2MASS Pipeline Processing System (2MAPPS)

Subsystem Tunable Parameter Summary


Contents


  1. Introduction
  2. RD_FRAME
  3. DARKS
  4. PIXCAL/DFLAT
  5. PIXCAL/FREXAS
  6. POSMAN
  7. PICMAN/PIXPHOT
  8. PICMAN/PROPHOT
  9. MAPCOR
  10. BANDMERGE
  11. GALWORKS
  12. MPCAT
  13. CALMON
  14. DBMAN
  15. QUALITY

1. Introduction


The design and implementation of the 2MASS Pipeline Processing System (2MAPPS) is described in the 2MAPPS Functional Design Document . 2MAPPS is a highly modular system comprised of individual subsystems that each handle a well-defined phase of data processing. Overall, processing within 2MAPPS is a linear function in that the output from one subsystem is the input for the next.

The document below describes the parameters within each subsystem that require tuning for optimal operation with the 2MASS data. Within each section, The name, a brief description, and initial estimates for the value of each parameter are listed. At the beginning of each section, the names of the cognizant 2MASS Science Team members and Design Engineers for the subsystems are given. Where relevant, links are also provided to the Subsystem Design Specification Documents (SDS) that provide detailed explanations of each subsystem.

The goals are to 1) converge on a complete and prioritized, yet manageable set of tunable parameters, 2) determine reasonable initial estimates for those parameters, 3) determine what if any further engineering test data should be acquired to perform the analysis necessary to tune the parameters, and 4) ultimately to converge on appropriate values for each parameter. Steps 1) and 2) are largely complete, less the prioritization. Steps 1)-3) will be topics for discussion at the November Science Team Meeting in Boston. The schedule for completion of step 4) coincides with the delivery of 2MAPPS v2.0 in February of 1998.

2. RD_FRAME


Cognizant Team Member - None

Cognizant Engineer - John White

Function:

Converts raw telescope data to rectified FITS images. Flags saturated pixels.

RD_FRAME is a subroutine called by by several other subsystems within 2MAPPS. Consequently, RD_FRAME's tunable parameters are actually parameters passed to it by other routines.


1)saturation values ; Currently single value per band.  This will change to
		      256x256 FITS array per band.

  saturat(j,h,k)   =  43000.,     43000.,    43000.,

2) t1, t2 ;           A pair of dates/times indicating the time range of 
		      validity of the above.

3)in and cross scan sos/eos position offsets  

  xsoff1(j,h,k)    =  0.0,        0.0,        0.0,
  isoff1(j,h,k)    =  0.0,        0.0,        0.0,
  xsoff2(j,h,k)    =  0.0,        0.0,        0.0,
  isoff2(j,h,k)    =  0.0,        0.0,        0.0,
 
4)Axis 1 and Axis 2 Pixel Size (degs) - for FITS headers 
               
  rcdelt1(j,h,k)   = -0.0005589, -0.0005585, -0.0005541,
  rcdelt2(j,h,k)   = +0.0005562, +0.0005580, +0.0005536,



3. DARKS


Cognizant Team Member: Mike Skrutskie

Cognizant Engineer: John Fowler

Retrieve Subsystem Design Specification Document

Function:

Compiles trimmed average dark images from nightly dark scans. Compares with archival dark frames. Compiles nightly responsivity images (flat-fields) from twilight sky scans. Compares with archival gain correction. Generate bad pixel masks.
                                 DARKS
Name      Description                          Dim  Type  Units  Default
--------- ------------------------------------ ---- ----  -----  -------
AndMsk    If AndMsk(I,J) = T, then for band I   3,2   L     -      6*T
          and read-type J, the new and old
          masks will be logically ANDed
DRLim     Maximum number of pixels failing      3,2  I*4    -      6*25
          the MaxRDif test for a responsivity
          image to be accepted [1]
Max1Sat   Maximum number of saturated pixels    3    I*4    -      3*200
          in Read1 to accept Read2-Read1
          frame for use in responsivity 
          computation (one per band)
Max2Sat   Maximum number of saturated pixels    3    I*4    -     3*2000
          in Read2 to accept Read2-Read1
          frame for use in responsivity 
          computation (one per band)
MaxDfMn   Maximum absolute mean of new-old     9,2   R*4   -/DN  3*0.003
          pixel values for J/H/K responsivity,                    6*10.0
          J/H/K Read1 darks, and J/H/K Read2-                      3*0.1
          Read1 darks, respectively [2]                           6*99.9
MaxDfSD   Maximum sigma value of new-old       9,2   R*4   -/DN  3*0.005
          pixel values for J/H/K responsivity,                    6*25.0
          J/H/K Read1 darks, and J/H/K Read2-                      3*0.2
          Read1 darks, respectively [2]                           6*99.9
MaxMed    Maximum permitted dark-subtracted     3    R*4     DN  3*27000
          frame median to keep in stack for
          responsivity calculation; one/band
MaxMEMn   Maximum absolute mean of evening     9,2   R*4   -/DN  3*0.003
          minus morning pixel values for                          6*10.0
          J/H/K responsivity, J/H/K Read1                          3*0.1
          dark, and J/H/K Read2-Read1 dark                        6*99.9
          images, respectively [2]
MaxMESD   Maximum sigma value of evening       9,2   R*4   -/DN  3*0.005
          minus morning pixel values for                          6*25.0
          J/H/K responsivity, J/H/K Read1                          3*0.2
          dark, and J/H/K Read2-Read1 dark                        6*99.9
          images, respectively [2]
MaxPass   Maximum number of linear least-            I*4    -       10 
          square passes per pixel allowed in
          the responsivity calculation
MaxRDif   Maximum absolute morning-evening     3,2   R*4    -      6*0.1
          and new-canonical responsivity
          difference for a pixel not to be
          included in sum for DRLim test
          (see above) [1]
MaxSD     Maximum sigma value for a pixel's     3    R*4    DN    3*25.0
          dark average for each band; pixel is
          masked if exceeded
MinFrac   Minimum fraction of raw pixel values       R*4    -       0.3
          to retain in final least-square pass
MinFrms   Minimum number of usable frames to    3    I*4    -       3*30
          proceed with responsivity computation
          (one per band)
RngMin    Minimum range in frame median to      3    R*4    -    3*20000
          proceed with responsivity computa-
          tion (one per band)
RspMax    Maximum responsivity value in J/H/K;  3    R*4    -     3*2.0
          if exceeded, pixel is turned off in
          the corresponding masks
RspMin    Minimum responsivity value in J/H/K;  3    R*4    -     3*0.5
          if less, pixel is turned off in the
          corresponding masks
SigRej    Sigma rejection level for linear           R*4    DN      2.0
          least-square passes after the first
TrmFrac   Fraction of total number of dark           R*4    -      0.25
          values for a pixel to be trimmed from
          each end of sorted array before
          averaging

Notes
 
[1] The first index in the (3,2) dimension is band (J, H, K), and the second
index has the value 1 for morning-evening differences and 2 for new-canonical
differences.

[2] The second index in the (9,2) dimension is a severity index; the first value
causes warning messages to be displayed if exceeded, the second causes DARKS to
return an error code of 4 after finishing its computations, which halts the
pipeline.


4. PIXCAL/DFLAT


Cognizant Team Members: Richard Elston & Roc Cutri

Cognizant Engineer: John Fowler

Retrieve Subsystem Design Specification Document

Function:

Compiles additive sky offset correction frames. Applies dark frame, flat-field and sky offset corrections to each data frame. Refines image masks.

Name      Description                          Dim  Type  Units  Default
--------- ------------------------------------ ---  ----  -----  -------
DLMax     Maximum acceptable deviation from     2    R*4    -      10.0,
          linearity of a pixel in an Sj image                      50.0
          with respect to time variation (in
          units of the pixel's mean standard
          deviation about its trimmed average
          offset correction; if exceeded and
          SetBlank = T, the pixel will be
          turned off in the Sj image; the index
          corresponds to read type
DSigMax   Maximum non-anomalous sigma of        2    R*4    9.9e25,
          linear fit of Sj pixel over time                       9.9e25
          (in noise units; see ZSigMax for 
          ""); if above and GenMask = T,
          pixel will be turned off in the mask;
          index corresponds to read type
MaxBlnk   Maximum no. of times a given pixel    1    I*2    -       3 
          may be blank in an Sj image before
          it is turned off in the mask if
          GenMask = T
NFrPerFl  No. of frames between flat            1    I*4    -       44
          computations                        
NFrWdth   No. of frames to use to compute       1    I*2    -       42
          an Sj image (stack size) 
NRMDR     Number of RMDR trims to do before     1    I*2    -       20
          the symmetric trimming
NTrim     No. of pixels to trim symmetrically   1    I*2    -        0
          from each end of the sorted list of 
          stacked pixels before averaging them        
SetBlank  If T, Sj pixels will be set           1     L     -        T
          blank if anomalous                  
SMax      Maximum acceptable value of a pixel   2    R*4   DN       500,
          in an Sj image; if exceeded and                           500
          SetBlank = T, pixel will be blanked 
          in that Sj image; index corresponds
          to read type
SMin      Minimum acceptable value of a pixel   2    R*4   DN      -500,
          in an Sj image; if not exceeded and                     -500
          SetBlank = T, pixel will be blanked
          in that Sj image; index corresponds
          to read type
TSigMax   Maximum non-anomalous sigma of        2    R*4   DN    9.9e25,
          trimmed average in Sj computation;                     9.9e25
          If above and SetBlank = T, pixel 
          will be blanked in that Sj image;
          index corresponds to read type
ZSigMax   Maximum non-anomalous z-sigma of      2    R*4       4,4
          trimmed average in flat computation;
          "z-sigma" in units of "" means 
          a pixel's standard deviation relative 
          to the standard deviation averaged
          over all non-masked pixels; pixels
          over the ZSigMax threshold will be 
          turned off in the mask if GenMask=T
          and TAStat=T; index corresponds to
          read type


5. PIXCAL/FREXAS


Cognizant Team Member: John Carpenter

Cognizant Engineer: Gene Kopan

Retrieve Subsystem Design Specification Document

Function:

Read_1 and Read_1-Read_2 frame source extractions. Read_1 photometry. Frame statistics measurement. Frame offset estimation.

   variable  dim   type     default      description
-----------------------------------------------------------------------------
    mag0     2,3    R4       17.4  20.8  mag0(i,j) = template magnitude zero point
                             17.2  20.6  offset for read i of band j
                             16.5  19.9
    amag0    2,3    R4       17.4  20.8  amag0(i,j) = aperture magnitude zero point
                             17.2  20.6  offset for read i of band j
                             16.5  19.9
    doaph      2    L          T     T   if doaph(i) = .true., call apphot for aperture
                                         photometry for read i
    par       14    R4        2.0   4.0  par(i)(i=1,12)= radii of aperture i.
                             12.0  15.0  par(13,14)= inner,outer radii for background
    hibad      1    R4       32000.0     maximum valid ADU
    phpadu     1    R4          10.0     gain (e-/ADU)
    readns     1    R4          48.0     read noise (e-)
    minrad     1    R4           3.0     minimun radius for apphot pixel fix-up
    pfrac      1    R4           0.5     initial pfrac for pass 1
    rpk        1    R4           0.3     maximum radius from peak for pfrac histo
                                         in pixels
    tpk        1    R4          20.0     minimum snr for pfrac histogram
    prmax      1    R4           0.85    maximum peak ratio for pfrac histogram
    thres    2,3    R4        15.0  10.0 thres(i,j) = extraction snr threshold for
                              15.0  10.0 read i of band j
                              15.0  10.0 
    siglo      1    R4        -1.0       noise is computed from difference in 
    sighi      1    R4        +1.0       gaussian percentiles of sighi and siglo
    sigmed     1    R4         0.0       background is computed from gaussian
                                         percentile of sigmed
    nb2do      1    I4          3        if: nb2do = 1, process read 1 only
                                         nb2do = 2, process read 2 only
                                         nb2do = 3. process both reads
                                         nb2do = 0. process neither read
    psiga,   2,3    R4       6*0.6       modeled position error = 
    psigb    2,3             6*2.0       sqrt(psiga^2 + (psigb/snr)^2)
    thrdmp   2,3    R4       4*50,2*30   dump 7x7 pixel patches for sources snr>thrdmp


   *Also, psfs need to be generated over a range of seeing. This is currently
    my highest priority.


6. POSMAN


Cognizant Team Member: Dave Monet

Cognizant Engineer: Howard McCallon

Retrieve Subsystem Design Specification Document

Function:

Frame and source position reconstruction. Final determination of frame/band offsets. Reconstruction into J2000 reference frame. Association with Optical source catalogs (Tycho and USNOA).

|parm  |current est.|Description                                               |
--------------------------------------------------------------------------------
                     CONTROL PARAMETERS
 mfrac   .5          Match fraction required for GOOD pattern match
 nrpmax  ?           Max number of reference peg-pair trys allowed
 nepmax  ?           Max number of extraction peg-pair trys allowed
 mxtry2  20          Max number of trys with same first reference peg allowed
 mnmfor  10.         Min ratio of matches found over expected random matches
 drmch   3. asec     Match acceptance radius 
 dxcan   90. asec    Max x-scan search radius for peg-pair matching
 dycan   90. asec    Max in-scan search radius for peg-pair matching
 ysepfr  .33         Min in-scan separation of pegs in fractional scan length
 maxjmv  1.5 mag     Max magj-magv allowed for match
 mxeovr  500.        Max ratio of extractions over references used for
                     pattern match. Only the n brightest extractions are
                     used where n is determined from mxeovr.
 dsymx   .01         Max in-scan scale change from apriori allowed.  Used to
                     limit search space.  
 necmx   100         Max number of extraction candidates used per reference peg.
                     Will consider the necmx brightest within the dxcan/dycan
                     window.

 mintyc  4           Minimum number of tychos required in scan for fit
 tcfrac  .33         Tycho coverage fraction of scan length required for fit

 sigap   5. asec     Sigma on the three apriori pseudo points which tie
                     each frame to the next.  Needed to bridge frames
                     without extractions.

 wxbox    3. asec    X-scan window used for merging extractions
 wybox    3. asec    In-scan window used for merging extractions

 mxdpfrj  200        Max detections per frame in J-band used for reconstruction
 mxdpfrh  200        Max detections per frame in H-band used for reconstruction
 mxdpfrk  200        Max detections per frame in K-band used for reconstruction

 chisqmx   32.       Max ChiSq for final point source associations 

                     INITIALIZATION PARAMETERS
 cxr10   .0 asec     READ1-READ2 x-scan difference
 cyr10n  .3 asec     READ1-READ2 in-scan difference for northward scans
 cyr10s -.2 asec     READ1-READ2 in-scan difference for southward scans

 cdxh2j   2.7 asec   X-scan J-H difference at nf=0
 mdxh2j    ?   as/nf Slope of x-scan J-H difference w.r.t. frame number
 cdyh2j   3.4 asec   In-scan J-H difference at nf=0
 mdyh2j    ?   as/nf Slope of in-scan J-H difference w.r.t. frame number
 cdxk2j  -0.7 asec   X-scan J-K difference at nf=0
 mdxk2j    ?   as/nf Slope of x-scan J-K difference w.r.t. frame number
 cdyk2j   1.4 asec   In-scan J-K difference at nf=0
 mdyk2j    ?   as/nf Slope of in-scan K-K difference w.r.t. frame number

 wpxj    1.9971 asec J-band x-scan pixel width 
 wpxj    1.9971 asec J-band in-scan pixel width
 dthej   .388 deg    J-band rotation angle w.r.t. scan direction
 wpxh    2.0061 asec H-band x-scan pixel width 
 wpxh    2.0061 asec H-band in-scan pixel width
 dtheh   .451 deg    H-band rotation angle w.r.t. scan direction
 wpxk    1.9837 asec K-band x-scan pixel width 
 wpxk    1.9837 asec K-band in-scan pixel width
 dthek   .344 deg    K-band rotation angle w.r.t. scan direction

 dra0n    -3.68 asec Bias in observatory est. of RA at northward scan start
 dra0s    -5.58 asec Bias in observatory est. of RA at southward scan start
 ddec0n  156.34 asec Bias in observatory est. of Dec at northward scan start
 ddec0s -177.14 asec Bias in observatory est. of Dec at southward scan start

 xstepjn   0.49 asec J-band x-scan frame-to-frame step size for northward scans
 xstepjs  -0.48 asec J-band x-scan frame-to-frame step size for southward scans
 istepjn  82.64 asec J-band in-scan frame-to-frame step size for northward scans
 istepjs -82.64 asec J-band in-scan frame-to-frame step size for southward scans

MODEL PARAMETERS
-----------------------------------

 Distortion model - TBD
 Uncertainty model - TBD


POSPTS PARAMETERS
-----------------
 
Name         Description                        Type  Dim  Units    Default
------   -------------------------------------  ----  --- -------  --------
DXAB     Coarse X window (ASOC/BMG matching)     R*4   1   arcsec      3.0
DXMax    Max abs mean N-Scan X offset to         R*4   1   arcsec      0.3
         accept first ASOC/BMG match set [1]
DXPB     Coarse X window (PFMP/BMG matching)     R*4   1   arcsec      3.0
DYAB     Coarse Y window (ASOC/BMG matching)     R*4   1   arcsec      3.0
DYMax    Max abs mean N-Scan Y offset to         R*4   1   arcsec      0.3
         accept first ASOC/BMG match set [1]
DYPB     Coarse Y window (PFMP/BMG matching)     R*4   1   arcsec      3.0
ThrAB    Threshold for ASOC/BMG match, pass 1    R*4   1  [2] -       25.0
ThrAB2   Threshold for ASOC/BMG match, pass 2    R*4   1  [2] -       25.0
ThrPB    Threshold for PFMP/BMG match            R*4   1  [2] -       25.0
XSigMx   Max N-Scan X sigma of matched objects   R*4   1   arcsec      2.0
         to accept first ASOC/BMG match set [1]
YSigMx   Max N-Scan Y sigma of matched objects   R*4   1   arcsec      2.0
         to accept first ASOC/BMG match set [1]
				      
Notes
-----
1.) If any of the DXMax, DYMax, XSigMx, or YSigMx thresholds are exceeded
by the best-match pairs of associated sources after the first pass
through the catalog association loop, the associations are done over
from the beginning with the mean offsets in N-Scan X and Y from the
first pass subtracted out, and a separate threshold parameter (ThrAB2)
is used. This applies only to ASOC/BMG matching.

2.) Thresholds ThrAB, ThrAB2, and ThrPB are 2-D chi-square upper limits.


7. PIXPHOT/PICMAN


Cognizant Team Member: John Carpenter

Cognizant Engineer: Gene Kopan

Retrieve Subsystem Design Specification Document

Function:

Assembles and registers Atlas Images from calibrated frames. Deep source detection.

   variable    type     default   description
-----------------------------------------------------------------------------
    coff        L          T      if = T, match frame background levels
    nlino       I        1024     # lines in full size coadded image
    ncolo       I         512     # columns in coadded image
    dlpim       F         970     # lines between coadded images
    ncmin       I           2     if < ncmin frames overlap a coadd pixel, set it blank
    zroxof      L          T      if = T, center coadds on x=0 Uscan coordinate
                                  else center on data
    solrad      F          2.5    blanking radius for solos (camera pixels)
    marzap      L          F      if = T, mark zapped solos in coadds (debugging)
    siglo       F         -1.0    Noise estimator uses the interval between the
    sighi       F          0.0    Gaussian percentiles of siglo,sighi /(sighi-siglo)
    dethr       F          3.0    Detection threshold (sigma)
    fwhm        F          4.0*   FWHM of Gaussian filter used for detection
                                  * should be a (TBD) function of seeing
    readns      F          5.0    Camera read noise (dn) - used for noise estimate
    phpadu      F         10.0    Camera gain (photons/dn) - used for noise estimate
    nolap       I          6      frame overlap depth - used for noise estimation
    lobad       F         10.0    Maximum valid sigma below background in input frames
    hibad       F        40000    Maximum valid dn in input frames
    shrplo      F          0.10   Minimum value of sharp parameter for valid detection
    shrphi      F          1.00   Maximum value of sharp parameter for valid detection
    rndlo       F         -2.00   Minimum value of round parameter for valid detection
    rndhi       F          2.00   Maximum value of round parameter for valid detection
    seep        F          1.1    Value of seeing parameter (overridden by seeman)
    wpc         F          1.0    width of coadd pixeli (arcsec)
    kernel      A          wkern  Interpolation kernel type:
                                    'wkern' - Weinberg kernel
                                    'bilin' - Bilinear kernel
                                    'aolap' - Area overlap kernel
                                    'kerfil' - Read kernel from kerfil
                                    ' '     - kernel specified in namelist
    hsmo        F          0.35   smoothing length for Weinberg interpolation kernel
    xpdel(ib)   F          0.0    PSF cross-scan offset correction for band ib (pixels)
    ypdel(ib)   F          0.0    PSF    in-scan offset correction for band ib (pixels)

8. PROPHOT


Cognizant Team Member: Martin Weinberg

Cognizant Engineer: Bob Light

Retrieve Subsystem Design Specification Document

Function:

Point source aperture and profile fit photometry. Optimal brightness, position and uncertainty estimation.

A. Apphot Parmeters

                   Current
        Name        Value          Description

   1.  SKYIN       7.0 pix      Inner radius for sky annulus, within which
                                pixels are used for a clipped sky mean.

   2.  SKYOUT     10.0 pix      Outer radius for sky annulus, within which
                                pixels are used for a clipped sky mean.

   3.  MINRAD      3.0 pix      Radius outside of which bad pixels are
                                replaced with a reasonable PSF+sky value.



B.  PSF fitting Parameters

                   Current
        Name        Value          Description

   1.  RADIUS      2.0 pix      Radius within which pixels are used in
                                the profile fitting process.

   2.  MASKYR      3.0 pix      Radius within which all pixels near all
                                detected sources are masked from the sky
                                (background) determination routine.

   3.  THRESH        1%         Fraction of PSF flux level that a detected
                                neighbor has at the center of a star before 
                                two detections are profile-fit simultaneously.

   4.  ITER         100         Limit on the number of iterations for the 
                                chi-squared minimization to achieve a solution.



C.  Deblending Parameters

                   Current
        Name        Value          Description

   1.  CHIP          5.0        Reduced chi-squared limit, above which a
                                single source may be "split" into two possible
                                sources.

   2.  XMAG         12.5        Magnitude limit for splitting sources; do not
                                split fainter sources.

   3.  NSOURCE       TBD        Density limit for splitting sources, the 
                   (b<20)       number of stars in a 6-frame overlap segment; 
                                do not split sources in very dense regions.

   4.  WEIGHT2       0.2        Weighting function parameter applied to ratio 
                                of peak pixel values in primary and secondary 
                                sources to determine the preliminary 
                                separation of double.

   5.  CHIP2     10000.0        Reduced chi-squared limit, above which a
                                double source may be "split" into three 
                                sources.



C.  PSF Manufacture Parameters

                   Current
        Name        Value          Description

   1.  NSTARS         50        Number of bright (non-saturated) stars used
                                to make the PSF for a scan.

   2.  FWHM          1.2 pix    FWHM of a Gaussian function used as a starting
                                point to extract PSFs, when selecting out
                                aberrant data such as doubles and galaxies.

   3.  SNMIN        50.0        Minimum acceptable S/N for a stellar image
                                as a PSF candidate.
 
   4.  DIFFMX        0.3 mag    Maximum acceptable magnitude difference;
                                |PSF fit - aperture| .
 
   5.  CHIMAX       10.0        Maximum acceptable reduced chi squared for
                                a source fit.


Notes:

  1)  The unit of a "pix" comes from a frame pixel, being 2 arcsec on a side.

  2)  Several of the "Profile Fit" parameters are used by the PSF
      Manufacture routines, for similarity in the handling of PSFs,
      e.g. RADIUS and MASKYR.


9. MAPCOR


Cognizant Team Member: Jay Elias

Cognizant Engineer: Tracey Evans

Retrieve Subsystem Design Specification Document

Function:

Photometric aperture corrections. Read_1 and profile fit photometry normalization. Artifact and confusion identification.

Look-up table values:
---------------------

Read_1 magnitude normalization value & sigma

For each PSF:
    aperture correction & sigma
    PSF-fit magnitude normalization & sigma 


Namelist values:
----------------

 det_lim     = j:17.0, h:16.5, k:16.0,
   pt src detection limit mags, beyond which no artifact searches are made.

 std_ap      = 2.0,
   The size of the "standard" aperture radius (in camera pixels) for the
     standard aperture mag, to use for the aperture correction determination.

 lg_ap       = 5.0,
   The size of the "large" aperture radius (in camera pixels) for the
     large aperture mag to carry along; useful for possible extended srcs.

 brstar_thr  = j:10.5, h:10.0, k:9.5,
   A source brighter than brstar_thr causes MAPCOR to
     search for resulting bright-star artifacts.

 starchi_thr = 6.0,
   A source is considered a star (for MAPCOR's purposes) 
     if its PSF-fit chi^2 parameter from the Read2-Read1 source data file 
     is less than starchi_thr.

 psfsrcs_min = 10,
   A "PSF group" must have at least this many sources for calculating
     aperture corrections and PSF-fit magnitude normalizations

 n_rej       = 2.0,
   The number times the rms to use in the root-mean-square rejection
     when iteratively calculating medians.

 loss_max    = 0.30,
   The max fractional loss of srcs from rejection to allow when iteratively
     calculating medians.

 iter_max    = 20,
   The maximum number of iterations allowed in calculating medians.

 del_med     = 0.0005,
   Iterative median calculation stops when 
     the change in the median drops below delmed. 

 n_check     = 3.0,
   When checking calculated values compared to lookup values, a warning is
     generated if they don't agree to within +/- n_check*sigma.

 satr1_c     = j: 0.973, h: 0.886, k: 0.663,
 satr1_d     = j:-5.702, h:-4.385, k:-1.444,
   A corrected magnitude for a saturated Read1 source
     will be calculated by m_c = c * m_p1 + d,
     where m_p1 is the magnitude of its first persistence source 
     in the Read1 source list

 satr1_rms   = j:0.443, h:0.601, k:1.11,
   rms of real R1 mag vs. 1st persistence src mag relation, to be rss'ed 
     with the persistence src mag sigma to determine the new sigma value.

 pers_nmin   = 4,
   min # of persistence srcs to search for

 pers_nmax   = 10,
   max # of persistence sources to search for

 chain_dens  = 5000.0,
   An unbroken chain of persistence sources may be required if
     the source density (per sq. deg) of the scan is above chaindens,
     to lessen the chance of real srcs being called persistence.

         (npers)   1    2    3    4    5    6    7     8     9    10
 pers_dm     = j: 5.8, 7.6, 8.4, 8.8, 9.1, 9.4, 9.8, 11.0, 12.0, 13.0
               h: 5.8, 7.6, 8.4, 8.8, 9.1, 9.4, 9.8, 11.0, 12.0, 13.0
               k: 5.8, 7.6, 8.4, 8.8, 9.1, 9.4, 9.8, 11.0, 12.0, 13.0
 pers_sigm   = j: 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,  1.0,  1.0,  1.0
               h: 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,  1.0,  1.0,  1.0
               k: 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,  1.0,  1.0,  1.0
   A progression of predicted persistence sources 
     should be fainter in magnitude than the "parent" source by 
     the empirically determined series pers_dm(band,npers)
     with uncertainties of pers_sigm(band,npers).

 pers_r      = 3.0,
   pers_r is the persistence search radius (in pix)

 pers_sigr   = 1.0,
   Predicted persistence locations have an empirically determined position
     uncertainty pers_sigr from camera jitter and frame-to-frame offset
     determination, which is used in the persistence chi^2 probability
     calculation. 
     chi^2 = [(x_src - x_pred)/sig_x]^2 + [(y_src - y_pred)/sig_y]^2 + 
             [(m_src - m_pred)/sig_m]^2 ;
     sig_x^2 = sig_x_parent^2 + sig_x_src^2 + pers_sigr^2 ;
     sig_y^2 = sig_y_parent^2 + sig_y_src^2 + pers_sigr^2 ;
     sig_m^2 = sig_m_parent^2 + sig_m_src^2 + pers_sigm^2 ;
     which is then converted to a probability.

 pers_pthr   = 0.50,
   If a persistence probability is above pers_pthr, the src is considered a
     persistence src within MAPCOR when determining if it is a "good" src,
     and keeping counts of artifacts.

       (glint #)   1    2
 glint_delm  = j: 6.4,
               h: 6.4,
               k: 6.4, 6.4,

 glint_sigm  = j: 1.0,
               h: 1.0,
               k: 1.0, 1.0,

   Filter glints have magnitudes that are glint_delm mag fainter than the
     "parent" bright sources with an uncertainty of glint_sigm.

       (glint #)   1    2
 glint_e     = j: 1.0,
               h: 1.0,
               k: 1.0, 1.0,
 glint_f     = j: 0.5,
               h: 0.5,
               k: 0.5, -2.0,
 glint_g     = j: 1.0,
               h: 1.0,
               k: 1.0, 1.0,
 glint_h     = j:-7.0,
               h:-7.0,
               k:-7.0, 8.0,
 glint_r     = j: 2.0,
               h: 2.0,
               k: 2.0, 2.0,
   Filter glints are located at the positions x = glint_e*x + glint_f, 
     y = glint_g*y + glint_h
     within an uncertainty radius of glint_r,
     where (x,y) is the position of the "parent" source.

 diffrlsc_dm = j:3.0, h:3.0, k:3.0,
   If a source is located on a bright-star diffraction spike but is brighter
     than m+diffrlsc_dm, where m is the bright star's magnitude, it will
     be accepted as a real source but flagged as contaminated by the spike.

 diff_width  = j:2.5, h:2.5, k:2.5,
   A diffraction spike has a predictable width diff_width. 

 diff_l0     = j:30., h:30., k:30.,
 diff_m0     = j:7.0, h:7.0, k:7.0,
 diff_a      = j:0.1, h:0.1, k:0.1,
   A diffraction spike has a predictable length
     (the distance it extends out from the center of the bright source)
     that is empirically modeled as l = diff_l0*{10^[diff_a(diff_m0-m)]},
     where m is the "parent" source's magnitude.

 diff_angle  = j:0., h:0., k:0.,
   diffraction spike angle to horizontal in deg.

 diff_lmin   = j:  1., h:  1., k:  1.,
 diff_lmax   = j:110., h:110., k:110.,
   Calculated diffraction spike lengths are limited by diff_lmin and diff_lmax.

 conf_r0     = j:2.0,   h:2.0,   k:2.0,
               j:2.0,   h:2.0,   k:2.0,
 conf_m0     = j:9.0,   h:9.0,   k:9.0,
               j:9.0,   h:9.0,   k:9.0,
 conf_b      = j:0.125, h:0.125, k:0.125,
               j:0.125, h:0.125, k:0.125,
   A "parent" source's confusion radius 
     is empirically modeled as r = conf_r0*{10^[conf_b(conf_m0-m)]},
     where m is the "parent" source's magnitude.

 conf_rmin   = j: 2.0, h: 2.0, k: 2.0,
               j: 2.0, h: 2.0, k: 2.0,
 conf_rmax   = j:30.0, h:30.0, k:30.0,
               j:30.0, h:30.0, k:30.0,
   Calculated confusion radii are limited by conf_rmin and conf_rmax.

 see_fact    = j:0., h:0., k:0.,
   To account for seeing effects, the confusion radius is root-sum-squared
     with see_fact, a function (TBD) of band and the seeing parameter.

 brap_thr    = j: 8.0, h: 8.0, k: 8.0,
 faap_thr    = j:13.5, h:13.5, k:13.5,
   A point source with a magnitude between brap_thr and faap_thr may be used 
     in calculating the aperture curve-of-growth and aperture correction.

 del_conv    = 0.015,
   The aperture curve-of-growth converges when rho_i,
     the median of m_i - m_{i+1}, is <=del_conv for a certain aperture i.
     The value of 0.015 mag corresponds to the amount of the photometric error 
     budgeted for aperture corrections.

 del_div     = -0.0005,
   The aperture curve-of-growth diverges when rho_{i-1} - rho_i <= del_div. 

 optap_max   = 5.0,
   The maximum "optimal" aperture (in camera pixels) allowed
     in the aperture correction determination.

 brpsf_thr   = j: 8.0, h: 8.0, k: 8.0,
 fapsf_thr   = j:12.0, h:12.0, k:12.0,
   A source with a PSF-fit magnitude between brpsf_thr and fapsf_thr may be
     used  for determining the normalization of the PSF-fit magnitudes 
     to the aperture-corrected aperture magnitudes.

 r1_rad      = 3.0,
 r1_dm       = 3.0,
   A Read1 source and a Read2-Read1 source are an acceptable match in
     the Read1 and Read2-Read1 merge if their positions are separated by
     less than r1_rad and their magnitudes differ by less than r1_dm.
     (r1_dm may become a function of Read1 magnitude, to account for
     Read2-Read1 magnitude errors from saturation; now it is simply a
     large value intended only to loosely constrain the Read2-Read1 match's
     magnitude.)

 r1mag_thr   = j:9.0, h:9.0, k:9.0,
   (Note: this parameter will probably be deleted
     and the saturated pixel flag always used instead.)
     If the Read1 magnitude is brighter than r1mag_thr,
     the Read1 magnitude data should be used instead of 
     the matching Read2-Read1 source's magnitude data.

 r1pos_thr   = j:7.0, h:7.0, k:7.0,
   (Note: this parameter might be deleted
     and the saturated pixel flag always used instead.)
     If the Read1 magnitude is also brighter than r1pos_thr,
     the Read1 magnitude and position data should both be used instead of 
     the matching Read2-Read1 source's data.

 brr1_thr    = j: 8.5, h: 8.5, k: 8.5,
 far1_thr    = j:10.0, h:10.0, k:10.0,
   A source with a Read1 magnitude 
     between brr1_thr and far1_thr may be used 
     for determining the normalization of the Read1 magnitudes
     to the Read2-Read1 magnitudes.


10. BANDMERGE


Cognizant Team Member: Jim Liebert

Cognizant Engineer: John Fowler

Retrieve Subsystem Design Specification Document

Function:

Merge J, H and Ks detections. Update positions.

Name      Description                          Dim  Type  Units  Default
--------- ------------------------------------ ---  ----  -----  -------
ChSqMax   Array of values for maximum allowed   2,   R*4   [1]      [1]
          chi-square used to specify density-   10
          dependent match thresholds
CnfVec    Confusion coefficient vector for      ?    R*4    -       [2]
          dot product with status-flags
          vector; added to chi-square test
          parameter when confusion has been
          diagnosed; see note [2]
ComFac    Scale factor for threshold applied    1    R*4    -       1.5
          to detections linked to a common
          detection but not to each other
Density   Array of source density values used   10   R*4   [1]      [1]
          to specify density-dependent match
          thresholds
FChiSq    Scale factor for position term in      1   R*4    -       1.0
          the pseudo-chi-square formula used
          for confusion match testing
FPh       Scale factor for photometric term in   1   R*4    -       0.0
          the pseudo-chi-square formula used
          for confusion match testing
SigUx     1-sigma uncertainty in X due to error  3   R*4  arcsec 3*0.05
          in mapping band-peculiar coordinates
SigUy     1-sigma uncertainty in Y due to error  3   R*4  arcsec 3*0.05
          in mapping band-peculiar coordinates
          to U-Scan coordinates, one per band

          Notes

[1]  The density-dependent threshold arrays are defined as follows.
ChSqMax(1,N) specifies the maximum allowed chi-square value for the source
density Density(N), where the latter is specified in total sources per square
degree in the two bands being tested, and the former applies to the unconfused
test. ChSqMax(2,N) is the corresponding value for the confused test, i.e., the
pseudo-chi-square. The arrays must be specified with Density(N) increasing as
N increases. The default values are as follows.

                 N  Density   ChSqMax(1,N)   ChSqMax(2,N)
               1          0        36.0         9.9e25
               2      95000        36.0         9.9e25
               3     100000        34.8         9.9e25
               4     200000        17.4         9.9e25
               5     300000        11.6         9.9e25
               6     350000        10.0         9.9e25
               7    9999999        10.0         9.9e25
               8    9999999        10.0         9.9e25
               9    9999999        10.0         9.9e25
              10    9999999        10.0         9.9e25

[2]  When confusion is diagnosed, the position-test chi-square parameter is
replaced with a pseudo-chi-square parameter that is a sum of the position-test
chi-square rescaled by   FChiSq, a photometric chi-square scaled by FPh, and a
dot product of the CnfVec and Flags vectors, where CnfVec is user-specified,
and Flags is formed from various status parameters in the point-source record;
see Appendix D for a detailed description of the Flags vector, how it is used
with CnfVec, and the default values for the components of CnfVec.


11. GALWORKS


Cognizant Team Members: Steve Schneider/John Huchra

Cognizant Engineer: Tom Jarrett

Retrieve Subsystem Design Specification Document

Function:

Extended source discrimination. Extended source photometry, position and uncertainty determination. Seeing estimation. Band-filling.

General Parameters:

   Mag limit for bright star blanking

   Mask sizes for bright star blanking, including
      diffraction spikes, persistence, horizontal stripes
      & ghosts
    {note: all are a function of the brightness of
    the object and the confusion noise or equivalently the stellar
    number density}   

   Variable mag thresholds for gal candidacy (function
    of stellar density -- critical to processing efficiency
    as the density becomes large)

   Score thresholds, both preliminary and final thresholds         

   Algorithm 2: mag thresholds and BSNR thresholds



A more complete list of tunable parameters are given below:
--------------------------------------------------------------

deblend_density = 3.1 (default)
 

mag limits for gal candidacy (function of stellar num density)
	Klim = 14.25
	Hlim = 14.75
	Jlim = 15.5  


mag limits for star subtraction and blanking

mags limits for ellipse fitting; otherwise use "super" coadd vals
	Klim_ellipt = 13.5
	Hlim_ellipt = 13.9
	Jlim_ellipt = 14.2

minima/maxima limits for profile fitting and scale length measurement
	radmax = 15.0
	radmin=4.0
	radmax_elfit= 12.0
	radmax_wedge=8.0
    
    
mag limit for bright star masking 
	brightlim = 9.0 (function of stellar num density)


other bright star masking parameters:
	mask radius (function of star brightness, and stellar num density)
	npersistence masks (function of star brightness)
 	diffraction spike length/width (function of star brightness)
	horizontal stripe masks (function of star brightness)

 mag limit for bright star processing (algor 3)


 fraction area limit of bright galaxy processing
 maximum size of known gal to process (diam in arcmin):  BGALDIAM
 minimum ratio between gal size and gal appearance on coadd: BGALRATIO
    (i.e.,  at least 75% of gal must be on coadd)

	BGALRATIO = 0.75


 detection threshold (nsigma)

 initial score thresholds
 P_scorelim = score limit for preliminary mxdn scores
 C_scorelim = score limit for mxdn, sh, & wsh tests
 T_scorelim = score limit for trip killers: vmean, vint, sctrip
 R_scorelim = score limit for r23
    
	P_scorelim = 0.00
	C_scorelim = 2.00
	T_scorelim = 2.00
	R_scorelim = 0.00


 mag limits for star subtraction and blanking
 radius of star subtraction / blanking (function of star brightness)
 radius of isophotal substitution area

 dx,dy from edge to work
	Edge_lim = 10

 stellar density limit at which to attemp galaxy deblends
	deblend_density = 0.1

 stellar density limit at which stars are blanked before ellipse fitting
	ellfit_density = 3.5

 mag limits at which to jettison K as fiducial

 stellar density limit at which confusion noise is used instead of poisson noise

 SNR limit at which to extract postage stamp

 n-sigma contour for ellipse fitting (default = 3)

 curve of growth convergence criteria
   a)  flux growth
   b)  stellar contamination

 petrosian ratio to compute petrosian radius and flux

 LSB detection thresholds
 LSB mag limits
 LSB SNR limits for star blanking
 LSB blocked-snr limits for gal candidacy
 LSB SNR limits for stamp extraction

 blanking radius after object processed (function of curve of growth)


12. MPCAT


Cognizant Team Member: Steve Price

Cognizant Engineer: John Fowler

Retrieve Subsystem Design Specification Document

Function:

Identify candidate minor planets in survey scans. Search for minor planet detections within source lists.

Name     Description                                           Default
-----------------------------------------------------------------------------

DecBox  Half-width DEC coarse matching window (half-width of   8.333e-3 deg.
	coarse RA window is calculated from this).              

ChiSqMx Chi-squared threshold value for minor planet - 2MASS   16.0
	planet - 2MASS source match.

13. CALMON


Cognizant Team Members: Pat Seitzer and Jay Elias

Cognizant Engineer: Sherry Wheelock

Retrieve 2MASS Photometric Calibration Strategy Document

Function:

Evaluate photometric transformations for a night. Estimate photometric quality. Apply photometric transformations to source data.

Name       Default        Description
---------------------------------------------------------------------------

&acj =     0.0887 mag/am  Slope of extinction correction j channel
&ach =     0.0873 mag/am     "                           h channel
&ack =     0.0763 mag/am     "                           k channel
 maxnrms = 0.04 mag       Rms limit for nightly average
 maxlrms = 0.04 mag       Rms limit for linear fit
*maxlslp = 0.0083333 mag/hr  .2/24 hr slope limit for linear fit
*jmin =    11. mag        Brightest j mag for smerge output - precal
*jmax =    14. mag        Faintest j channel for smerge output - precal
*hmin =    10.5 mag       Brightest h mag for smerge output - precal
*hmax =    13.5 mag       Faintest h mag for smerge output - precal
*kmin =    10 mag         Brightest k mag for smerge output - precal
*kmax =    13 mag         Faintest k mag for smerge output - precal
*jbs =      3 mag         Magnitude bin size for smerge - precal - j channel
*hbs =      3 mag         Magnitude bin size for smerge - precal - h channel
*kbs =      3 mag         Magnitude bin size for smerge - precal - k channel
 minfld =   3 mag         Minimum number of calibration observations for night
 nsf =     38             Number of calibration tiles

 sstrf1='90004v1 90009v1 90013v1 90021v1 90067v1 90091v1 90121v1 90161v1 90191v1 90217v1 90234v1 90247v1 90273v1 90279v1 90290v1 90294v1 90301v1 90312v1 90533v1 90547v1 90565v1 90808v1 90860v1 90867v1 90868v1 90893v1 92026v1 92202v1 92397v1 90028v1 90023v1 90221v1 90330v1 90266v2 90565v1 90272v1 90182v1 90813v1' 
 fldnm='FS4 Oph_n9a FS13 BRI0021-0214 M67 P091-D S121-E P161-D LHS191 S217-D S234-E P247-U S273-E S279-F P290-D S294-D S301-D S312-T P533-D L547 P565-C S808-C S860-D S867-V T868-53850 S893-D LHS2026 BRI2202-1119 LHS2397a FS28 FS23 P221-C P330-E P266-C P565-C P272-D P182-E S813-D'

 MINNFDSETS =  2          Minimum number of calibration field sets to make a 
			  calibration observation
*MINNCS   =    6          Minimum number of calibration scans in field to make 
			  a valid field set
 MAXSEP   =    2.2 hours  Time span between endtime of last calibration 
			  scan and beginning of next calibration scan
*MAXDSPNSCN = .04 mag     Maximum magnitude dispersion for all sources between 
			  same-field calibration scans
*MAXDSPFD =   .04         Maximum calib delta magnitude dispersion for all 
			  standards within calibration scan

Other parameters which may need to be defined in addition to the above.

 exmaglk    |               | extinction calc. magnitude lower limit k band
 exmaguk    |               | extinction calc. magnitude upper limit k band
 exmaglj    |               |  "          "      "       lower limit j band
 exmaguj    |               |  "          "      "       upper limit j band
 exmaglh    |               |  "          "      "       lower limit h band
 exmaguh    |               |  "          "      "       upper limit h band


 * Values that may need to be tuned. 
 & Values that need to be computed and require some effort
   look at several nights where a cal field was taken at a range of airmass
   and fit linear function to delta mag vs. airmass. 


14. DBMAN


Cognizant Team Member: Carol Lonsdale

Cognizant Engineer: Tracey Evans

Function:

Load processed source and image data into survey databases.


DB architecture (table fragmentation, etc.)

intervals at which data is loaded and re-indexed; update stats run

QA parameter threshold to determine whether a scan is "good"

For extended srcs -- which mags are "default" mags (see below); which
  parameters to get rid of.
     For each band, a "default" magnitude is selected by DBMAN and is used to
     calculate the mag colors J-H,...R-K,...etc.  For the "regular" ext. srcs,
     the < b>_m values are the mags in each band at 20 mag/sq.arcsec w/fiducial
     elliptical aperture (< b>_m_i20fe).  If, however, this mag in any band is
     > 75 (neg. flux or no mag found), the < b>_m values will be the mags from
     fixed circular apertures with radii = 10 arcsec (< b>_m_10).
     If one of these mags is > 75, the "default" mag
     will be set to NULL.  
         For the LCSB srcs, the < b>_m values are the mags
     from fixed circular apertures with radii = 10 arcsec (< b>_m_10).  If one
     of these mags is > 75, the "default" magnitude will be set to NULL.  
         If there is no detection of the src in one or two bands, the value
     of the mag(s) in that band(s) is the band-fill aperture-photometry mag
     (not yet available).
         The < b>_msig fields contain the corresponding default mag sigmas.

15. QUALITY


Cognizant Team Members: Mike Skrutskie & Roc Cutri

Cognizant Engineer: Linda Fullmer

Retrieve Subsystem Design Specification Document

Retrieve 24HR QA Requirements Document

Retrieve 2MASS Quality Assurance and Data Validation Overview

Function:

Telescope and camera system health monitoring. Data validation and quality assurance.



1. Thresholds for 24-Hour Monitor
___________________________________________________________

        Observation Date  970731n

-------------- pointing summary --------------

PctNpm (Percent failed position reconstruction) <  15%

Pointing:
  AvDra (Average delta nominal RA to reconstructed)   <  20.000
  SigDra (RMS on above)  <  15.000

  AvDdec (Average delta nominal Dec to reconstructed) <  20.000
  SigDdec (RMS on above) <  15.000

  AvXSDif (Average X-Scan difference - south edge)  <    10.000
  SigXSDif (RMS on above)  <  10.000

  AvISMrg (Average In-Scan difference - either edge) >  -10.000
  SigISMrg (RMS on above) <  10.000

-------------- scanning summary --------------
Scanning:
  dtheta (Average rotation angle) 0.250 - 0.350
  sigthet (RMS on above) < 0.020

Offsets:
  avyoff (Average in-scan step size) | 82.400| - | 82.800|
  sigyoff is not tested.

  avxoff is not tested.
  sigxoff is not tested.

Tracking:
  maxdra (Average max. delta reconstructed frame center to 1st frame) <  15.000
  rmsdra_first (RMS on above) <  10.000

-------------- seeing/focus  summary --------------

  Avshape (Average image shape parameters) tested only if sigZP <  0.050
   in which case Avshape    <  1.10
                  OR   1.04 < FWHM_j / FWHM_h < 1.09
                    and
                       0.95 < FWHM_h / FWHM_k < 1.05
  sigshape is not tested.

  AvAspct (Average Moment Ratio) is not tested
  MnAspct (Minimum Moment Ratio) >   0.900
  MxAspct (Maximum Moment Ratio) <   1.100

-------------- photometric summary --------------

  AvZP (Average photometric zero point offset) is tested only if sigZP >   0.050
   in which case AvZP  > -0.300
  sigZP (RMS on above) <  0.050

---------- dark/responsivity summary ------------

  diffdark (Average delta nightly & canonical dark frames)  < 150.000
  sigdark (RMS on above) < 150.000
  Ndrkmask (Number pixels masked) <  2500
  diffresp (Average delta nightly & canonical responsivity) <   0.010
  sigresp (RMS on above) <   0.050


2. Scan Overlap Parameters
--------------------------------------------------------------------------

psf clearance angle = 4.0 arcsec
low coverage area   = 428.0 arcsec at ends of scans
match radius        = 2.0 arcsec
hi/low snr cutoff   = 20.0 snr
number of bad frames= 2 (2 bad frames is max allowed)
persistence thresh. = .50 (if Tracey says % chance it's persistence)

3. Nightly Q/A Thresholds
--------------------------------------------------------------------------

4. Scan Q/A Thresholds
--------------------------------------------------------------------------



Last update - 3 September 1997
Roc Cutri - IPAC