Photometric Catalogue for Space and Ground Night-Time Remote-Sensing Calibration: RGB Synthetic Photometry from Gaia DR3 Spectrophotometry
, Nicolas Cardiel 4,5, Eduard Masana 1,2,3, Jaime Zamorano 4,5, Sergio Pascual 4,5, Alejandro Sánchez de Miguel 4,5,6, Rafael González 4 and Jaime Izquierdo 4Round 1
Reviewer 1 Report
See attached pdf
Comments for author File: 
Comments.pdf
Author Response
========
Corrections/questions:
Page 2
line 34: “that the BP instrument has”
MODIFIED
line 51: do not need the “st” after “21”
MODIFIED
line 53: suggest “allows” for “is allowing”
MODIFIED
line 55: the current sentence is awkward; I suggest rearranging - “The homogeneity of the spectrophotometry over all sky positions..”
MODIFIED
line 56: “catalogue” is used – keep uniformity in spelling with either “catalog” or “catalogue”
MODIFIED. CATALOGUE IS NOW USED EVERYWHERE.
Page 3
line 73: suggest “In its turn”
MODIFIED
Page 4
Overruns on a number of lines
line 105: “variable sources”
Page 8
line 177: I believe that “accurate” (quality of data), rather than “precise” was intended
MODIFIED
Page 9
line 185: awkward phrasing: suggest “being more important for the latter” or similar
MODIFIED
line 188: was “gating” or “grating” intended?
GATING IS CORRECT. IT CORRESPOND TO THE STRATEGY IN THE GAIA MISSION TO AVOID SATURATION READING ONLY PART OF THE CCD. THE GATES ALLOW TO RELEASE THE CHARGE AT THE MIDDLE OF THE READING PROCESS AND EFFECTIVELY REDUCE THE EXPOSURE TIME. A REFERENCE TO GAIA PAPER IS ADDED TO THE TEXT FOR MORE INFORMATION.
lines 192-3: this sentence is awkward – I suggest that “The method for deriving the synthetic RGB magnitudes explained in Sect. 2 allows for the derivation of the associated uncertainties and it is possible…”
MODIFIED
lines 198 & 201 (also Fig 11 caption): “99th percentile”
MODIFIED
line 201: suggest replacing “which” with “with values of”
MODIFIED
Page 16
Table 2: (small point) the quoted precision of 10 micromagnitudes seems excessive given the issues in calibration etc. Would 1 mmag be more appropriate?
THREE DECIMALS ARE NOW KEPT FOR EVERY VALUE IN THE TABLE.
Page 18
line 283: suggest “rely better” be replaced with “more strongly”
MODIFIED
line 306: “rely” instead of “relay”
MODIFIED
Reviewer 2 Report
The authors present a very useful and carefully executed project to provide a very dense network of stars with improved photometric calibration in the native R,G,B system of many commercial cameras. The explanation of their methods is very clear, and they have made the catalog and a helpful search tool available for general use. The advantages of having some 200 million sources from Gaia with bandpass magnitudes synthesized from their spectra are clearly explored, along with the variations from the previous fits to Gaia bandpass magnitudes with more limited color range.
The authors are urged to consider the suggestions for minor revision below; the manuscript is suitable for publication, no matter how they choose to react to those suggestions.
The main question for consideration is this: are there certain classes of objects that should not really be included in a grid of calibration standards? The authors were clear in distinguishing certain subclasses of objects that are not well described as stars with single photospheric temperatures. One such class is that of quasar candidates. The best statistical models of large samples of quasars show that _all_ quasars vary in optical light on timescales as long as decades. Even though those quasars did not vary over the Gaia time baseline, they will. If the intention is to provide a standards grid for immediate, short-term use, then including them is fine. If the intention is for a dataset with persistent value, the authors should consider whether variables should be included.
Two subclasses of objects have measured magnitudes that depend on image quality and detector sampling. One is that of galaxies; their total measured light depends in detail on their intrinsic profiles, surface brightness sensitivity of the system and detector sampling. The other subclass is that of stars at very low Galactic latitude. The amount of blending, therefore the total brightness measured, depends on delivered image quality and sampling quite critically. It’s not obvious how to exclude those categories, because the range of applications of the catalog will vary considerably. The authors might consider providing a warning about those two classes that different applications might encounter different levels of accuracy when using such objects as reference calibrators.
More minor comments:
Fig. 1 – The legend needs to have a slightly clearer distinction between the solid vs. long-dash lines for G_Gaia and G_RP, and either the caption or the text should have some brief reiteration of the definition of the G-Gaia broadband magnitude.
Discussion of difference in calibration brighter and fainter than 11.5 mag on current line 186: was that in reference to the left-hand panels of Figs 9 and 10 rather than the bottom panels?
Those Figures 9 and 10 show lots of systematics at the sub-0.1 mag level. Is there any concern about the systematic offset at the ~0.4 mag level for the brighter, bluer stars in the R band between the spectrum-synthesized magnitudes and the bandpass-transformed (C21) magnitudes?
Author Response
The main question for consideration is this: are there certain classes
of objects that should not really be included in a grid of calibration
standards? The authors were clear in distinguishing certain
subclasses of objects that are not well described as stars with single
photospheric temperatures. One such class is that of quasar
candidates. The best statistical models of large samples of quasars
show that _all_ quasars vary in optical light on timescales as long as
decades. Even though those quasars did not vary over the Gaia time
baseline, they will. If the intention is to provide a standards grid
for immediate, short-term use, then including them is fine. If the
intention is for a dataset with persistent value, the authors should
consider whether variables should be included.
Two subclasses of objects have measured magnitudes that depend on
image quality and detector sampling. One is that of galaxies; their
total measured light depends in detail on their intrinsic profiles,
surface brightness sensitivity of the system and detector sampling.
The other subclass is that of stars at very low Galactic latitude.
The amount of blending, therefore the total brightness measured,
depends on delivered image quality and sampling quite critically. It’s
not obvious how to exclude those categories, because the range of
applications of the catalog will vary considerably. The authors might
consider providing a warning about those two classes that different
applications might encounter different levels of accuracy when using
such objects as reference calibrators.
WE DECIDED TO KEEP ALL THESE TYPES OF SOURCES AS FAR AS THEY WERE NOT IDENTIFIED AS VARIABLE BY THE GAIA TEAM IN ORDER TO INCREASE THE RANGE OF APPLICATIONS FOR THE FINAL CATALOGUE. SOURCES SHOWING STRONG DEVIATIONS CAN BE REJECTED AS OUTLIERS BY THE USER, THROUGH SOME STATISTICAL ANALYSIS, BEFORE BEING USED AS CALIBRATORS. IN ADDITION, WE HAVE ADDED TWO NEW COLUMNS IN TABLE 2, AS WELL AS IN THE FILE rgbloom_200m.csv GENERATED BY THE PYTHON CODE rgbloom, THAT PROVIDES TWO FLAGS INDICATING THE TYPE OF OBJECT AND THE GLOBAL QUALITY OF THE RGB ESTIMATES:
FLAG#1: objtype
1: object flagged as NON_SINGLE_STAR
2: object flagged as IN_QSO_CANDIDATES
3: object flagged as IN_GALAXY_CANDIDATES
0: none of the above
FLAG#2: qlflag
0: object with reliable RGB estimates
1: object with less reliable RGB estimates
THE AVAILABILITY OF THESE TWO NEW FLAGS SHOULD HELP THE POTENTIAL USERS OF THIS WORK TO MAKE A JUDICIOUS REMOVAL OF SUSPICIOUS CALIBRATING SOURCES. THIS INFORMATION IS NOW EXPLAINED IN SECTION 5.
TABLE 2 (AND ITS CAPTION) HAS ALSO BEEN UPDATED ACCORDINGLY.
More minor comments:
Fig. 1 – The legend needs to have a slightly clearer distinction
between the solid vs. long-dash lines for G_Gaia and G_RP, and either
the caption or the text should have some brief reiteration of the
definition of the G-Gaia broadband magnitude.
WE MODIFIED THE DARKNESS OF THE GREY COLOUR LINE FOR RP IN ORDER TO BETTER DISTINGUISH RP PASSBAND FROM THE REST OF GAIA PASSBANDS IN THE PLOT AND NOT ONLY RELY ON THE DIFFERENT DASHED LINE. THE DEFINITION OF GAIA VS RGB MAGNITUDES WAS ADDED TO CAPTION.
Discussion of difference in calibration brighter and fainter than 11.5
mag on current line 186: was that in reference to the left-hand panels
of Figs 9 and 10 rather than the bottom panels?
WE AIMED TO REFER THE RIGHT-HAND PANELS, NOT BOTTOM (OR LEFT, AS THEY SHOW THE DEPENDENCE WITH COLOUR, NOT MAGNITUDE). TEXT HAS BEEN MODIFIED.
Those Figures 9 and 10 show lots of systematics at the sub-0.1 mag
level. Is there any concern about the systematic offset at the ~0.4
mag level for the brighter, bluer stars in the R band between the
spectrum-synthesized magnitudes and the bandpass-transformed (C21)
magnitudes?
(?) NO 0.4 MAG LEVEL RESIDUALS CAN BE SEEN IN THE FIGURES. ONLY +-0.2 RANGE IS PLOTTED. AND THE BLUE AND BRIGHT STARS ARE NOT SPECIALLY HIGH RESIDUAL LEVELS IN R.
Reviewer 3 Report
This paper presents a new all-sky catalog with synthetic RGB photometry for more than 200 millions sources with low-resolution Gaia spectra. The analysis and scientific validation is solid and very convincing, and the importance of such a catalog for both the astrophysical community and the pro-am collaborators is undeniable, as commercial imaging detectors are becoming more and more performing and commonly used in our field. I think this paper is already in a very good shape and is suitable for publication with only minor modifications or clarifications, listed below.
-
Page 4, Fig. 3 & 4. I don’t see the point of discussing and showing the properties of the variable stars since they are eventually excluded from the photometric sample. Rather, I think it would have been more useful for the readers to give an estimate of the completeness ratio of the final sample, in particular as a function of magnitude.
-
Page 4: “we also aimed to select only sources with phot_variable_flag=’CONSTANT’ but no sources with Gaia DR3 XP spectra have this flag value”. Is this something unexpected, or that should have been reported to the Gaia collaboration? Is there any information from the Gaia documentation that can explain this?
-
Page 4: “...after removing 40 616 sources with unreliable RGB magnitude estimates”. What is the exact, quantitative criteria aplied for this selection process?
-
Page 4: “In a small subsample of this collection the RGB magnitude estimates are missing in some of the 3 bands [...] but we have kept the affected sources because in all these cases a magnitude estimate is available in at least one of the RGB bands”. Again, is there an identifiable reason for this? Can we trust the remaining band(s) if we are not sure they are affected to some extent by the same technical issue?
-
Page 4, Sec. 3.1, Fig. 5. The authors mention that there is a double peak in the histogram of the source density. The origin of the first peak on the C21 catalog is identified as due to directions of high extinction excluded by that work, but there is no discussion on why we see a double peak in the 200M catalog. Also, about the statement “99.7% of the celestial sphere is covered with a source density above 100 sources/(square degree) by this catalogue”: it should be mentioned that the 0.3% regions at very low densities are not randomly spread over the sky, but rather are concentrated on two well defined stripes at high Galactic latitudes (Fig. 5). Is there any technical reason for this?
-
Page 8. The 200M catalog explicitly contains stellar binaries and non-stellar sources such as QSOs and galaxies. Is there any reason not to discard these sources, since they are easily identifiable by the corresponding DR3 flags? I don’t see how an extended source, for instance, can be considered a good calibrator.
-
Page 8, line 177: I guess that “accurate” should be more appropriate than “precise” in this context.
-
Fig. 9. In the Delta G vs. BP-RP plot, there is a curious “bridge” at red colors (BP-RP>1.0). Is that due to some surface gravity effects in that band, i.e., dwarfs vs. giants separation? If so, it would be nice to mention that in the text.
-
Fig. 11. The red points are almost overlapped to the horizontal axis over most of the plotting range. These plots would be much clearer with a log scale on the vertical axis, but this is up the authors to implement.
END OF REPORT
Author Response
Page 4, Fig. 3 & 4. I don’t see the point of discussing and showing
the properties of the variable stars since they are eventually
excluded from the photometric sample. Rather, I think it would have
been more useful for the readers to give an estimate of the
completeness ratio of the final sample, in particular as a function of
magnitude.
Page 4: “we also aimed to select only sources with
phot_variable_flag=’CONSTANT’ but no sources with Gaia DR3 XP spectra
have this flag value”. Is this something unexpected, or that should
have been reported to the Gaia collaboration? Is there any information
from the Gaia documentation that can explain this?
THIS IS KNOWN BY THE GAIA COLLABORATION. THEY DID IT ON PURPOSE FOR GAIA DR3. THE GAIA DR3 CATALOGUE IS NOT YET THE FINAL ONE. IT IS EASY TO DETECT A VARIABLE SOURCE, BUT THE "CONSTANT" FLAG CAN ONLY BE SET WHEN THE WHOLE MISSION IS CONSIDERED IN ORDER TO BE SURE THAT THE SOURCE IS REALLY CONSTANT. A SENTENCE HAS BEEN ADDED TO THE TEXT IN ORDER TO CLARIFY THIS.
Page 4: “...after removing 40 616 sources with unreliable RGB
magnitude estimates”. What is the exact, quantitative criteria aplied
for this selection process?
Page 4: “In a small subsample of this collection the RGB magnitude
estimates are missing in some of the 3 bands [...] but we have kept
the affected sources because in all these cases a magnitude estimate
is available in at least one of the RGB bands”. Again, is there an
identifiable reason for this? Can we trust the remaining band(s) if we
are not sure they are affected to some extent by the same technical
issue?
THIS IS PROBABLY DUE TO VERY LOW SIGNAL PRESENT IN THESE WAVELENGTH RANGES IN GAIA SPECTROPHOTOMETRY. THIS EXPLANATION IS NOW INCLUDED IN THE TEXT.
Page 4, Sec. 3.1, Fig. 5. The authors mention that there is a double
peak in the histogram of the source density. The origin of the first
peak on the C21 catalog is identified as due to directions of high
extinction excluded by that work, but there is no discussion on why we
see a double peak in the 200M catalog. Also, about the statement
“99.7% of the celestial sphere is covered with a source density above
100 sources/(square degree) by this catalogue”: it should be mentioned
that the 0.3% regions at very low densities are not randomly spread
over the sky, but rather are concentrated on two well defined stripes
at high Galactic latitudes (Fig. 5). Is there any technical reason for
this?
THE LACK OF SOURCES IN SOME SPECIFIC AREAS OF THE SKY ARE DUE TO THE LACK OF BP/RP SPECTRA IN GAIA DR3 DUE TO LESS NUMBER OF TRANSITS FOR THOSE SOURCES. FUTURE GAIA RELEASES WILL INCLUDE MORE SPECTROPHOTOMETRY IN ALL AREAS OF THE SKY, AND ALSO IN THESE PATCHES. A SENTENCE WAS ADDED TO THE TEXT WITH THIS EXPLANATION.
Page 8. The 200M catalog explicitly contains stellar binaries and
non-stellar sources such as QSOs and galaxies. Is there any reason not
to discard these sources, since they are easily identifiable by the
corresponding DR3 flags? I don’t see how an extended source, for
instance, can be considered a good calibrator.
WE DECIDED TO KEEP ALL THESE TYPES OF SOURCES AS FAR AS THEY WERE NOT IDENTIFIED AS VARIABLE BY THE GAIA TEAM IN ORDER TO INCREASE THE RANGE OF APPLICATIONS FOR THE FINAL CATALOGUE. SOURCES SHOWING STRONG DEVIATIONS CAN BE REJECTED AS OUTLIERS BY THE USER, THROUGH SOME STATISTICAL ANALYSIS, BEFORE BEING USED AS CALIBRATORS. A SENTENCE ON THIS SENSE IS NOW ADDED TO THE TEXT.
Page 8, line 177: I guess that “accurate” should be more appropriate
than “precise” in this context.
MODIFIED
Fig. 9. In the Delta G vs. BP-RP plot, there is a curious “bridge” at
red colors (BP-RP>1.0). Is that due to some surface gravity effects in
that band, i.e., dwarfs vs. giants separation? If so, it would be nice
to mention that in the text.
YES. THE BRIDGE WITH LARGER RESIDUALS CORRESPOND TO MAIN SEQUENCE STARS AND GIANTS ARE SHOWING SMALLER RESIDUALS. THIS IS NOW INCLUDED IN THE TEXT.
Fig. 11. The red points are almost overlapped to the horizontal axis
over most of the plotting range. These plots would be much clearer
with a log scale on the vertical axis, but this is up the authors to
implement.
