Minutes:2MASS Point Source Workshop

The goals of the meeting were:

1. to determine whether the current baseline processing approach and the sub-system algorithms are on track toward enabling the project to achieve its level 1 requirements for source positions, photometric limits and accuracy, completeness and reliability;
2. to discuss and define a raw-data simulator which will provide insight into the limitations of the processing algorithms under a wide variety of sampling, seeing, source crowding, etc. conditions;
3. to begin a discussion of prototype camera and other (higher sensitivity; higher angular resolution) observations which will challenge the current point-source extraction algorithms, and provide the basis for refining extended source algorithms (this issue will be discussed in greater detail both at the upcoming 16-17 February science team meeting, and by the galaxy "tiger team" meeting in early March);
4. to understand the extent to which current hardware (prototype camera; telescope; observing procedures) and software approaches limit positional and photometric accuracy for point sources;
5. to develop a formal mechanism for identifying the origin of errors througout the system and how they propagate;
6. to discuss a survey calibration strategy and data processing approaches to ensure survey uniformity.

1. Based both on analysis of the June, 1994 observations of the strips centered on M92 and on simulations, the current KAMPHOT-based approach appears very promising. Even at this relatively early date, the preliminary results come close to and in some cases could eventually exceed the level 1 requirements for point sources (astrometry, photometry, reliability, completeness).

   --- With modifications, the current approach appears capable of exceeding the level 1 photometric requirements; for sources with 8 < K < 12, 2.5% repeatability has been achieved in uncrowded fields using aperture photometry; simulations suggest that 1-2% photometry should be possible in uncrowded fields.

   --- Both analysis of the M92 field and preliminary results from simulations (see below) suggest that the level 1 astrometric specifications can be met and/or exceeded over a large fraction of the sky. Whether this statement applies to the entire sky depends on both the number of proper motion stars (currently drawn from the PPM catalog) in scan overlap regions and the efficacy of post-processing algorithms in carrying out a "global" astrometric and photometric solution; further analysis is required.

   [See viewgraph summary documents from: Beichman (Point Source Processing for 2MASS; overview) Chester and McCallon; Evans and Wheelock; Moshir] --- Kopan reported on a "hybrid" approach (using the DAOFIND algorithm to locate candidate point sources and the KAMPHOT algorithm to define final centroid and photometric parameters) which appears capable of a possible 6-fold increase in point-source extraction efficiency. The group urged that a careful comparison of the DAOPHOT and KAMPHOT source location algorithms be undertaken particularly in simulated crowded fields. This approach appears promising and may provide a significant saving in processing time/hardware.

   --- Schneider, Seitzer, Cutri and Elias (in the context of their calibration star summary) will summarize the arguments in favor of setting as a project GOAL (as opposed to a new level 1 requirement), achieving 2% photometry for stars K < 12. It is anticipated that the arguments will center on dwarf/giant population discrimination, but may also include photometric discrimination of masses and possibly chemical composition among low mass dwarf stars). We recommend discussing this at the upcoming Science Team meeting.

   --- There will also need to be be work (involving the IPAC simulator) and the survey simulator (currently under development by Rudenko at UMass) aimed at refining the survey strategy (tiling strategy; field overlap) to optimize photometric and positional accuracy in all parts of the sky. We recommend that this be discussed by the Science Team; a coordination plan needs to be established.

2. Jarrett and Light described the design and early results from a simulator designed to numerically simulate star fields (i.e. raw data) observed by the 2MASS survey. The star fields are "observed" via a sequence of frames to form a scan sequence. The goal is to use the simulation as ONE OF SEVERAL CHECKS on the photometric and positional accuracy, completeness and reliability of the data, under a variety of conditions. The simulator promises as well to be of great value in evaluating the consequences of various scanning strategies; telescope errors on the accuracy of measured quantities.

   [see summary documents by Jarrett and Light; Light and Jarrett; Fundamental Limits of Photometry] --- Jarrett and Light were urged to carry out simulations which would provide a context for the Science Team's eventually evaluating the suitability of the KPNO 1.3m telescope for the Northern Hemisphere survey should financial considerations drive us to consider this possibility. We have recommended that if possible, they have results available in time for the team meeting. (See 2MASS Protocam Numerical Simulations:Stars)

   We recommend that the Science Team carefully discuss the simulations which will prove most challenging to the current algorithms. Jarrett and Light indicate that a version of the simulator should be available for remote use relatively soon. In order to both organize the approach to simulations, and to control the processing load at IPAC, we recommend that the Science Team develop a coordinated, prioritized list of simulation/reduction tasks.

3. Lonsdale summarized the IPAC view of (a) prototype camera data observations; and (b) higher angular resolution, higher sensitivity data that could provide "ground-truth" for completeness, reliability and extended source identification.

   We recommend careful discussion of the the April/May prototype camera run at the upcoming Science Team meeting. The Lonsdale document should be reviewed prior to that discussion. (See Prototype Camera Observation Plan for April 1995 Run)

4. There was an extensive discussion (based on analysis of flat fields; current prototype camera data; simulation data) of the hardware factors which limit photometric and positional accuracy.

   Cutri and Skrutskie will develop appropriate tests and a flat-fielding strategy during the upcoming prototype camera run (April/May). We recommend their discussing the flat-fielding issues at the upcoming Science Team meeting.

5. The discussion in (4) above fed naturally into a discussion by Fowler aimed at establishing "error trees" for 2MASS. (See 2MASS Error Tree)

   This is viewed as critical to understanding the fundamental sources of error (and thus of potential cost/performance trades) and requires serious study by IPAC and the Science Team.

   We recommend that the Science Team discuss how best to develop a procedure for scrutinizing the assumptions underlying the error tree and for developing a list of simulations and observations aimed at understanding the errors affecting the derived photometric and positional uncertainties.

6. Seitzer presented [see handwritten viewgraph copies] a summary of proposed calibration strategies. The assembled group developed a strong consensus in favor of establishing a grid of equatorial standards (approximately 30 in number; strong overlap with extant Landolt standards) to provide a fundamental system accurate at the ONE PERCENT level. Standard fields will be observed between once/hour and once/90 minutes.

   Schneider, Cutri, Seitzer and Elias will develop a brief (2 page) document summarizing the rationale for this decision. They will also summarize the advantages of setting a goal of 2% photometry for sources K < 12 in uncrowded fields (see summary under 1 above).

   It is essential that the calibration strategy and the overall requirements and desirable goals be discussed thoroughly at the upcoming Science Team meeting.