V3 Verification of GALWORKS
Notes & Log

T. Jarrett, IPAC
updated Sept 5, 2001

Subject: Symmetry and Edge Flares ; Date: Late March, Early April

Cong Xu and Jarrett briefly explored the issue of galaxy symmetry. GALWORKS computes several "symmetry" measures, including CHI^2 fits to the ellipse modeling, flux ratios across the semi-major axis, and cross-corration (across semi-major axis). Symmetry details can be found here.

Our goal was to find galaxy pairs and "disturbed" galaxies, as well as take a census of artifacts, in particular those generated from bright stars. Our qualitative study did reveal asymmetric galaxies (no surprise) and did bring to the fore sources generated from bright stars. The brighter the asymmetric source, the more likely it is an artifact. The most annoying artifacts are the "edge flares". These derive from bright star located just beyond the scan edge. Here is an example of an edge flare. Jarrett modified GALWORKS to better deal with edge flares based on this work. The module edge_flare is designed to eliminate these artifacts without killing inclined galaxies. This work has been suspended until Cong returns from China.

Subject: Bright Star Masking ; Date: April 20-26

The original bright star masking method relied upon PIPELINE V2 photometry for bright stars. Very bright stars that saturate in R1 will not have reliable mags (nor by association, masking radii) for V2 reduction. The improved PIPELINE V3 estimates the integrated flux of R1-saturated stars by integrating and extrapolating the radial profile with a functional fit (work by Cong Xu and Gene Kopan). Very bright stars will have reliable mags under most conditions. Analysis of the bright star masking for these types of stars reveals underestimation of the halo radius. A slight scaling adjust is therefore made for the bright end of the halo-radius masking scale. The results look very good. See for some details.

Subject: Abell 3558 Repeatibility Tests ; Date: April 26

The Abell 3558 repeat data sets and RTB tests are detailed in Repeatibility Tests. The tables below represents the latest (mostly stable) GALWORKS V3, and comparison with V2 repeatibility. Note: these tests were carried out on data sets generated by the V2 pipeline (but with V3 GALWORKS) and hence are not a comprehensive test of the V3 pipeline. These tests will come later when the RTBs have been generated.

ABELL 3558 Repeatibility, GALWORKS V3
-- -- -- non-contaminated ... | contaminated ... |
fname mag- mag+ #r #p C #r #p C
J7_comp_1.tbl 14.00 15.00 2518 2676 0.94 814 969 0.840
H7_comp_1.tbl 13.70 14.20 1190 1374 0.87 452 518 0.873
K7_comp_1.tbl 13.00 13.50 1415 1398 1.01 200 233 0.858
J7_comp_2.tbl 14.00 15.00 20 239 0.08 16 197 0.081
H7_comp_2.tbl 13.70 14.20 10 135 0.07 1 11 0.091
K7_comp_2.tbl 13.00 13.50 5 13 0.38 0 0 0.000
fname val- val+ # valbar RMS # valbar RMS
J7_repeats_1.tbl 14.00 15.00 118 14.45 0.068 35 14.58 0.093
H7_repeats_1.tbl 13.70 14.20 53 13.95 0.104 18 14.04 0.118
K7_repeats_1.tbl 13.00 13.50 75 13.23 0.092 10 13.22 0.094
Jf20e_repeats_1.tbl 14.00 15.00 110 14.48 0.079 36 14.57 0.101
Hf20e_repeats_1.tbl 13.70 14.20 52 13.94 0.098 14 14.04 0.111
Kf20e_repeats_1.tbl 13.00 13.50 71 13.24 0.111 10 13.18 0.112
Jf20c_repeats_1.tbl 14.00 15.00 113 14.44 0.072 38 14.60 0.117
Hf20c_repeats_1.tbl 13.70 14.20 53 13.94 0.107 21 14.04 0.143
Kf20c_repeats_1.tbl 13.00 13.50 71 13.23 0.101 12 13.19 0.137
RKf0e_repeats_1.tbl 10.00 20.00 31 12.78 1.297 27 13.62 2.210
RKf0c_repeats_1.tbl 10.00 20.00 11 12.38 1.073 19 13.03 2.259
J_ba_repeats_1.tbl 0.20 0.80 30 0.58 0.072 106 0.57 0.073
K_ba_repeats_1.tbl 0.20 0.80 32 0.59 0.070 110 0.59 0.082
S_ba_repeats_1.tbl 0.20 0.80 26 0.55 0.073 107 0.57 0.076
J20e_Jext_repeats_1.0.01 0.80 192 0.18 0.089 96 0.22 0.096
H20e_Hext_repeats_1.0.01 0.80 204 0.18 0.104 90 0.20 0.113
K20e_Kext_repeats_1.0.01 0.80 210 0.16 0.096 89 0.18 0.097
RkronE_repeats_1.tbl10.00 20.00 80 12.51 1.947 71 13.20 2.292
JkronE_repeats_1.tbl14.00 15.00 94 14.41 0.088 49 14.56 0.127
JkronE_repeats_1.tbl13.70 14.20 41 14.01 0.072 23 14.01 0.093
JkronE_repeats_1.tbl13.00 13.50 17 13.25 0.049 16 13.27 0.076
RkronC_repeats_1.tbl10.00 20.00 25 12.01 1.707 40 13.02 2.006
JkronC_repeats_1.tbl14.00 15.00 99 14.41 0.089 49 14.52 0.119
JkronC_repeats_1.tbl13.70 14.20 39 14.01 0.070 22 13.98 0.090
JkronC_repeats_1.tbl13.00 13.50 19 13.27 0.059 16 13.25 0.086
J20e_Jkron_repeats_10.01 0.50 174 0.08 0.043 92 0.11 0.050
H20e_Hkron_repeats_10.01 0.50 181 0.08 0.050 86 0.11 0.051
K20e_Kkron_repeats_10.01 0.50 175 0.08 0.061 80 0.11 0.064


V3 vs. V2 for Abell 3558 Repeats
-- -- -- non-contaminated ... | contaminated ... |
fname mag- mag+ #r #p C #r #p C
J7_comp_1.tbl 14.00 15.00 2518_2023 2676_2180 0.940_0.930 814_1269 969_1536 0.840_0.826
H7_comp_1.tbl 13.70 14.20 1190_1075 1374_1193 0.870_0.900 452_658 518_795 0.873_0.828
K7_comp_1.tbl 13.00 13.50 1415_1277 1398_1276 1.010_1.000 200_449 233_514 0.858_0.874
fname val- val+ # valbar RMS # valbar RMS
J7_repeats_1.tbl 14.00 15.00 118_93 14.450_14.470 0.068_0.069 35_52 14.580_14.480 0.093_0.076
H7_repeats_1.tbl 13.70 14.20 53_51 13.950_13.970 0.104_0.105 18_25 14.040_13.940 0.118_0.110
K7_repeats_1.tbl 13.00 13.50 75_62 13.230_13.240 0.092_0.095 10_18 13.220_13.240 0.094_0.098
Jf20e_repeats_1.tbl 14.00 15.00 110_83 14.480_14.500 0.079_0.079 36_55 14.570_14.480 0.101_0.098
Hf20e_repeats_1.tbl 13.70 14.20 52_44 13.940_13.970 0.098_0.105 14_24 14.040_13.950 0.111_0.122
Kf20e_repeats_1.tbl 13.00 13.50 71_59 13.240_13.240 0.111_0.123 10_18 13.180_13.200 0.112_0.132
Jf20c_repeats_1.tbl 14.00 15.00 113_83 14.440_14.470 0.072_0.075 38_58 14.600_14.460 0.117_0.089
Hf20c_repeats_1.tbl 13.70 14.20 53_44 13.940_13.960 0.107_0.110 21_31 14.040_13.940 0.143_0.116
Kf20c_repeats_1.tbl 13.00 13.50 71_55 13.230_13.240 0.101_0.113 12_22 13.190_13.210 0.137_0.130
RKf0e_repeats_1.tbl 10.00 20.00 31_30 12.780_12.020 1.297_1.607 27_47 13.620_13.260 2.210_1.706
RKf0c_repeats_1.tbl 10.00 20.00 11_9 12.380_12.420 1.073_1.452 19_35 13.030_12.690 2.259_1.632
J_ba_repeats_1.tbl 0.20 0.80 30_28 0.580_0.490 0.072_0.061 106_75 0.570_0.550 0.073_0.075
K_ba_repeats_1.tbl 0.20 0.80 32_30 0.590_0.520 0.070_0.068 110_84 0.590_0.570 0.082_0.096
S_ba_repeats_1.tbl 0.20 0.80 26_27 0.550_0.470 0.073_0.054 107_77 0.570_0.550 0.076_0.061
J20e_Jext_repeats_1.0.01 0.80 192_89 0.180_0.320 0.089_0.115 96_104 0.220_0.350 0.096_0.114
H20e_Hext_repeats_1.0.01 0.80 204_76 0.180_0.310 0.104_0.119 90_83 0.200_0.320 0.113_0.119
K20e_Kext_repeats_1.0.01 0.80 210_49 0.160_0.290 0.096_0.116 89_52 0.180_0.270 0.097_0.106


V3 Repeatibility

Subject: V2 vs. V3 Photometry ; Date: April 27

A direct comparison between V2 photometry and the new V3 photometry can reveal potential problems, such as (1) systematics in the calibration, (2) background removal differences and (3) non-uniformities due to the new V3 methods (e.g., star removal). Star-removal non-uniformities are revealed by comparing the photometry across fields with different stellar number density. The first set of results will be separated according to the night (and hence, hardware and calibration period dependent).

Note the slight, ~2%, systematic difference in photometry. This is undoubtedly due to the zero-point calibration correction. Here is a note from Roc: There is no evidence for a flux bias with stellar number density.

Subject: M33 & M51 and Extended Sources ; Date: April 27

M33 brings out the worst in GALWORKS. We detect the galaxy all right, but every piece of the galaxy. The chaff is extensive.

Many of the detections are extended (probably H II regions and clusters), most are point-like (although associated with M33). The mags are undoubtedly grossly incorrect due to the poor background subtraction (M33 itself). Question: do we want or need this in the extended source DB? It is a byproduct of attempting "normal" reduction of large galaxies. We do have the option of eliminating these sources during the cat generation process, where we would apply a masking radius around M33.

Subject: Disk Usage ; Date: May 1, 2001

Two major changes to GALWORKS may significantly effect the disk usage. (1) GALWORKS V3 generates "coverage" maps; see Coverage Maps for some information. The image cubes are 6 planes deep, the first three planes are the GALWORKS masked areas and the last three planes are the PIXPHOT "coverage" maps. The images are 8-bits deep (i.e., short integers) to save space (they compress very efficiently). Here is an example of a GALWORKS coverage cube. The top three panels show the GALWORKS masking for J, H and Ks, respectively, of this Abell 3558 field (J-band coadd). The bottom three panels show the corresponding PIXPHOT coverage for J, H and Ks, respectively. The darkest grey scales correspond to a maximum coverage of 7. The typical coverage is 6. The lighter greyscales correspond to 5 and less coverages; (2) the galaxy postage stamp images are now 6 planes deep, the first three are nominal and the last three include star removal. Here is an example of a 3-color image showing the first three planes and the second three planes. Here is an example of a new postage stamp FITS header.

Coverage Maps

Gstamp Images Extrapolation

Summary

We can expect about 120 Gb of compressed "coverage maps" from GALWORKS, assuming a total of 70,000 scans. For galaxy stamps, we can expect between 100 and 200 Gb of compressed images. The low number corresponds to "typical" scans (low density, field galaxies). The high number incorporates galaxy clusters and fields full of nebulosity and complex structure (e.g., Orion). The actual disk usage will fall somewhere between these two limits.

Subject: V2 vs. V3: Ellipse Fit; Date: May 9

Low Density Fields

Subject: Coverage/Masks Date: May 23

Subject: Follow up on Coverage/Masks Date: July 9

Subject: V3 Ops Testing; Date: July 16

Abell 3558 Repeatibility from the OPS Testing Phase

Repeatibility results show that the performance (completeness, reliability, photometry) of GALWORKS is nominal.

Outstanding problems:

Both issues have been resolved. Delivery of modified GALWORKS in progress.

See OPS tests results below.

Subject: V3 Ops Testing: coadd masking Date: Aug 1

Report from Cong Xu:

I checked 2 most recently processed nights:

They look quite OK. The problem of too small masks for high density scans disappeared. For very bright stars (perhaps brighter than 3 mag) the mask size looks just fine (example: dino:/b2/V3test/000406s s128:f198). Though for less bright saturated stars the masks look a bit too large (examples: dino:/b2/V3test/000406s 031:257, 128:033, 128:245), I guess it is OK. The low density scans are fine.


Subject: final V3 Ops Testing; Date: Sept 5

Abell 3558 Repeatibility from the OPS Testing Phase

Repeatibility results show that the performance (completeness, reliability, photometry) of GALWORKS is nominal. The completeness and reliability level-1 requirements (for |b| > 20 deg) are achieved. The 10% photometric precision requirement is achieved (for Jiso < 14.5)