Minutes of AXAF Users' Committee Meeting; July 23-24, 1998

Committee members present: Keith Arnaud (for Rich Kelly), Jill Bechtold, Webster Cash, JP Caillault, Lynn Cominsky, Megan Donahue, Steve Kahn, Shri Kulkarni, Craig Sarazin, Jane Turner, Mike Watson, and Ray White.

Ex-officio members present: Fred Seward and Allyn Tennant.

NASA HQ personnel present: Paul Hertz

Status of Project

Harvey Tananbaum, ASC Director, reported on the status of AXAF. Due to problems with the spacecraft assembly at TRW, the launch was initially delayed to December 3, 1998. Subsequent problems with the shuttle launch schedule and the assembly of Space Station have led to a further delay until January 21, 1999. However, it is possible that AXAF could still be launched on December 3, 1998 if the schedule of Space Station changes.

During the thermal vacuum test, the door to ACIS could not be opened due to the failure of an actuator. The nature of the failure is not completely understood. Although the door only needs to be opened once in orbit, it must open to use ACIS-I, ACIS-S, or HETG, which account for a large fraction of the AXAF observations. This is a serious problem which might affect the schedule for the AXAF launch.

During tests, some heaters on the Science Instrument Module failed. There were also some problems with spurious commands, and with the controls for the reaction wheels.

The contest to name AXAF produced over 6000 entries, with most coming from the public. A committee, including media people, has been assembled to recommend a choice by September 1998.

AO-1 Proposal Review Process

Fred Seward (ASC) reviewed the AO-1 proposal review and selection process, The statistics of received and accepted proposals were discussed. A list of improvements to the review process had been sent to the Committee prior to the meeting, and the Committee generally supported the suggestions for changes.

There was a discussion of methods of increasing the coordination between review panels covering similar science areas. The Committee recommended that the chairs of these panels meet on the morning of the second day of the review to compare coverage of science topics and overlaps. The Committee also suggested that each panel be given the titles, abstracts, and targets for all proposals from other panels, and access to information on previously accepted observations.

Problems which were encountered with completing the written reviews were discussed, and suggestions for speeding up the writing were considered. A "check off" of a few standard reasons for rejection, such as "a good proposal but a better one was selected for the same target(s)" or "not technically feasible" should be included on the forms, in addition to space for more detailed, individual comments.

The composition of the review panels was discussed in general terms. The panel expressed a desire to see an increased level of international participation in the review. Partly to insure that the prior human technical reviews by the ASC were not the primary assessment of technical feasibility, the Committee recommended that the reviews include a higher proportion of panelists with detailed technical expertise and a knowledge of X-ray astronomy. In the AO-1 review, due to a policy decision by NASA and the ASC, there were no panelists from the ASC staff, from the GTO teams, or from the Project Science staff. In votes without dissent, the Committee recommended that people from all three of these groups be allowed to participate as panelist in future reviews. However, the Committee felt that, if possible, no more than one person from any given institution should serve on any one panel.

There was a suggestion that the review panels merely assign time to accepted proposals, and that the PIs be allowed to choose the targets from among those not previously assigned. The Committee did not feel this proposal was either practical or beneficial.

Paul Hertz (NASA HQ) asked whether review panels should be told to avoid removing targets for proposals. The Committee did not support such a restriction, but preferred the present policy where the removal of targets and even changes in the exposure time was allowed.

Harvey Tananbaum asked whether the current definition of conflicts between proposals was adequate. The current definition is that a conflict occurs when two or more proposals wish to observe the same target with the same instrument for comparable times. The definition might be extended to observations of the same target with similar instruments (ACIS and HRC images, or LETG and HETG spectra). The Committee felt that the current definition should be retained. There was also a discussion of "science conflicts" (i.e., different targets but the same scientific goals). The Committee felt that this was too ambiguous to be used as a definition of conflicts, and that these questions are best handled within the scientific review panels.

The Committee felt that longer reviews will be needed in the future, particularly as the merging process is expected to be more complicated because of Large Proposals.

Technical Reviews

Belinda Wilkes (ASC) described the technical reviews of the AO-1 proposals. These reviews included computer checks for inconsistencies and errors, and prior human technical reviews. Technical reviewers were also present to help the review panels. Several improvements to the technical reviews were suggested for AO-2; one of the most important was distributing the technical reviews to the panelist with the proposals, well before the panels meet.

The value of the prior human technical reviews was discussed. It was clear that many ASC staff members put a great deal of effort into these reviews. Belinda Wilkes and Fred Seward suggested that the required technical reviews might be eliminated. In the proposal which was discussed most extensively, there would be computer checks and technical reviewers would be present to help the panels at the review. However, prior human technical reviews would be done only for cases which were flagged by the computer review, or where a panelist requested the technical review.

After considerable discussion, the Committee was split on this proposal, and voted 7 to 5 to recommend retaining prior human technical reviews. However, the Committee said that these reviews should focus on flagging those proposals with serious technical concerns, rather than reproducing the technical justification in the proposals or simulating the observations.

The Committee also felt that it was important to make sure that the review panels understand that they are responsible for ensuring the optimum scientific program based not only on the scientific justification but also for the final evaluation of the technical feasibility.

Budget (Stage 2) Reviews

Fred Seward and Paul Hertz discussed the status of the nearly complete Stage 2 Budget Review for AO-1. Because of adequate funding for the analysis of AO-1 data and a significant international proportion of accepted proposals (which do not receive funding), the review has been straightforward. Paul Hertz suggested that it might be possible to simplify the budget review, by having it done by a committee at the ASC or by some simple formula. The Committee felt that the current Stage 2 review process should continue, at least through AO-2. However, to aid in determining budgets for approved proposals, the Committee suggested that the Stage 1 scientific review panels might provide a check-off indicating the level of effort required to accomplish the science goals of each accepted proposals, on three levels: more than average effort; average effort; or less than average effort compared to other accepted proposals.

Large Proposals

Previously, it had been decided that AXAF would have a policy to encourage Large Proposals starting with AO-2. This policy requires that at least 20% of the observing time go to proposals with at least 200 ksec. Based on the distribution of requested exposured times in AO-1, the Committee felt that the minimum total exposure time for a Large Proposal should be increased to 300 ksec.

The Committee discussed methods of evaluating the Large Proposals at the review. Harvey Tananbaum pointed out that a single Large Proposal will be a significant fraction of a single panel's assignable observing time; thus, it is difficult to treat them just like normal proposals. The Committee considered having a separate panel to judge Large Proposals, but felt that it might be difficult to create a panel with a sufficiently broad perspective. Instead, the Committee recommended that the Large Proposals be assigned to science panels, and be reviewed and graded on the same scale as the normal proposals. The final decisions on the Large Proposals would be made during the merging session after the individuals panels met.

The Committee recommended that the panels be assigned a "bogey total exposure time" of 70% of the prorated observing time. This reserves 20% of the time for Large Proposals, with an extra 10% for negotiating room in the merging session.

The Committee said that the individuals panels should have the option to delete targets from Large Proposals, perhaps to assign them to normal proposals. If this resulted in a Large Proposal falling below the 300 ksec limit, it could be judged as a normal proposal.

The Committee also considered whether the page limits for the Scientific Justification should be increased to 6 pages for Large Proposals, but narrowly voted to recommend no change in the page limits.

In-Flight Calibration

Larry David (ASC) presented a summary of the current plans for in-flight calibration. The calibration observations and data will go directly into the archive. Given the early access of the general astronomical community to these data, the committee suggested that a mechanism be established for public input on calibration data and re-calibration.

The committee was particularly interested in the plans for the "first light" science observations; at present, the suggestion is that this would be an image of Cas A. The committee suggested that considerable thought go into the choice of the first observations for public relations and their presentation.

Director's Discretionary Time

At the suggestion of Richard Mushotzky of the AXAF Science Working Group (SWG), the SWG had recommended that some fraction of the observing time be allocated to the ASC Director for use at his discretion, which is a common practice at other observatories. The Committee also received a letter from David Helfand (Columbia) supporting this idea. The Committee unanimously voted to recommend that up to 5% of the total observing time be reserved for ASC Director's Discretionary Time observations, beginning immediately. This allocation of time would include non-peer reviewed TOOs; currently, the Director is allowed to accept these for up to 3% of the observing time. The funding and proprietary period for Director's Discretionary Time observations would also be left to the discretion of the ASC Director.

The Committee asked that the ASC Director report on the use of the Director's Discretionary Time at future AXAF User's Committee meetings.

Coordination With Other Missions

Concern was expressed that a significant opportunity to enhance the scientific output of both AXAF and FUSE might have been lost by the lack of coordination. The FUSE GTO target list was not announced until after the AXAF proposal deadline, and the AXAF GO target list was not announced until after the FUSE proposal deadline. While it is difficult to coordinate various proposal dates among the large number of space observatories, the Committee felt that it was important for AXAF to announce in advance the date for the distribution of the approved target list, and then adhere to this schedule.

AXAF Fellows Program

Nancy Evans ASC) reported on the first year selection of AXAF Fellows. Because of the strong set of applicants, the selection panels suggested that the number of fellows be increased from the planned three to five, and the ASC Director agreed to this.

Data Distribution and the Public Archive

Arnold Rots (ASC) reported on the plans for the AXAF data formats, data products, data distribution, and the public archive. The data products will all be distributed and archived in FITS formats, compliant with the recommendations of the HEASARC FITS working group. As a result, the data should be compatible with FTOOLS, XSPEC, etc.

The normal scientific data products are divided into primary products, secondary products (which most observers are likely to use to start the analysis of their data), and supporting products (which include many of the housekeeping data files). The PIs of GO and GTO observations will receive the primary and secondary products (the supporting products are available electronically by request). Following completion of the data pipeline analysis, these proprietary data will first be made available through password-protected WWW access. Then, they will be sent to the PIs on physical media, which might be 4-mm DAT tape, 8-mm Exabyte tape, or CD-ROM. When the data become public, the data will be available electronically (physical distribution by request only and as time permits). A mirror site for the public archive is planned for Leicester.

The calibration data would be maintained in HEASARC/CALDB compatible files and directories.

The Committee requested that they be given simulated science data sets as soon as available (by September 1998). This would allow "real user" testing of compatibility with existing software.

Science Data Analysis

Dave Huenemoerder (ASC) reported on the plans for the data analysis pipeline processing of grating data. Because of differences between dispersed and nondispersed photons, an iterative aspect solution will be necessary. He also discussed techniques to use ACIS pulse height for order discrimination in HETG/METG spectra.

John Houck (ASC) discussed software which is being developed for the interactive scientific analysis of grating spectra. These include techniques for selecting, identifying, and characterizing lines. He discussed progress on the Emission Line Project to improve the atomic physics of X-ray line emission.

Martin Elvis (ASC) reviewed the overall status of the AXAF Science Data analysis software. Prior to the meeting, the Committee received the 22 June 1998 revision of the AXAF Project Data Management Plan. Martin Elvis emphasized the complexity of the spatial-spectral data sets which AXAF will produce. The priorities for software development are that the highest priority is for flight operations, mission planning, and data receipt. The second priority is the pipeline software to produce and distribute the standard science data products. The third priority is the development of science analysis tools. Some of the software tools are simple adaptations of existing tools, but areas which represent new challenges include grating spectroscopy, combined spatial-spectral analysis of data, and source detection in complex fields with a wide range of source surface brightnesses.

Ultimately, the data analysis software would be accessible as stand-alone FTOOLS from a Unix prompt, from within an integrated GUI, and as "tasks" within IRAF. The ASCFIT program will incorporate the XSPEC library of spectral models as well as spatial models, and employ a powerful model algebra to combine models.

The Committee asked about the progress on software to deal with the effects of ACIS pile-up. Martin Elvis discussed some ad hoc interim solutions, and described an effort involving the Harvard Statistic Department to develop more general approaches. The Committee recommended that the ASC hold a short workshop on exploring solutions to the pile-up problem.

At present, the ASC staff is working towards a 1 November 1998 release (the GTO release) of portions of the science data analysis system. This would include the software for the pipeline production of the standard science data products, software for producing instrument response files, software for source detection, a portion of the ASCFIT program for spatial-spectral data fitting, and some data model tools. This first release will be in the form of Solaris and Linux binary executables. A 15 December 1998 release (the GO release) should include the First Look GUI interface to the data analysis programs, some of the interactive spectral analysis software, some integral transform programs, version 2 of ASCFIT, and more data model tools.

The Committee was concerned that the release dates for the software may not allow enough time for user testing and input. The Committee felt that it was vital that they have a chance to try the software as soon as possible. A schedule for testing of the software by the Users' Committee was determined, with access by physical visits by committee members or their colleagues to the ASC in August 1998, by remote logins to the ASC in September 1998, and by the distribution of GTO release of binary executables of the software in Solaris and Linux by October 1998.

The Committee felt that the status of the software was very important, and decided to schedule a special Committee meeting in late fall 1998 specifically to review the results of testing the GTO software release.

The Committee also expressed some concern with the lack of responsiveness to reported problems with the AO-1 proposal preparation software. The Committee asked that a plan to improve this software be presented at the next meeting.