|
|
EUROPEAN SOUTHERN OBSERVATORY |
| Re-Engineering Project: Astronomy Operation REVIEW LSO-PLA-ESO-90000-3 |
| Prepared | Olivier Hainaut | 2002-03-20 |
| Reviewed | Rene Mendez |
2002-04-25 |
| Released | Jorge Melnick | 2002- |
| 0.1 | 2002-03-10 first draft, ohainaut |
| 0.2 | 2002-04-02 2d draft, ohainaut |
| 0.9 | 2002-05-11 2d includes revision, ohainaut. Includes Review Discussion |
This document discusses astronomical aspects of the LSO Science Operation Department. It will eventually become part of a larger document describing the Science Operation Department as a whole.
This document is the result of the discussion that took place on Thu. 2002 Mar. 18 at la Silla and Vitacura. Were present at that meeting, in Vitacura, Emmanuella Pompei, John Pritchard, Leonardo Vanzi, Malvina Billeres, Michael Sterzik, Rene Mendez, and in La Silla, Lisa Germany, Martin Kurster and Olivier Hainaut. Gaspare Lo Curto had sent some written input.
In this document, "he" and "his" refer to the position described. In practice,
these positions (LVA) can be occupied by people of any gender.
| DMD | Data Management (and Operation) Division |
| MPE | Max Planck Institute |
| MTS | Medium Term Scheduling |
| OB | Observation Block |
| OPC | Observation Program Committee |
| OpA | Operation Astronomer, aka 80/20 Astronomer |
| P2PP | Phase II Preparation Package |
| QC | Quality Control |
| QC | Quality Control |
| SM | Service Mode |
| USG | User's Support Group |
| VA | Visiting Astronomer |
| VM | Visitor Mode |
| VisAs | Visiting Astronomers Department |
| WFI | Wide Field Imager |
The tasks and duties of the La Silla SciOp Astronomers are
COMMENT: KBA: I do not agree that the role of duty astronomers should include support to process the data of a visiting astronomer (VA) after their run. This can easily be abused by the VA to the extent that the duty astronomer has to reduce the entire dataset or provide lessons using packages such as IDL or midas. It also has the potential to extend the support to more than the duration of the VA run. Certainly the duty astronomer should ensure that the VA can analyze the quality of their data in real time (ie via pipeline procedures) and of course the VA can seek advice from the duty astronomer if there are problems with the data - but what the astronomer does with their data after it haas been taken is their business and is not of concern to the duty
astronomer."
REPLY: RME: I agree with this concern, in principle. On the other hand, if we
are using these type of data for our own science (are we?), we should
also be able to provide tips and generic input to the VA. Could you
please make this point crystal clear, Oli?
CHANGES: Right. Support astronomer do not reduce the data, he
- he gives info (text changed, cf below)
- he prepares tools to analyse the calibration plan and scientific data (i.e. the routines that would build a Pipeline)- cf below
(hints/tips to process their data, etc -but not to process the data for them).
It is considered desirable that, in the case of VM, the Phase II (P2PP,
strategy, etc) and the night support are provided by the same person, i.e.
as done at present. Indeed, having these tasks split
between 2 persons would cause a psychological chock on the VA...
COMMENT: LVA: the point of the psychological shock is at least weird I would
rather say that for continuity and to optimize the communication level
it is desirable that the two tasks are not split..
REPLY: RME: I agree with Leo's point. It looks more
professional the way he phrased it, and indeed that is exactly what we mean....
TEXT CHANGED:
It is considered desirable that, in the case of VM, the Phase II
(P2PP, strategy, etc) and the night support are provided by the same
person, i.e. as done at present, in order to have continuity and to
optimize the communication level. However, it should be noted that
some of the Telescope Introduction could possibly be delegated to the
TIO (e.g. tour of the telescope). This has to be experimented with.
The possibility of having the Phase II preparation
could take place in Vitacura, e.g. in cases where that would permit the support
astronomer to come later on the mountain.
COMMENT: LVA: previous sentence is gramatically wrong
REPLY: fixed and improved
COMMENT: KBR: I am skeptical about the success of Phase II preparation
taking place in Vitacura. I see potential problems with space - a designated
area and computer needs to be provided. Furthermore the flexibility of the
duty astronomer's time in Vitacura is threatened - the duty astronomer must
ensure they are at the office at a certain date and time. I strongly am in
favor of this mode remaining the exception.
REPLY RME: True. Emphasize that this will be the exception,
rather than the rule. Also, if the leave this door open, we should organize
a "VA
office" in Vitacura (along with PAO if they are interested?), with the proper
connectivity, p2pp, jot,.ip, etc, so that this can actually take place in
a proper setting.
CHANGE:
The Phase II preparation could take place in Vitacura, e.g. in cases where that would permit the support astronomer to come later on the mountain. This was experimented with success at the NTT. It should however remain an exception, the normal mode of operation being Phase II intro at La Silla. Indeed, the astronomer's time at Vitacura is for Science, but it is better to "lose" 2h of science for an introduction than to "lose" 2 days. The hardware and software set-up in Vitacura will need to be improved (i.e. have a "Visitor Computer" with P2PP and up-to-date instrument packages).
It is also stressed that the VAs should do their homeworks before the Phase II preparation, i.e. read the manuals and think about what they want to do. This will have to be made clear to the VAs before they arrive.
COMMENT: LVA:May be it is worth to add the figure of the "periodic
visiting astronomer" or "friend of La Silla". A VA could be defined
"friend of La Silla" at his third visit over a period of 3 or 4 years,
this person usually requires a very limited support if any at all and
make us save
energy."
REPLY: RME: While I like this idea I have 2 concerns:
1) This depends a lot on the VA, some of them are good at remembering how
to get around, others tend forget very quickly (specially those that have
observing runs at several observatories in one year...),a nd 2), Some VA
may feel that the
quality of service they are getting is not as good as if they were newcomers,
instituting thus a kind of double standard, which is dangerous. TEXT ADDED:
Even "experienced" observer will have to go through all the point of the standard introduction. The support astronomer
can of course go over some of these points very quickly if the VA is very
familiar with them, but the support astronomer has to make sure that nothing
is forgotten!
The possibility of (limited) "remote observing" or "remote
operation" from Vitacura is also considered. This would permit more
flexibility, e.g. in case of trouble-shooting, tests, and even
observations. This is to be experimented with.
Every astronomer (staff and OpA) should belong to one Instrument Force (cf [2]), and be able to support all instruments in that force (to the level of a good generic introduction), and be specialist (i.e. be able to provide in-depth support and good level of trouble-shooting) of at least one instrument. After some time at La Silla, the Astronomers should also be able to provide basic support on at least one instrument of another Instrument Force.
Most astronomers are likely to be at some point Instrument Scientist (cf [2]) of an instrument.
As stated in [1], it is expected from the Astronomers that they will
remain informed while being in Vitacura, by reading emails and
replying to urgent emails related to their duties, and perform/complete
urgent tasks related to their duties for SciOp. This should take a very
small fraction of their research time.
COMMENT: KBR: I have never been in support of an OpA position
(ie 80% support). Based on my previous experience, if one wants to remain
active in science then it is impossible to do this in the remaining 20%time.
One option is to utilize the compensation time, but this is not a happy and
healthy solution. Instead I would have preferred to have had 100% support
(150 nights per year) without the added pressure of having to produce science
in my compensation time. Therefore I would recommend "staff 50%" positions
for those astronomers who want to do science and "OpA 100%" positions for
those who don't. It is not fare to offer something in between."
REPLY: RME: Deep waters. I have heard from severalOPAs (and
some fellows!) that they would rather have 100% duty...What do we do about
this one Oli?
Text added:
ESO considers that a "support astronomer" should be an
"active astronomer". This is the justification of the 20% of the 80/20. Nevertheless,
ESO acknowledge that it hard/difficult/impossible to be a world-wide top
scientist with only 20% of one's time to do research. As a consequence, the performance appraisal of the
OpAs is based only on their functional work, not on their science (as opposed
to the 50/50 astronomers).
Every Fellow should be specialist (to the level of providing in-depth introduction and basic trouble-shooting) of one instrument, and be able to provide generic introduction on at least one (JME) another instrument of the same instrument Force.
COMMENT: GHA: "If I understand it correctly, Fellows won't
be instrument scientist anymore, but will more similar to the support
astronomers in Paranal?".
REPLY (RME): I think this should be clarified in
the docu. Indeed, you said: "...no - we'll try to make the doc more clear.
We will have 10 instruments and 12 astronomers (incl. Fellows), so everybody
will end up being instrument scientist (the 2 extra corresponding to people
in training)."
CHANGE: text added:
Most fellows are likely to be at some point Instrument
Scientist (cf [2]) of an instrument. Indeed, as La Silla will have 10 instruments
and 12 astronomers (inc. Fellows), their role will be crutial.
The contract of the Fellows include 80n of duties on the mountain, and
35 days of duties in Vitacura. These 35d are used to ensure the
continuity of their service to SciOp (e.g. read SciOp email, reply to
those concerning them directly, complete urgent tasks, etc.), and
perform background tasks such as documentation, analysis, etc. The
Fellows are not expected to keep detailed track of their time in
Vitacura, but experience (+ time record from some template fellows)
shows that this does not amount to 35d. The remainder is considered as
a compensation for possible over-run of their 80n quota on La Silla
(not to be abused, and to be kept at a small level, e.g. no more than
10%), required by emergency situations or schedule impossibilities.
Currently, the ESO Students are not included in the La Silla
Operation, although some experienced was gained from Students who
volunteered to work for the Teams. Students have no obligation (by
contract) to perform duties at the Observatories. It is also noted
that the Observatories are not a training camp so, if they work at La
Silla, the Observatory should get some return from that work.
COMMENT: DAL: Maybe you could remove the sentence in the first paragraph:
"It is also noted that the Observatories are not a training camp etc..."
which I find useless and out of place.
REPLY: OK, Rephrased
It must be noted that the most important goal of the
Observatory is to support the observing projects, not train students. It
is therefore important that the Observatory should get some return from the
resources invested in the students.
The current experience with students showed that involving them in long term projects is a failure: they don't come enough to La Silla, and they are not supposed to work in such project in Vitacura. On the other hand, they proved to provide very efficient help for very well defined tasks that can be completed in very short time (few hours to few days, to be competed before they finish their "turno"), i.e. "here-and-now" projects.
Based on this, it is proposed that the Students will be proposed to volunteer to perform a small amount of duties (of the order of 40n/yr). These duties will be of the "here-and-now" kind, i.e. should be completed at the end of their "turno". These will be decided on a trimester by trimester bases (i.e. each scheduling period). Once accepted (i.e. once the schedule is accepted and published), the Student are committed to perform the scheduled "turnos" and to follow the instructions they are given.
The implementation of this scheme will be negotiated with the relevant
instances (Science Division, etc), and the budget to cover their
travel and stay expenses will have to be secured for >2002. For 2002,
an arrangement will have to be found between the different parties
COMMENT: DAL: I agree with the content of the section about
students: No obligation for duties. But, if a student volunteers to have
some functional work (on a trimester basis), then she/he is committed to
do so.
For the support: in 2002, Science will continue to pay for the flights (max
4 per student). In 2003, the flights for "duties" should be on the LaSilla
budget.
Reply: ok, thx.
For 2002,
an arrangement has been found between the different parties: Science/Vitacura pais the trips, SciOp pais the stay.
Currently, the staff of SciOps includes 5 Staff Astronomers
(>=105n/yr), 4 OpAs (>=120n/yr, with some variation from contract to
contract) and 7 Fellows (=80n/yr). In the long run (>=2004), SciOp
will have 4 Staffs, 4 OpAs and 2 Fellows (assuming one of the
remaining Fellow position is converted in an OpA, which is the current
plan). This will ensure 980 man-night/yr, i.e. an average of 2.7
astronomer per night. It is therefore not possible to ensure
full coverage of 1 astronomer/night/telescope.
COMMENT: JME: The common control room should enter here. If you have it, why would you have 1Ast/night/telsc?
REPLY: Right. Text added.
In that framework, the Common Control Center will
play an important role: one astronomer will be able to support more than
one telescope, provided that
Currently, the OpAs have 1-yr "Paid Associate" contracts with no retirement benefits, with a duty level varying between 120 and 135 nights/yr.
The duty level should be uniformized for all new contracts, e.g to 125n/yr, which is the number currently used for scheduling purposes. Also, the OpA contracts should be as much as possible made similar to those of Paranal, i.e. 3yrs with retirement benefits. The implementation of this point will have to be negotiated with the relevant instances (Pers. and Co).
COMMENT: JME: we [upper management] are working on this
REPLY: Thx.
COMMENT: KBR: I have never been in support of an OpA position
(ie 80% support). Based on my previous experience, if one wants to remain
active in science then it is impossible to do this in the remaining 20%time.
One option is to utilize the compensation time, but this is not a happy and
healthy solution. Instead I would have preferred to have had 100% support
(150 nights per year) without the added pressure of having to produce science
in my compensation time. Therefore I would recommend "staff 50%" positions
for those astronomers who want to do science and "OpA 100%" positions for
those who don't. It is not fare to offer something in between."
REPLY: RME: Deep waters. I have heard from severalOPAs
(and some fellows!) that they would rather have 100% duty...What do we do
about this one Oli?
REPLY: OLI: ESO considers that a "support astronomer" should
be an "active astronomer". This is the justification of the 20% of the 80/20.
Nevertheless, ESO acknowledge that it hard/difficult/impossible to be a world-wide
top scientist with only 20%. As a consequence, the performance appraisal
of the OpAs is based only on their functional work, not on their science
(as opposed to the 50/50 astronomers).
In order to ensure the quality of the support, it will become a necessity to constrain the program schedule as follow
COMMENT: JBr: AS WAS ATTEMPTED FOR THE VLT WE CAN TRY TO RESTRICT TO THREE NIGHTS MINIMUM THE LENGTH OF A VM RUN. HOWEVER, WE CANNOT PREVENT THE OPC TO REDUCE THE TIME IF THE LIST OF PROPOSED TARGETS IS NOT CONVINCING. A 3-NIGHTS RESTRICTION WILL HAVE THE IMMEDIATE EFFECT TO INCREASE THE DEMAND FOR SM OBSERVATIONS. THIS IS WHAT HAPPENED AT THE VLT.
REPLY: Unfortunately, we cannot increase the fraction of service.
COMMENT: JBr: ALTHOUGH WE SHOULD INDEED REDUCE TO AN ABSOLUTE
MINIMUM SHORT RUNS. I DOUBT THAT WE CAN REALLY FORBID THEM.
REPLY: OK:
Discourage short runs: 1-night runs should either not be scheduled (c.f next section "Service Observing"), or reserved to extremely experienced observer, considering that it is possible that such run would not get support.
COMMENT: JBr: THE SCHEDULING OF ONE TELESCOPE IS ALREADY A DIFFICULT TASK IN VIEW OF ALL THE PARAMETERS THAT HAVE TO BE TAKEN INTO ACCOUNT. THE COMBINATION OF TWO OR THREE IS EVEN MORE COMPLEX. ADDING THE "INSTRUMENT FORCE" CONSTRAINT MAY SIMPLY BE TOO MUCH!
IT WILL, OF COURSE, ALWAYS TRIED TO OPTIMIZE THE OBSERVING SCHEDULES FOR THE VARIOUS TELESCOPES. HOWEVER, IT CANNOT BE GUARANTEED THAT TWO TELESCOPES WILL ALWAYS HAVE INSTRUMENTS FROM THE SAME "IF".THE REAL ISSUE THERE IS: WHAT HAS THE HIGHEST PRIORITY, SCIENCE OR TELESCOPE OPERATION? A POLICY DECISION IS OBVIOUSLY NEEDED.
REPLY: OK- I realize that more constrains will be difficult to accomodate. If "smart scheduling" (not implying that the current one is "dumb") is possible for the majority of the runs, we should be able to support them. About the last comment: I fully agree. This document describes what we can offer with what we have available. CHANGE:
Schedule the various telescopes so that at least two telescopes have instruments from the same Instrument Force mounted. As this adds one (difficult) constraint to the schedule, this will obviously not always be possible, but it should be true for the majority of the time.
The implementation will have to be negotiated with the relavant instances.
The following is considered as what SciOp can reasonably and
efficiently support in terms of Service Observations. Its
implementation for P>=70 will be negotiated with the relevant
instances
(OPC, VisAs and Co).
Visitor Mode: a Visiting Astronomer to perform his own observations on defined nights.
Flexible Observing: aka "queued observing", aka Q-mode: during some pre-defined nights, observations are selected for execution from a pool of pre-prepared OBs according to the weather conditions, instrument availability, priorities, etc.
Delegated Observing: on pre-defined nights, a specific
observing program is executed by SciOp staff. From the scheduling
point of view, this is similar to VM.
COMMENT: JBr: Good to have these clear definitions
REPLY: Thx
Currently, ~3-5% of the time is performed in Service Mode, exclusively as Delegated Observing. As the 3.6m does not receive support from DMD/USG for Phase II, QC and data distribution, this should remain at the ~5% level. As there is no instrument flexibility, and as the number of night in SM is very small, Flexible Scheduling is not an option, so the SM should be exclusively performed as Delegated Observing, and strictly be reserved to
In particular, programs of >= 2 nights should not be accepted in
Service Mode any more, since there is no justification for this to
happen. More specifically, if the PI does not want to come in person,
he should find an observer and not rely on SM.
COMMENT: JBR: I DO NOT QUITE UNDERSTAND THE STATEMENT: "PI DOES NOT
WANT TO COME IN PERSON"
REPLY: We currently have SM runs that have no good justification for
SM. The only reason is that the PI is "not absolutely needed at the telescope".
I rephrase:
More specifically, the justification that "an observer
from the proposal is not absolutely needed at the telescope" should not be
sufficient to grant Service Mode.
COMMENT: JBR: I BASICALLY AGREE, BUT WHAT DO WE DO WITH 2-NIGHTS RUNS?
THEY ARE TOO SHORT FOR VM AND TOO LONG FOR SM!
REPLY: Discourage them... If the proposers propose >=3n, and the
OPC does not cut too much, we should end up with less short runs than currently.
A small fraction of the time, about 5-10%, is performed in Service Mode. Thanks to the fact that, in general, all 3 NTT instruments are available at all times, this SM can be offered in full Q-Mode. Nevertheless, it should be strictly reserved to
The latter two will constitute the "filler" needed for when conditions are not exceptional (i.e. the majority of the time).
Programs of >= 2 nights with no specific constraints should not be accepted in Service Mode any more. More specifically, if the PI does not want to come in person, he should find an observer and not rely on SM.
See above:
More specifically, the justification that "an observer
from the proposal is not absolutely needed at the telescope" should not be
sufficient to grant Service Mode.
It must be noted that NTT receives no support from DMD for Phase
II, MTS, QC, and data distribution. This is fine provided that the
fraction of SM time remains <10%.
30% of the time belongs to MPE, 60% to ESO. The agreement with MPE is
that 100% of its time is on WFI in SM. It is expected that about 50%
of the ESO time will be scheduled with FEROS.
COMMENT: JBr: MPE IS GETTING THREE MONTHS PER YEAR, I.E. 25% TO MPE AND
75% TO ESO.
ACCORDING TO J. MELNICK, MPE IS ALSO ENTITLED TO USE FEROS.
COMMENT: JME: 1/3 FEROS.
REPLY: OK, thx, corrected:
Three months per yrs (i.e. 25%) belong MPE, 75% to ESO. MPE can use both WFI and FEROS, and agreement with MPE is
that 100% of its WFI time in SM. It is expected that about 1/3
of the ESO time will be scheduled with FEROS.