Standard QA Diagnostics
Shows the zero-point determination for J (top panel), H (middle), and Ks (bottom) at a function of UT time. Fits are shown for --
Computed uncertainties are shown above each photometric interval --
Note: if there are fewer than five cal sets, the polynomial fit will be shown by a flat line at jzp=0.
Listed above these uncertainties are the maximum allowable time interval between adjacent cal sets (black bar) and the fit that was actually used (text). For the latter, defaults are piecewise for J and linear for H and Ks.
Under each cal set are four fields --
Color coding for the cal sets is --
Photometric Overlaps Summary Shows the average delta magnitude between adjacent scans if they have a region of common sky coverage, plotted against UT time. Statistics are shown for J (top panel), H (middle), and Ks (bottom) and are computed for source having SNR>20. (Actually, specific magnitudes are hard coded into this averaging rather than an actual SNR computation, but those mags closely correspond to the SNR=20 limits we derived in v2 processing.) Only sci scan overlaps are shown. Photometric downgrades occur whenever the peak-to-peak difference within a sci block exceeds 0.05 mag. (This downgrade can be overridden if, for example, the sci block consists of two or more contiguous areas that are distantly separated on the sky.)
The first page of this plot has a vertical scaling of +/-0.1 mag. The second page has a scaling of +/-1.0 mag and is used to look for bad outliers lost off the first plot.
Shows the Photometric Sensitivity Parameter, or PSP value, as a function of scan number for J (top panel), H (middle), and Ks (bottom). More on the derivation of this value can be found in Section VI.2.
This parameter calculates the probability that we have reached our level-1 specs for sensitivity. Horizontal dashed lines on the plot show regions above which there is a strong possibility that we have not met our requirements.
Points are coded as follows --
The horizontal lines are hemisphere and date dependent, as detailed in http://spider.ipac.caltech.edu/staff/hurt/2mass/docs/qa_rules.html. Also discussed there are special rules sometimes used in scoring.
Shows in the three panels the aspect ratio (I2min/I2maj), orientation of major axis (Position Angle, East of North), and average seeing (Seeing Shape) for each scan. Values for each camera - J (blue), H (green), and Ks (red) - are plotted.
Downgrades are warranted when the aspect ratio drops below 0.81 or the average J-band seeing shape exceeds 1.25.
Meteor trails are identified and actively masked out in the pipeline. The meteor diagnostic allows the reviewers to visually scan all incidents of blanking to verify the nature of the blanked object and to allow for overriding masking instances that are inappropriate and lead to unacceptably low coverage.
Each frame triggering meteor streak removal is displayed in 3 adjacent J/H/K panels. Endpoints of the identified streak are marked with green circles. Under the frames is the output from the corresponding streakr table summarizing useful reference information including:
Streaks will be blanked in all bands even if only detected in one or
two of them. The masked region will be extended along the streak to the frame
edges to catch entire meteor trails even if only a portion is identified.
The exception to this rule is when the streak is identified in three or more
consecutive frames; in this case only the identified length of the streak
will be masked (this prevents slow-moving satellites from creating 0 coverage
areas along their path as they show up in sequential frames).
This plot summarizes overall astronometric consistency in the scan-scan overlaps. RA offsets are in red and Dec offsets are in blue. Diagnostics are plotted vs. scan number to aid in identifying questionable overlaps.
The top panel shows the worst average offset within a scan. Trimmed averages are computed for 12 bins (evenly-spaced in Dec) along the scans and the worst offset is plotted. Reference fidutials are drawn at +/- 0.5".
The bottom panel shows the fraction of the overlapping stars in the adjacent scans that exceed the offset thresholds of 0.3" (crosses) and 0.6" (boxes). This plot gives an overall idea how extensive the astrometric offsets are in the scan.
Shows histograms of 2MASS vs. Tycho differences in the left two panels:
Tycho-1 = solid lines
Tycho-2 only = dotted lines.
Right panels include scan overlap differences for sources with SNR>20. Cross-scan histograms are on the top and in-scan histograms on the bottom. Means, sigmas, and counts are shown at the upper left in each panel. Howard says that standard deviations exceeding 150 milliarcseconds may indicate a problem on each panel (ignoring Tycho-2 stats in the left plots).
Shows mean position differences as a function of Ks-band magnitude. This plot is intended primarily to monitor any position differences arising from the different Read1 and Read2 paths through the pipeline processing. On the left are plotted mean differences between 2MASS and adjusted* USNO-A2.0 positions as a function of Ks magnitude. Cross-scan differences are in the upper-left plot with in-scan differences in the lower-left. The same plots are repeated on the right, only this time the differences are between final 2MASS positions coming out of PosPts and the positions determined back in PosFrm from frame extractions. Ideally these plots would be flat lines at zero. Deviations from zero for the left plots could indicate problems with either 2MASS or USNO-A2.0 (or both). Deviations from zero in the right plots reflect 2MASS internal problems. A step near mag 8.5 in the right-hand plots is indicative of a bias between the Read1 and Read2 astrometry.
* USNO-A2.0 positions have been adjusted to remove systematic in-scan and cross-scan differences as a function of in-scan position.
Shows the quality of distortion removal. The two plots on the left monitor mean differences between 2MASS and adjusted* USNO-A2.0 positions as a function of cross-scan position. Cross-scan differences are in the upper-left plot with in-scan differences in the lower-left. The two plots on the right monitor the consistency of the distortion removal done in PosFrm and ProPhot. Mean cross-scan differences can be found in the upper-right plot and mean in-scan differences in the lower-right. Ideally these plots would be flat lines at zero.
* USNO-A2.0 positions have been adjusted to remove systematic in-scan and
cross-scan differences as a function of in-scan position.
Sigma vs. Magnitude
This plot shows the estimated uncertainty (sigma) for stars with abs(glat)>20 and which have no flags set to indicate artifacts. The blue points (mag<5) are Read-1 saturated stars. Sigma's of >0.2 mag are expected for these stars. The green points show the uncertainty for Read-1 aperture magnitudes. These sources are saturated in Read-2, and their sigma's reflect the scatter in the six individual aperture magnitudes. The black points are the Read-2 sources. At high signal to noise ratio, they should have sigma's of about 0.02-0.03 magnitudes.
Chi^2 vs. magnitude
This plot shows the chi^2 values for stars with abs(glat)>20 which are clean Read-2 sources in all three bands. The red points show the trimmed averaged. The mean chi^2 value should stay near 1 for all magnitudes.
Chi^2 vs. Cross-Scan Position
This plots shows the trimmed average as a function of cross-scan position. Bright stars (J<=13, H<=12.5, K<=12) are shown in blue and fainter stars in red. All sources must be clean and detected in all three bands. Ideally, chi^2 would be near one throughout the detector, but in practice focus variations may manifest themselves.
(PSF-AP) vs. PSF Magnitude
This plot shows the difference between the PSF magnitudes and the Aperture magnitudes for Read-2 sources. The red points are the trimmed average. The mean should be zero (note that each PSF used in a scan is normalized to zero difference for bright sources) as a function of magnitude. The green lines mark differences of 0.01 magnitude. In some nights, we see a non-linearity at the few percent level.
(PSF-AP) vs. PSF Magnitude (East/West)
The trimmed averages as a function of magnitude as computed in the (PSF-AP) vs. PSF Magnitude plot, but now for all stars (black), the eastern (X<-200), and western (X>200). Ideally, all should be consistent and stay near-zero. In practice, this may not be the case if there are image variations across the chip.
(PSF-AP) vs. Cross-Scan Position
The trimmed averages as a function of cross-scan position for bright stars (J<=13, H<=12.5, K<=12) and fainter stars. The DMAGCOR correction should make the differences for bright stars zero across the array. The fainter stars should ideally also be near-zero as a function of cross-scan position but are not forced to be so.
This plot shows the differences between the Read-1 aperture magnitude and the Read-2 PSF magnitude. The red points are saturated in Read-2, so values away from zero are not a concern. The blue points are unsaturated in Read-2. Only single (non-deblended) clean sources are shown. The black points show the trimmed averages for the unsaturated (blue) stars. The means should be near zero and the points should transition between blue and red approximately near J=9.5, H=9.0, K=8.5. Deviations at the faintest magnitudes, where the Read-1 photometry is unreliable, are not a concern.
These plots show a square-root-stretched plot of the density of stars in the magnitude-color plot (left) and color-color plot (right). High galactic latitude is abs(glat)>20; moderate galactic latitude is 5<abs(glat)<20; low galactic latitude is abs(glat)<5. Only clean (no artifact flags) stars detected in all three bands with sigma<0.05 in at least one band are shown. If no stars fall into the galactic latitude range the plot is black.
Color vs. Cross-Scan Position
This plot shows colors as a function of cross-scan position for clean sources with abs(glat)>30. The requirements are 9.0<J<12.5; 8.5<H<12.5; 8.0<K<11.0: the J-H plot used the J and H magnitude limits, the H-K plot uses the H and K magnitude limits, and J-K plot uses the J and K limits. The trimmed averages are also shown. Note that H-K is most useful, since the typical J-H distribution is bimodal. No color trends should be evident.
Color vs. In-Scan Position
The same filtering as Color vs. Cross-Scan Position. Note that due to the glat cut, some nights may have lack points at certain in-scan positions if all the scans were near the glat cutoff.
Bright Star Color-Color Plot
Color-color plots for saturated Read-1 sources derived from profile fitting compared to standard Bessell and Brett color (green lines). The key gives the color coding by saturated band. The points should lie near the standard sequence. Errors of ~0.2 magnitudes are expected.
Bright star K vs J-K
Stars saturated in both J and K (black), J only (blue), and K only (red) are compared to the Read-1 unsaturated stars (small black points). Typically, The distribution of staturated stars should be a continuation of the unsaturated stars.
Bright star J vs. J-H
Stars saturated in both J and H (black), J only (blue), and H only (red) are compared to the Read-1 unsaturated stars (small black points). Typically, The distribution of staturated stars should be a continuation of the unsaturated stars.
Bright star H vs. H-K
Stars saturated in both H and K (black),
H only (blue), and K only (red) are compared to the
Read-1 unsaturated stars (small black points).
Typically, The distribution of
staturated stars should be a continuation of the
Galaxy Quality Diagnostic
This three panel plot presents on a scan-by-scan basis basic indicators of extended source extractions. The top two panels present averaged J-K and J-H colors with one point per coadd. Typical galaxy colors are indicated by the magenta fiducial marks. The points are color-coded by the total number of sources going into the average:
Galaxy Color-Color Plot
Extended source H-K vs. J-H colors are plotted. Bessel-Bret tracks and extinction vector are included for reference. Color coding indicates Galactic Latitude:
Galaxy Color-Mag Plot
Extended source H-K vs. J-H colors are plotted. Marker sizes/colors are as
Version 2 Comparison Diagnostics
Shows the changes in zero-points, PSP (sensitivity) values, and source counts between version2 and version3 processing. The different bands follow the standard color coding - J (blue), H (green), and Ks (red).
In the top panel the zero-points are shown for both the cal and sci scans. Any difference exceeding 0.05 mag is highlighted with a black star. The zero-points for cal scans are expected to change between the two versions and to be better determined for version3 due to a larger network of secondard standards within each cal field.
In the middle panel the PSP values should fall near zero except in cases where the dark subtraction between version2 and version3 is substantially different. Especially in the north, the average canonical dark for a period can be significantly different from the single-night value used before. This will result in displacements either upward or downward in one or more bands.
In the bottom panel shows the percentage gain or loss of total sources (no
filtering; all sources in the .cal file are counted) vs. the number found in
version2 processing. Also shown by the light blue points are the log of the
number of sources found at J-band during version2 processing.