cold" and thus available for unlimited scheduling, considering any observation at a solar elongation greater than 105 degrees to be warm (limited to short observations that would not affect later pointings) and greater than 110 degrees as hot (only to be scheduled if scientifically critical and, only then, taking all necessary scheduling precautions, including scheduling at the end of the day and placing only pointing-robust observations immediately afterwards). From late July 2012 HSpot showed 75% of the visibility window of a fixed target as "cold" and thus available without limits for scheduling and 25% as "hot", in which progressively tougher scheduling limits were imposed the further that the object penetrated into the hot region.

Efforts are continuing to improve the automatic processing of observations affected by pointing drift from observation at hot attitudes, particularly those from early in the mission before the importance of the effect was recognised. The treatment is complicated and still requires some work before it can be implemented.