Comparison of Second V3 Ops Test to V2 Processing for 2MASS Galaxies
T. Chester and T. Jarrett Conclusions as of 8/23/01:
BOTTOM LINE: we have found no problems that impede declaring galworks ready to go, and using the current v3 ops db as the true beginning of the reprocessing.
Summary of analysis results:
- There is a "dupe source bug" in galworks that is of no consequence, being fixed by the dupe removal processor.
- V3 sources have substantially the same position and aperture flux as in V2. The k20fe mags are systematically fainter by 2-4% as a function of magnitude due to incorporating the actual calibration, which results from slightly smaller, but more accurate, radii to that isophote. These are all good things.
- V3 goes deeper into the galactic plane, again a good thing, as intended, since all the sources produced by GALWORKS in the plane can be checked manually.
- V3 produces about 10% more sources in low density areas above k20fe = 13.5 mag. The extra sources are largely good sources, as verified by visual examination by TJ.
- V3 is much improved for big galaxies. Their positions are now correct, which means their fluxes are much more accurate.
We intend to complete the comparison of v3 and v2, dotting all the i's and crossing the T's above, as well as do an analysis of how many v2 sources are still found in v3. (A database bug has prevented an accurate comparison of the entire V3 ops db with the V2 processing.) This analysis can proceed during the ops processing.
The above does not include the analysis of the abell repeats, done by tj, which also shows no problems with v3.
Change in Number of Sources from V2 THIS IS A PRELIMINARY, WORKING DRAFT. WE ARE WORKING ON ALL THE STATEMENTS IN RED BELOW. DO NOT BELIEVE ANY OF THEM UNTIL THEY BECOME STATEMENTS IN BLACK!!
We have begun our comparison of the v3 processing to the v2 processing by analysis of the second v3 ops test, processed in August 2001. See Preliminary unfinished analysis of the first v3 ops test for some analysis that has not been repeated.
For this analysis, we have analyzed all sources, galaxies plus other extended objects, that have k_m_20kfe < 13.5 mag. In order to unambiguously match sources with their v2 counterparts, we selected sources away from any coadd edge via:
50 < x_coadd < 460 and 50 < y_coadd < 974.
A total of 34,472 sources satisfied these constraints. Of those, only 27,406 found matches in the v2 wsdb from the same hemisphere, date and scan with positions within 5" of the v3 position. Thus only 80% of the sources are unambiguously matched in the v2 processing, a surprisingly low number.
Investigation reveals that all sources from nights 9/2/2000 and later did not find matches in the v2 WSDB. We have forwarded this to Tracey Evans and the IRSA staff for followup.
A plot of the percentage of unmatched sources vs. night and a histogram of the number of unmatched sources for each night (truncated at 25%) shows that most nights have a percentage of unmatched sources below 10%, with a tail out to ~20% unmatched sources. The nights after 7/7/2000 account for the point with 100% unmatched sources, and the night of 3/21/1998 is the outlier with 76% unmatched sources. In the following analysis, we exclude the nights after 7/7/2000 as well as the anomalous night of 3/21/1998.
A plot of the number of unmatched sources vs. density shows a smooth distribution. (See the number of total sources vs. density for comparison.) Low-density scans, density < 3.5, have < ~10% unmatched sources. In contrast, the percentage of unmatched sources increases for scans with densities above 3.5, becoming almost 100% at the highest densities. This is an expected result of the V3 processing, which was designed to go a bit deeper into high source density regions than was done in V2.
(analysis to come of the % unmatched vs. mag)
Curiously, two sources each found two identical matches in the v2 wsdb. Tracey Evans tracked these down for us, and found that these sources were output by GALWORKS from the same coadd in the same scan, with identical positions. Similar sources have been found by Tracey in the current processing, and T. Jarrett is investigating them.
Photometry Of the matched sources, we selected only sources with clean jhk_k20fe flags, leaving 20,993 sources. Of those, 19,539 sources had matches within 1", necessary to provide good photometric comparisons. Of those, 19,474 sources had measured fluxes in all three bands (no nulls).
(need more intro here to the comparison below)
The V3 ops database includes a fairly wide range of density: Plot of k_k20fe vs. density (expanded scale). It is curious that the brightest sources are found in the low density areas in both V2 and V3. This is probably due to a few bright stars being intentionally processed in low density areas in this database.
Plots of the difference in the k20fe magnitudes are given below:
Selection J H K All sources J H K All sources, expanded scale J H K Density < 3.5, expanded scale J H K Plot of the median change in delta mag vs. mag.
Comparison of the photometric measures shows that the radius = 7" aperture magnitudes are essentially unchanged for h and k for all magnitude bins, with means of ~0.001 ± 0.001 mag, and only slightly changed for j, with means of ~-0.005 ± 0.001 mag.
However, the magnitudes for the K=20 mag isophotal elliptical apertures show a significant change of 0.020 ± 0.002 mag for brighter sources, which increases to ~0.040 ± 0.002 mag for fainter sources. Because the changes are highly correlated in all three bands, the change is likely due to the definition of the k fiducial ellipse.
V3 GALWORKS processing uses an improved calibration to define the 20th magnitude isophote. V2 was forced to use an isophote defined using a fixed calibration which did not take into account nightly changes in the calibration, because the nightly calibration was not available at the time the processor was run. For V3, GALWORKS uses the actual calibration derived from the V2 processing.
(add reason for bias and calculation of effect)
Hence our conclusion is that all is well with the V3 photometry.
leftover from old comparison to be updated:
A total of 3058 sources satisfy the above constraints. We matched these against the v2 WSDB using a maximum radius of 6", and demanded that the date, scan and hemisphere be identical for each source. Only 2096 v2 matches were found, 69% of the v3 sources.
It is likely that most of the other 31% of the v3 sources also have counterparts in the v2 processing, but simply have changed positions by more than 6". A key change for the v3 processing was to no longer use positions for the brightest sources that were pre-determined from optical catalogs. A spot check of some of these sources indeed found matches just outside a radius of 6". need to check a few more sources - only one source checked so far. Furthermore, a check by Steve Schneider of one of the sources found that the new position was correctly centered on the galaxy, whereas the v2 position was off-center.
Our goal in this first analysis is the compare the photometry between v3 and v2. Specifically, we analyze the changes in the 7" radius aperture photometry, jhk_m_7, and in the elliptical k fiducial 20th mag per square arcsec isophotal magnitude, jhk_m_k20.
Therefore, we demand that the positions be identical to within a radius of 1". Accepting larger position changes would make the apertures significantly different in their center position and complicate the comparison. According, of the 2096 sources with matches, we eliminated the 225 sources with position changes between 1 and 6".
We wish to eliminate any effects of source contamination, and hence eliminate the 707 sources that were flagged in one or more bands for the isophotal magnitude in the new data set.
Finally, 5 sources were eliminated that had null mags at one or more bands in the v2 processing. TJ: what changed?
These demands left 1159 clean gals, 38% of the original 3058 sources.
The plot of delta mag vs. mag looks similar for all three bands. The vast majority of sources have photometry that is very little different between v2 and v3.
Only four sources have changes larger than 0.25 mag:
delta j_20 delta h_20 delta k_20 -0.6 -0.7 -0.4 0.2 0.2 0.3 0.3 0.3 0.3 0.4 0.4 0.4 Because the outliers are highly correlated in all three bands, the change is likely due to the definition of the k fiducial ellipse. Followup revealed that this was the case??? And if so, why did the ellipse change? Subtracted stars? In contrast, the 7" radius aperture photometry has no outliers.
Expanding the scale of the plot of delta mag vs. mag reveals a bias between v3 and v2 isophotal mags of ~0.02 mag, which does not depend on magnitude. The v3 photometry is ~2% fainter on average. This bias does not show up in the 7" aperture magnitudes.
The bias and population sigma for each photometric measure in each band are given in Table 1:
Statistic j20 h20 k20 j7 h7 k7 mean 0.02 0.02 0.02 -0.01 0.00 0.00 sigma(pop) 0.05 0.05 0.04 0.03 0.03 0.02 There is no apparent trend vs. density in the photometric changes (see also plot with expanded scale).
Investigation shows that the bias results from the improved calibration in v3 processing??? Expand if so.
The scatter of 2-3% for the 7" photometry is also due to calibration??. The increased scatter of 4-5% for the isophotal photometry is due to the propagated error of the radius determination (see Error Tree For 2MASS Galaxy Photometry).
Thus we conclude that v3 processing is better than the v2 processing, and has not introduced any problems in photometry.
It is interesting to note that the new star subtraction algorithm in v3?? the star subtraction in v3 not leaving a flag?? has greatly decreased the number of flagged sources. Of the 1159 galaxies, only 818 of these were also not flagged in v2; 341 (29%) were flagged in one or more bands in v2.
Positions A plot of position angle vs. radial distance separation for these 1159 sources shows that the changes in position were mostly less than 0.2", and fairly uniform in position angle for those sources. However, for sources with changes in position > 0.2", the position angle is primarily 0 or ±180 degrees, with a few at ±90 degrees. Spot checks show that this change is a result of the position reconstruction.
As an aside, the curious half circles in the above plot are a complete mystery. Gene Kopan suggested making a delta RA / delta dec plot, which reveals that delta ra tends to be quantized in steps of ~0.05" (not 100%, but often), but delta dec seems not to be! We are still tracking down the source of this quantization.
