The observatory schedule was defined by the database of accepted observations. The HSC carried out a careful study of the observation database to define a long-term mission plan that would accommodate all constraints and would maximise the scientific return. The Long Term Mission Planning tool developed at HSC was a very powerful aid to identifying potential future scheduling problems at a very early stage.

Following the agreed long term mission plan, short term observing schedules, together with the corresponding instrument commands, were produced with the Mission Planning System at the HSC, and transferred to the Mission Operations Centre (MOC), at ESOC. MOC added the satellite commands and produced the final detailed mission timeline that was uplinked to the spacecraft.

The basic time unit for Mission Planning was the Operational Day, or OD, defined as the interval of time between the start of two consecutive DTCPs. The DTCP, or Daily TeleCommunication Period, was the time interval when the spacecraft antenna was be pointed to the Earth to receive telecommands and send the recorded data. The duration of an OD was normally about 24 hours, but depended on the availability and detailed schedule of the New Norcia Ground Station, which was shared with other ESA missions; the longest OD during the mission was over 35 hours, although such cases were exceptional and occurred when there was a switch from using New Norcia to the back-up station at Cebreros.

The operational constraints on the Herschel instruments determine that only observations that use a particular sub-instrument were scheduled in a single OD. For sub-instruments that require cooler re-cycling, only observations with the cooled sub-instrument (e.g. PACS photometer) would be scheduled for the duration of the cooler hold time: two to two and a half consecutive ODs. Occasional exceptions were made to the single sub-instrument rule at the end of the mission, as it was often more efficient when the pool of targets was very small, to use the PACS spectrometer during the DTCP of HIFI days and accept the penalty for switching on and switching off a sub-instrument, than to try to fill the DTCP very inefficiently with HIFI observations.

For a wide range of reasons, from safety to calibration needs, the instrument assignation for each OD was standardised. It consisted of the repetition of 28 ODs, i.e. four weeks, during which the instruments followed one another and were used for a different number of consecutive ODs. The standard instrument distribution used during the routine phase is the one shown in Figure 7.1, which reflected the relative usage of each sub-instrument in the approved proposals. It would be revised according to the each instrument observing time pressure at different epochs of the year. This is what we call the "planning cycle", which is the foundation of most of the Ground Segment activities related to Mission Planning. This translates into an additional difficulty to accomodate observations that had been defined with timing or grouping constraints shorter than a few weeks.

As demand for SPIRE was relatively lower at the end of the mission, because many SPIRE programmes had been completed early, the default planning cycle for most of the final year of helium had SPIRE spectroscopy in one cycle and SPIRE photometry in the next. This allowed the SPIRE Instrument Control Centre to plan their calibration strategy knowing approximately when each instrument would be scheduled for several months in advance.

The SPIRE cooler hold time was very close to 48h, or two ODs, so both SPIRE photometry and spectroscopy were scheduled in two-day blocks. In contrast, the PACS cooler hold time was approximately 2.5ODs. Rather than waste precious helium, the third day after a PACS cooler re-cycle was always split between PACS photometry and spectroscopy. The expected hold time was calculated carefully and a contingency factor added to it so that if helium consumption were a little greater than expected, the final observation(s) would not be affected by temperature variations; from that amount of time after the end of the re-cycling, only spectroscopy observations would be scheduled in the OD.

Figure 7.1. The default Mission Planning Cycle that was the basis for scheduling in routine phase. Especially towards the end of the mission the Mission Planning cycle was a strong function of the sky distribution of targets as the remaining visibility and the need to complete scheduling before the end of helium was driving the Mission Planning Schedule.

When a Parallel cooler re-cycle was made, PACS would stay cool on the third day after the re-cycle even after the SPIRE cooler hold had run out. Thus the third day after a Parallel cooler re-cycle would also be split between PACS photometry (first half) and PACS spectroscopy (second half) as shown in the sample schedule in Figure 7.1.

Although Figure 7.1 shows the standard observing cycle, it was essentially the average assignation over a year. There was a twice-yearly observing window for the Galactic Centre and Orion: at these times of year, there was a far greater demand for PACS spectroscopy and HIFI observations, hence these instruments were more intensively scheduled. At other epochs when the Ecliptic Poles were visible, cosmological surveys dominated the telescope schedule and there was greater demand for PACS and SPIRE photometry, hence more time was scheduled with these observing modes and less time for spectroscopy.

With the end of the mission approaching towards the end of 2012 and the supply of observations in the database diminishing with time, the driver was the need to schedule all Priority 1 observations before helium exhaustion. This meant that there was considerable adaption of the instrument sub-schedule from month to month in order to complete scheduling. Some epochs were heavily dominated by a single instrument (for example, late summer and early autumn 2012 which was the last guaranteed visibility window for Orion and the Galactic Centre, thus these months were heavily dominated by HIFI to ensure that all outstanding observations were completed then, to avoid taking a risk that helium might end unexpectedly and leave some observations un-executed).

The use of the different HIFI bands was an additional constraint on the optimisation of observatory time as there was a significant overhead in switching between sub-bands that made it extremely inefficient to observe for short periods only in a particular sub-band. Very limited HIFI band changes were allowed in a given OD in normal circumstances, preferably no more than two. Therefore HIFI observations using different bands and with time constraints shorter than an OD were very difficult to schedule. Observations were scheduled to group sub-bands as much as possible in each OD and thus limit the need for band transitions.

As indicated when requesting the visibility window of an observation with HSpot, visibility was limited by the so-called "warm" attitudes, i.e. solar aspect angles between -30 and -15 degrees, in which the Sun warms the star tracker baseplate and the pointing accuracy may be de-graded (see Section 2.4). For this reason the scheduling of observations within this area was to be avoided where possible. Only in certain, justified occasions, were solar aspect angles between -20 and -30 degrees allowed for less than one hour, at the user's own risk and only when it could be demonstrated that it would not affect later observations, while potential scheduling at solar aspect angles from -15 to -20 degrees was decided strictly on a case-by-case basis.

During the DTCP, spacecraft poimting was highly restricted. At the end of each OD the spacecraft would slew to its DTCP attitude. This allowed its medium gain antenna to point towards Earth and make communication possible. To ensure that the signal was strong enough, the spacecraft attitude was constrained such that the antenna was always pointed within 15 degrees of the Earth. This restricted spacecraft pointings to an extremely limited area of the sky during the DTCP; filling the available science time in the DTCP efficiently with observations became a major task towards the end of the mission and each OD required careful study.

Normally cooler re-cyclings were placed within the DTCP (occasionally, for operational reasons, a cooler re-cycle had to be placed outside the DTCP, but such situations were rare). These provided a stable spacecraft attitude that could be used for spacecraft maintenance and housekeeping activities, with an absence of instrument commanding. Similarly, instrument health-checks, switch-on and daily set-up were carried out during the DTCP, as well as the daily momentum dump from the reaction wheels (the daily "SOPS" -- Spacecraft Operations -- window that was unavailable for scheduling.

However, any time in the DTCP that was not required for maintenence, house-keeping and calibration activities was available for scheduling science. Observing time within DTCPs was essentially a bonus. During a DTCP, only targets within 15 degrees of the DTCP attitude could be observed, putting very strong constraints on scheduling.

Two DTCP points were available "cold" and "warm". In the "warm" DTCP part of the available sky area was at warm attitudes, while the entirety of the "cold" DTCP was a cold attitudes, allowing the spacecraft to cool properly from any excursions into the warm area of the sky during the previous OD, with no limitations on the available area within the DTCP region. Over the course of the year the DTCP areas would pass from being a circle, to becoming increasingly elliptical and, finally, the two would, briefly, join. During the mission the "cold" DTCP attitude was the only one that was used.

Efficient scheduling of observations during the DTCP depended on the sky distribution of observations within the database. Towards the end of the mission it became increasingly difficult to fill DTCPs for some instruments so, on many occasions, to use telescope time more efficiently, PACS spectroscopy would be used in the DTCP of HIFI days, as described in Section 7.1.3.