UKIDSS Newsletter No. 7

From: Steve Warren, Andy Lawrence

January 2003

Happy New Year to all UKIDSS Consortium members. Please read the list of work to be done below, followed by minutes of the Nov. 25th 2002 meeting.

Work to be done

1. Each survey working group to prepare a paper detailing their VST imaging goals, with a deadline of Feb 15.¹ [The UDS are planning to use VIMOS, and do not expect to submit a VST proposal.] This is our opportunity to set out the needs of UKIDSS for optical imaging in the S. Recall that a year ago the LAS prepared an outline VST proposal in response to the 3rd ESO WGS 'Call for ideas for future public imaging surveys'. The working groups are now in a position to make concrete proposals, with a detailed science case, and firm suggestions on depths/filters/areas. I suggest we use the old LAS proposal as a guide to layout. The emphasis should be on public utility of the data, particularly for maximising the scientific output of the VLT.
   ¹Submission date may be extended, subject to confirmation by ESO Working Group for Surveys.
   [Action on Survey Heads to coordinate input from Working Groups. All other interested consortium members to work through the Survey Heads.]
2. Decision by DXS on whether they wish to shift the 22h field. The CFHT Legacy Survey have reduced their Wide Synoptic survey to 3 fields: 02 18 00 -07 00 00, 08 54 00 -04 15 00, 14 17 54 +54 30 31. Only the first is a DXS field. However their Deep Synoptic survey includes a 1degx1deg field at 22 15 31.67 -17 44 5.7. Should DXS4 be centred there?
   [Action on DXS WG]
3. APEX, the ALMA prototype, is rumoured to be choosing fields of size a few sq. degs. Since the science goals are similar (particularly evolution of abundance of galaxy clusters), DXS and UDS should contact them and enlighten them.
   [Action on DXS, UDS WGs]
4. Propose sequential data release timetable for each of the surveys, separately, by end February (see minutes of Nov. 25 meeting, below, for notes of discussion on proprietary period).
   [Action on Survey WGs]
5. Each survey needs to create an implementation strategy paper. Each survey needs to think through in detail how the survey will get built up with time. At any time, given the observing conditions, which field+filter gets observed next e.g. does DXS build up depth or area? In what order are the filters done? In one night is it better to do JHK, or, instead J, H, K, on successive nights? What is done in non-photometric conditions, bad seeing, or bright sky? An elaborate decision tree needs to be created that determines the next observation to be made given the conditions, and the previous history of observations. A simulation needs to be created that demonstrates the feasibility of the strategy. All these details are required for the observation scheduler, and will feed into the overall survey simulator, which in turn will influence when the camera is on the telescope. In addition decisions on the dither/jitter sequences to be employed need to be made. Finally for surveys with more than two passes in each band, requirements for what needs to be stacked when, should be specified e.g. create stacks at end of every semester, and later stack the stacks, or create super stack from all frames completed by that time, once per year.
   
   I suggest the following timetable. The Working Groups each have a meeting and prepare a brief report by end of February, listing what they see as the relevant issues for their survey, and the outline strategy, and detailing the open questions that need to be answered before finalising the strategy. These outline reports to be circulated to the other WGs. Then a report structure to be agreed, and draft reports completed by end March, and final reports by end April.
   [Action on Survey WGs.]
6. Follow development of Survey Definition Tool, and interact with Martin Folger (ROE), testing beta versions.
   [Action on DXS WG]
7. Sign off on the science requirements for the pipeline and archive. Any final comments, by 31 Jan, to mike@ast.cam.ac.uk (pipeline) and nch@roe.ac.uk (archive), each copied to sjw4@ic.ac.uk.
   • For the pipeline the relevant documents can be found on the CASU WFCAM page (you need casu casu2 to get there). The two docs are the WFCAM Pipeline Design document (October 2002, seen before) and the list of parameters, which is new. The latter is the important one, containing the proposed list of image parameters, revised after the Nov. 25 meeting.
   • For the archive visit the WFAU WFCAM page, and the relevant document is the Science Requirements Document, which again you have seen before but is now in revised final form.
   These documents take into account the discussions at the Nov. 25 consortium meeting, and a subsequent meeting on Dec. 9 between CASU, WFAU, and the SS.
   [Action on all interested consortium members]
8. Prepare a report on the usefulness, cost, and feasibility of installing Skyprobe on UKIRT.
   [Action on SS]

Minutes of Nov. 25 meeting, Imperial College

1. Pipeline

The status of the pipeline was presented by Mike Irwin and Jim Lewis. The standard pipeline is the data reduction and object detection, plus measurement of parameters for each object. All subsequent analysis takes place through reference to the catalogues in the archive, and will in future be referred to as 'database-driven products' (previously 'advanced pipeline'). These include, inter alia, band merging, sophisticated stacking, difference imaging, low surface brightness objects. The timetable is for the standard pipeline to be ready by first light.

There are some uncertainties in the data reduction steps:

• remanence
• psf sampling/parameterisation/variability and effect of intra-pixel q.e. variation
• what dither (jitter) patterns will be used?
• what image parameters will be measured

There was a discussion of the list of image parameters to measure. The main modifications to the draft list were as follows:

• psf-fit magnitudes are required. Several options are possible: a) firstly detect peaks, and then fit psf at position of those peaks, b) as (a) but in the fit also allow centroids to be free parameter (relevant where components overlap), c) iterative multi psf decomposition of sources ('CLEAN' type algorithm). There was agreement that type b) fitting was preferred, and that type c) fitting would be a database-driven product provided at a later stage, subject to demand.
• add Sloan Petrosian mags
• add 2.5 R(Kron) mags
• add a large set of circular aperture mags with sufficiently close spacing
• Sextractor parameters were discussed, but no extra parameters from the Sextractor list were proposed
• deconvolved 2D exponential and de Vaucouleurs fits (as Sloan) were considered desirable, but not essential, and only to be considered if a simple solution (e.g. use the Sloan code) could be found.

The GPS wants a measurement of nebulosity. A more definite proposal will be put forward after service observations have been taken.

2. Archive

NCH explained plans for a phased release of archive software. V1.0 would be at first light, with V2.0 some 12 months later. V1.0 would include stitched individual tiles, object catalogues matched across bands, and would allow finder charts. The baseline is SDSS Skyserver. There will be a trial period where it can be tested on other datasets, before first light. It will have basic and advanced user options. V2.0 will have an enhanced GUI interface, and will include the possibility of creating superframes (across tile boundaries). Everyone agreed with this approach. [V3.0 is the grid-enabled version - too far ahead to concern UKIDSS.]

Points that came up:

• SDSS fluxes/Luptitudes for SDSS non detections were seen as a must, if at all reasonably possible, for V1.0. NCH to check with SDSS the way forward. [This is now stated as a requirement for V2.0, which means that it is a goal for V1.0, and the likelihood of it appearing in V1.0 will depend on the difficulty of implementation.]
• capability of federation with other (general) catalogues was seen as similarly desirable for V2.0 (e.g. GLIMPSE, ASTRO-F). The UDS foresee the need for multiwavelength queries - sort of precursor to AstroGrid.
• some sort of simple, general stacking option, and source extraction on stacked frame, is needed in V1.0, even if it is not the final solution.

3. Complementary optical imaging

We discussed complementary optical imaging needs, both in general, and ESO opportunities in particular. Joachim Krautter summarised the ESO schedule. Proposals for VST are due April 2004, but there is also an opportunity in April 2003, for example to use WFI. The plan to develop public surveys on VST is as follows. There will be a call for ideas very soon, with a deadline of Feb 15th, followed by an open workshop on public surveys in May 2003. The WGS then formulates a community proposal to submit to the OPC in April 2004. It was emphasised that ESO are keen to see proposals from the UKIDSS consortium, as we are seen as well organised, and they want to demonstrate demand in order to justify large slices of public survey time.

It was agreed that proposals should come from each UKIDSS WG as desired to meet the Feb 15 deadline.

4. Survey definition tool

Andy Adamson described the development of the survey definition tool software. There was little comment, suggesting it is fine so far. The software will be ready by first light, and AL reminded people that we have to get to grips with these tools soon to actually implement surveys, so we'd better be sure we are happy.

5. Proprietary period

Phil James summarised the Board's feeling that the propietary period should be shorter, as this will maximise the science done with UKIDSS, and keep us on our toes. Jim Dunlop suggested that for UDS the proprietary period should be longer. Several other people appreciated the general argument for a short proprietary period, but were nervous of being scooped by half-baked science from incomplete datasets, rather than by really good science. As we tried to analyse the situation, we realised a blanket statement was impossible, and was very sensitive to when the clock starts etc. What we should actually do is get each survey WG to propose a 'release plan'. This might involve for example a long delay to first release, and then frequent increments. We did however make one general decision, which was that the whole of the 2-yr-plan data should become public at T=3.0yrs.

6. Grant applications

AL described the worrying experiences with the Warren and Jameson applications, but people were largely unphased, and indeed mentioned some successes, such as Simpson's AF having a large UKIDSS component. The meeting re-affirmed the consortium's policy that this was outside our remit. AL proposed the idea that there could be a standard UKIDSS 'plug-in', but this wasn't popular. People felt that each case would be so different it should really be left to the proposers.

7. Skyprobe

JAC are prepared to take this seriously, but believe they probably have enough information to monitor photometricity anyway. The meeting discussed the issue briefly, but basically people didn't feel inclined to venture a strong opinion without having the pros and cons laid out before them clearly. Adamson suggested that Warren should formally ask him about the possibility, and he should think it through and make a written reply.