CXC response to CUC Recommendations following 2010 April CUC Meeting
> DISCUSSIONS and RECOMMENDATIONS
> 1. RPS. A number of CUC members and their colleagues have
> expressed some concern about the RPS system, specifically the
> e-mail interface for handling proposals with more than five
> targets. It appears that the RPS system, and this aspect of it
> in particular, has not been updated since Chandra's launch. The
> e-mail interface seems to be vulnerable to hidden special
> characters that are inserted by modern e-mail handling
> applications, but that RPS cannot strip out. Given the fact
> that the vast majority of Chandra proposers appear to be happy
> and familiar with RPS (including some users of the e-mail
> interface) we are reluctant to recommend any big changes.
> RECOMMENDATION Perhaps the best compromise would be to offer
> users a simplified front end to RPS that can fill in all fields
> required in the e-mail version. We are thinking of proposals
> with target lists where the only change from one target to the
> next will be in the coordinates and the count rate, with all
> other instrument parameters the same, where the user would be
> asked to upload a target list to clone RPS forms. It might also
> be worth looking at whether it is reasonable to reduce the
> number of supported ACIS modes in such circumstances to ease
Some email systems (those that encode text as quoted-printable)
were breaking lines in emailed target forms. We have repaired
this by identifying and decoding these using a perl script.
In addition, RPS has been modified to allow line splitting.
We are testing a text box on the RPS 'Add Target' webpage to
allow the user to input target information when cloning a target.
It's a very simple interface accepting comma-separated list of
RA,Dec,Count Rate, Observing Time, and Target Name. We are also
testing an 'RPS Email' button which will convert the web form to email
for those multi-target proposers who start with the web and then
decide it's too slow for them. These should all be implemented
for Cycle 13.
> 2. Calibration. The CUC remains a little confused about the
> precise status of the Chandra calibration, especially in
> comparison to other missions like XMM. We attempt to summarize
> as follows:
> The comparison between ACIS, PN and MOS gives ACIS fluxes
> higher than those for PN by ~10%, and MOS fluxes higher than PN
> by ~4-7%. Above 2 keV, these three instruments give the same
> plasma temperatures. Below 2 keV, ACIS can still be fit with a
> single temperature, although the 0.5-2 and 2-7 keV temperatures
> are inconsistent by about 10%. PN and MOS however both need 2
> temperature component to give acceptable fits to spectra from
> what are presumed to be isothermal sources (outer regions of
> galaxy clusters). The soft band in XMM generally gives a lower
> temperature than Chandra, indicating a soft excess, or a hard
> deficit, in XMM relative to Chandra.
> The comparison of spectral line fluxes (O VII, O VIII, Ne IX,
> Ne X) indicates higher fluxes at 1 keV and below with Chandra.
> A series of plots of blazar grating observations during the XMM
> and Chandra missions indicated the improved calibration in a
> series of energy bands, with the Chandra flux relative to XMM
> increasing slightly with energy.
> Is this accurate? Ideally we would have liked to see
> the temperature and flux for both the ACIS/pn hard band
> and the ACIS/pn soft band.
The EO102 data show that the ACIS-S3 flux at 0.5 keV is
in good agreement with that derived from the MOS and PN data.
At 1.0 keV, the ACIS-S3 flux is approximately 5% higher than
the MOS flux and 10% higher than the PN flux. This slope in the
flux ratio between the different detectors below 1.0 keV
is responsible for producing the different cluster temperatures
in the 0.5-2.0 keV energy band.
The series of AGN plots presented at the CUC meeting were shown
with an older version of the HRC-S QE. A new version was
released about one week before the meeting, and was mentioned during
the CAL talk. With the new HRC-S QE, the fluxes from the Chandra
gratings and XMM-Newton detectors and gratings are in good agreement
below 0.5 keV.
Thus, the following general statements can be made regarding
the flux ratios in different energy bands.
1) Below 0.5 keV, the flux ratios between Chandra and XMM-Newton
are in good agreement.
2) Above 1 or 2 keV, there is an energy-independent off-set in
flux ratios with
ACIS(flux)/PN(flux) = 1.1
3) Between 0.5 and 1 or 2 keV, the flux ratios are energy-dependent
and the exact slope of the energy dependence depends on which
detectors are being compared.
While we caution the CUC that this is still an active area of
investigation, we believe that this represents a fair assessment of
where we stand today.
> The CXC's ongoing efforts to 'scrub' its web pages to remove
> obsolete and/or confusing documentation (which the CUC
> applauds!) should focus on the Cal group's areas. These are in
> particular need of updating and reorganization. Such
> update/reorganization should be aimed at clearly identifying
> areas that are well-understood and stable (e.g., CTI correction
> for all modes except CC graded) vs. areas that are 'under
> study' and in flux (mission cross-calibration work; ACIS filter
> contamination; CC graded mode), and should clarify the status
> of the latter. For example, the various memos concerning ACIS
> contamination could be 'bundled together,' and the obsolete
> documents among them should be clearly labeled as such.
The calibration group agrees with the need to bundle together
some of the recent calibration efforts into summary papers or memos
for the general user and to post these on the calibration web pages.
In particular we plan to organize memos, technical notes and papers in
sections, so they can be more easily and readily accessible.
Currently we plan to divide them into:
- What's new (the latest calibration memos in final form)
- In progress (memos on work in progress or under study)
- Archival (all the valid calibration memos)
- Obsolete (old and superseded calibration memos - for reference only)
We also plan on publishing some of these memos as summary papers in
> We are pleased to hear of cross calibration papers nearing
> readiness for submission. We also support the ongoing efforts
> in calibration of the ACIS CC mode.
We have submitted to A&A our papers on the use of clusters of
galaxies, and thepulsar-wind nebula G21.5-0.9 to cross-calibrate
between Chandra and XMM-Newton. These papers will be posted on astro-ph
once accepted for publication. We will also show some of these
at the next CUC meeting, including flux ratios in different energy
> RECOMMENDATION Calibration of ACIS-HETG (including CC mode)
> should aim for a goal of 5% between the various orders in MEG
> and HEG. This is driven by the need to accurately measure
> continua in a variety of applications from spectra of accreting
> binaries (black holes, neutron stars, white dwarfs) and AGN. In
> particular the measurement of absorption edges, existence/contribution
> of thermal bremsstrahlung components, and fitting the red wing of the
> Fe K alpha line from accreting black holes all depend critically on fixing the
> continuum level. To a lesser extent, the interpretation of
> emission line spectra from neutron star and black hole binaries
> also place demands on the calibration, since the dominant lines
> come from H-like ions for which the atomic data are accurate at
> this level. Other coronal sources which depend more on a wider
> selection of spectral line and less on the continuum do not
> stress the calibration so much, because atomic data used in
> interpretation is less accurate.
The HEG/MEG calibration needs to be broken up into several
different observing modes.
1) TE mode
The AGN gratings plots presented at the CUC meeting show
that the HEG/MEG cross-calibration is already good to
approximately 5% in TE mode. This mode comprises the
bulk of archived HETG data.
2) CC-F mode with the new set of telemetered grades
Starting at the end of 2009, a new SI mode was created
for CC mode data that telemeters more flight grades then
before. This is required since CTI-induced grade morphing
is different in CC mode compared to TE mode. Approximately
10% of good events were left on board before this change
in the telemetered grade set. We have only obtained a few
data sets in this new SI mode at the present time and these data
are being analyzed with a test version of a CTI-correction
algorithm for CC-F mode data. We will give a status report
on the HEG/MEG flux cross-calibration in CC-F mode with
the new set of telemetered grades at the next CUC meeting
3) CC-F mode with the old set of telemetered grades
To improve the CC mode HEG/MEG cross-calibration with the
old set of telemetered grades will require the generation
of a separate QE file for these data. We will give a status
report at the next calibration report on the development
of this new QE table and the resulting implications for
4) CC-G mode for sources with fluxes less than about 100 mCrab.
As shown at the CUC meeting, the fluxes derived from the 4
arms of HETG spectra (i.e., the plus and minus first orders of
the HEG and MEG) are in reasonably good agreement for sources
fainter than about 100 mCrab. The remaining calibration issue
with these data is to derive a method for applying CTI-corrections
to graded CC mode data. A different algorithm must be developed
for CC-G mode data since the full charge distribution around each
event is not telemetered in graded mode observations.
5) CC-G mode for sources with fluxes greater than about 100 mCrab.
The plots presented at the CUC meeting showed that the derived fluxes
for the four arms of HETG spectra begin to diverge for sources
brighter than about 100 mCrab. A recent analysis shows that the
distribution of photons tends to broaden with increasing flux in
(m_lambda, PI) space. Using the standard OSIP (order separating and
integrated probability) file, which is applied in (mlambda, PI)
space, tends to discard an increasing number of photons with
increasing flux. So, in addition to a new algorithm for applying
CTI-corrections, further improvements in CC-G mode calibration may
require flux-dependent OSIP files.
> 3. CDFS Archive. The CUC are mildly concerned about the reduced
> oversubscription in archive proposals to study the CDFS,
> compared with other archive and theory proposals. Of course
> there is no way to know a priori what will happen in response
> to the AO.
> RECOMMENDATION For the time being we just comment that in the
> peer review, we hope that CDFS archive proposals will be held
> to the same standard for funding as other archive proposals.
The 14 CDFS proposals were assessed by the topical panels at the peer
review according to their subject area. They were graded according to
science on the same scale as all other proposals in the topical
panels. A CDFS merging panel was convened on Wednesday evening,
following the topical panel sessions, to combine the results into a
single, prioritized list. After the review, this list was used to
decide on the approved proposals based on the available budget of
$500K. With approval from the CXC manager, we increased the budget to
$532K in order to include the 7th proposal on this list which
otherwise would have been only partially funded. With this increase in
the budget, the final over-subscription was 2.6.
Similarly, the archive proposals were discussed and graded on the same
scale as all the other proposals by the topical panels. In this case
the merging was done by the CXC after the review, based on normalized
grade across all the panels. 17 proposals were approved with the total
budget being increased slightly to allow for a good balance across all
panels. In this case the over-subscription was 3.2.
Comparing the two, one approved CDFS proposal had a grade which was
significantly below that of the lowest accepted archive
proposal. Given the importance of the CDFS data and the support of the
merging panel for including this proposal to yield a balanced science
program, we decided to include this one with a small increase in the
CDFS budget. There remained 3 CDFS and 14 archive proposals with
passing grades that could not be approved due to lack of funds, which
is typical for the Chandra reviews.
> 4. One/Two CUC meetings/year? As the Chandra mission matures,
> the necessity to have CUC meetings twice a year has been
> RECOMMENDATION The CUC agrees that the meeting schedule could
> reasonably go to once a year, preferably in October after the
> peer review. We also agree that at each meeting, a time in
> April for a telecon should be set aside in case any issues
> arise that need discussion. Note that the telecon need not
> happen, but the CUC should be prepared in case it does become
CDO agrees as well that one meeting will most likely be sufficient,
with added benefits both for the Chandra and global carbon budgets.
The next meeting has been scheduled for 2010, Oct 25-26.
> 5. Catalog. We are impressed with the ongoing progress. Our
> discussions prompted a few further questions:
> What is likely to happen to the source catalog when the
> Chandra mission ends? Is CIAO envisaged to be an interface to
> the catalog?
The catalog is considered to be a part of the archive, and therefore
responsibility to provide continued access to the catalog will fall to
HEASARC after the Chandra mission ends. We expect that all catalog
interfaces will be 100% IVOA-compliant. We already provide support for
IVOA standards such as Simple Cone Search and SAMP messaging.
Back-end catalog interfaces are being migrated to IVOA standard
protocols (e.g., TAP) as such standards are defined. Catalog FITS
file-based data products (eg, event files, PHA spectra, RMF, etc,) are
HEASARC/OGIP compliant and can therefore, in principle, be analyzed by
any software that conforms to those standards. The CIAO data analysis
package is particularly well-suited for user analysis of Chandra
Source Catalog data products, although it is not a catalog interface
per se. Post-mission support of CIAO tools that are particularly
relevant to catalog users will be evaluated as part of the overall
long-term CIAO plan.
> RECOMMENDATION/QUESTIONS Although solid progress is being made
> on the CSC, we remain concerned about the CSC's visibility in
> the wider astronomical community. We recommend that (1) CSC
> releases be announced via the AAS; and (2) when the manuscript
> defining the CSC is accepted, it be posted to astro-ph (we note
> with satisfaction that the accepted version of the Evans et al
> paper on the Chandra Source Catalog has been posted at
> http://arxiv.org/abs/1005.4665). We would also like to see a
> batch search engine.
We would also like to enhance the visibility of the CSC in the wider
community. Announcements of future releases (including forthcoming
release 1.1) will be submitted to the AAS for inclusion in a AAS news
bulletin, as well as being announced via current Chandra-specific
channels such as the Chandra bulletin and the chandra-users email list.
As noted the CSC description paper has been posted on astro-ph, and has
now been published in ApJS. See Evans, I. N., et al. 2010, ApJS, 189,
We currently provide the command-line interface, which can be scripted
for batch queries. In addition, VO-compliant tools such as TOPCAT can
be used to perform batch position searches via the IVOA simple cone
search interface. We are actively working and have made very good
progress on adding a position cross-match capability for CSCview, which
will allow a user to upload a list of positions (and optionally,
position errors) to use as search criteria.
> 6. Low Energy Calibration. The CUC is pleased with the
> progress that has been made.
> RECOMMENDATION It would be desirable to have a physical model
> for the "gaussium" component, not just for the intellectual
> satisfaction, but to be able to predict how the contamination
> might behave in the future. We recognize that this is asking a
> lot of the resources of the CXC, and that the main aim of
> providing a calibration that observers can use has basically
> been met.
We agree with this goal. We are in the process of collecting data
which might shed more light on the issue. These include
LETG ACIS-I data and the fine structure of the C-K edge, as well
as monitoring of ACIS-S observations of the M87 jet away from the
knots, which provides a sufficiently bright (but not piled-up),
non-variable, power-law calibration source. However, our confidence
that these analyses will provide a definitive resolution to the issue
> 7. Senior Review. Second place to Planck in a review with the
> metric science/NASA$ is a very satisfactory outcome, and we
> commend Harvey and the team on a job well done.
Thanks from the CXC!