Transcript presentation source

Discussion - Survey Design
• Survey product equation:
#fields = fld/nt x useable x (%xnights/yr) x years
= 4 x 0.5 x (0.75 x 13 x 18) x 3
= 4 x 0.5 x
175
x 3
= 1050
• Fields to be observed = 1360 - 197 = 1163
(difference is an extra 4 months of observing)
• Duration: 1360 fields will take until end 2005
• How can this timescale be shortened?
Survey Design (2)
• More fields / night?
– Could we achieve 5 flds/nt? (shorter exposures? quicker
change-overs?) But stress on observers, quality control?
• Higher useable fraction?
– El Nino! If 0.5 is 0.6 then 1260 fields done in 3 years
– Fewer instrumental problems would also
be...instrumental
• More nights per year?
– Higher fraction of time? (85% allocation would do it)
• Longer duration?
– Perhaps possible, but scientifically undesirable
Survey Design (3)
• Current bottlenecks…
– Partial/missing input samples: what is the deadline for
including samples in the survey (very soon surely?)
– Morphological classifications: NIR photometry insufficient,
but we have SuperCosmos B + R, so can use optical/NIR
colours or even visual morphologies (who will do this?)
• Other issues…
– Standard stars for  measurements? Needed? If so,
which, how many, how to obtain?
– Order to observe fields: by Dec strip? Galactic latitude?
Discussion - Survey Operations
• Survey observations…
– Quality control: top priority! Need immediate checks on
spectral quality - S/N, PSF; what else? Who should do
this (observers? astronomers?) Should subjective
measures have names attached?
– Observing support: what are the needs for visiting
observers from the survey team? How best deployed?
• Survey reductions…
– Checks on 6dfdr: Careful examination of spectra to check
reductions is needed - who, how, when?
– Optimization of 6dfrunz: are needed - who, how, when?
Survey Operations (2)
• Survey dataflow (see chart)…
–
–
–
–
–
Where are the weak points in this plan?
Where are the bottlenecks?
Is information getting to where it’s needed?
Are the resources sufficient for each task?
What is missing?
Survey Operations (3)
• Current bottlenecks:
– Data reduction backlog: need all data to May 2002
reduced well before Early Data Release in Sep 2002;
can this be done by under current arrangements? Are
extra resources needed?
– Redshifting backlog: at present ~1hr/field, 150 fields to
do, ~3 people working on it volunteers welcome!);
tuning of 6dfrunz (esp. splicing V+R spectra) would
greatly reduce time/field; reasonable to expect 15min/fld
– Velocity dispersion: pipeline still to be written, much work
still to do (dispersions NOT available for the EDR).
Discussion - Survey Database
• Database content…
– Redshift survey: all input source catalogue parameters,
spectroscopic observation parameters, postage stamps
in as many bands as possible, spectra (and variance
array), redshifts and associated info…
– Velocity survey: Dn- parameters, distance, pec.vel., etc.
– What additional parameters should be included? (what
about useful/rationalized subsets? what about federation
with external datasets?)
– Need to deal with null values in catalogue parameters.
Survey Database (2)
• Database interface…
– Does the 2dFGRS database provide a suitable model?
(albeit with a number of functional improvements)
– Essentially a collection of FITS files with all information
on each object in headers/images/spectra plus an SQL
database of all parameters with a WWW interface and
multiple extraction methods (text, HTML, FITS)
– Serve images, spectra and user-specified subsets of
complete catalogue (queryable on all parameters?)
– What parameters should be indexed? (speed/storage
trade-off)
Survey Database (3)
• Practicalities…
– Location: at WFAU in Edinburgh (what about mirrors?)
– Manpower: what effort is available? is it sufficient?
– Data release: who will be responsible for this?
Discussion - Science Exploitation
• Data releases…
– Early Data Release in Sep 2002 (all input catalogues
plus limited set of spectra/redshifts for testing); should
this be limited to survey team and AT PIs, or fully public?
– Subsequent data releases: intention is to provide a
public release every 6 months; can we provide a shorter
timescale for AT PIs?
– Velocity survey data (distances, peculiar velocities) are
much more difficult, and will lag significantly - t.b.d.
– Format: releases (except final) will be WWW-only
Science Exploitation (2)
• Students and postdocs…
– Heath Jones, Lachlan Campbell, Tom Mauch, Daniel Burkey
– Do we need more? Are others expected?
• Avoiding/resolving scientific overlaps/conflicts (also
identifying productive collaborations/synergies)…
– Especially important for students/postdocs
– Communication the key! Survey participants need to spell
out proposed areas of research, identify potential overlaps
and negotiate solutions - how best to do this? (team email
exploder, ‘who is doing what’ WWW page, meetings…)
Science Exploitation (3)
• Other issues…
– Detection of optical Sne in 6dFGS spectra - can we do it
in real time? (perhaps include a SN template spectrum?)
– Detection of Galactic emission/absorption is common at
low |b|; is this of scientific interest? If so, what do we
have to do to make these detections useful?