Split-Night Policies

Justification

Support of split nights presents a distinct challenge to Observing Support. Normally Support Astronomers are available to help observers starting around 1 p.m. In order to properly support second-half observers they may be required to stay until 2 a.m. This long, 13 hour day is taxing on the SAs and prevents them from accomplishing more on other important Observatory work (e.g. Web page and infrastructure improvements). With some modest coordination of effort on the part of the split-night observing teams, we feel that this impact can be minimized.

Policy

• For configurable instruments (mostly LRIS and DEIMOS), the two PIs must coordinate a SINGLE configuration request. The configuration forms will be modified to allow entry of BOTH PI names. Both names should be filled out regardless of which PI submits the form. There exists a LRIS policy on this. The overriding reason for this is that the instruments cannot be reconfigured at night, so both teams must share the limited number of slots available in the instrument.
• If possible, both teams should be available for a single training session by the SA in the afternoon. Alternatively, if afternoon training is not required, the second-half team can refresh their memory of how to operate the instrument by sitting in on some of the first-half observing.
• If possible, one or more observer from either team is encouraged to provide support during the handover and subsequent startup phase. This normally would include the first slitmask alignment. For example, a first-half observer could agree to stay up to help the second-half observers make the transition to their observing directory, etc., then make sure that they understand how to operate the instrument, and wait until they get on to their first object. Alternatively, a second-half observer could participate during part of the first half of observing, at least at the level of monitoring the instrument operation.
• Of course, if the two observing teams cannot make suitable arrangements, the Support Astronomer will be available to both of the teams. Remember, we do not want to create any awkwardness or other problems for either observing team. We are only searching for ways of maximizing the time we can spend on other important tasks that will benefit the observers in other ways.

When does the split occur?

Nominally the night is split at the midpoint between 12° twilights (which is almost the same as the midpoint between sunset and sunrise, 18° twilights, etc.). This time is displayed on the telescope schedule by holding your mouse over the date in the second column. A pop-up should show the sunset, midpoint, and sunrise times.

This implies that the first half observers should be done their final exposure at this time, and the second half observers and OAs should start switching to the new program and/or instrument, as appropriate.

Note that there are many reasons to argue for a switchover time either before or after the midpoint of the night: LGS-AO checkout takes time only at the start of the night; lost time tends to be at the start of the night; switching instruments includes overhead that might be shared between the groups; (the former two arguing for a switchover later than the midpoint, the latter arguing for one earlier). To keep it simple, we have chosen to simply use the midpoint for all instrument combinations.

Nonstandard switchover times

On occasion a TAC will award less than a half night to one observer and more than a half night to another, and pair them up. The TACs should let each observer know when the split is to occur. WMKO will not be responsible for arbitrating disputes over when the split should occur, although for both standard and nonstandard split times, observers are free to work out between themselves their own handover time. It is useful for observers to communicate such arrangements to their support staff, both OAs and SAs, so that the support staff can manage their time more effectively.

Protecting Your Data

The Keck Observatory Archive contains data from all instruments. This data is kept private for a certain proprietary period, nominally 18 months. (Observers can request extensions; see the KOA Web site for further details.) On nights with only one PI, assigning data to the correct PI and observing program is easy. On split nights, it can be harder. You can help protect your own data by setting the OBSERVER keyword when you are taking data. Ideally, include the PI's name in the keyword, even if the PI is not present. This will allow the automated ingest systems of the archive to properly associate your data with your observing program.

A GUI is being developed that will aid this, allowing the observer to change observer name, and possibly output directory (if the two observing teams are using different directories) with a single click.

Mahalo for your help,

Randy Campbell
Observing Support Manager