REVIEW 3 major objections 4 minor 25 references
Digitization of Astronomical Photographic Plate of China: Photometric Calibration of Single-exposure Plates and Release of Photometric Catalogs
T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read The paper shows that about 15,680 century-spanning Chinese photographic plates can be calibrated to the modern JKC system at 0.11–0.23 mag precision, releasing 33 million calibrated stellar brightnesses.
desk verdict A valuable century-scale data release with an in-sample precision estimate that needs an out-of-sample check. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the per-plate zero-point model together with the standard-star sample it is fitted on. The standard stars come from the BEST database: roughly 200 million all-sky stars brighter than $G \approx 17.65$ whose JKC $UBVRI$ magnitudes are synthesized from Gaia DR3 BP/RP spectra whose systematic errors were corrected in Huang et al. (2024). Each plate is calibrated in four steps — construct the standard-star sample, select the closest JKC passband by the flattest residual slope versus $\mathrm{BP}-\mathrm{RP}$, remove false detections, and fit the zero-point — with the fitted terms iterated three times after a large-scale flat-field pre-correction. The model explicitly absorbs the three dominant error sources of photographic photometry: the S-shaped non-linearity of the emulsion (magnitude term), the mismatch between the plate response and the standard passband (color term), and position-dependent sensitivity variations (flat-field term).
What would settle it
Cross-check the catalog against stars with known large historical color changes: for long-period or semiregular variables with published color curves, the residual between calibrated and standard magnitude should not track the star's known color excursion at the epoch of each plate. A second, cheaper test is to plot residuals versus epoch for non-variable stars observed on many plates across the full 1901–1999 baseline — if the modern-color proxy degrades with time separation, scatter or offsets should visibly grow with baseline length, especially for red stars.
Extended reading notes
Core claim
The paper claims that a single per-plate zero-point model of the form $m = m_{\rm inst} + f_m(m_{\rm inst}) + f_c(\mathrm{BP}-\mathrm{RP}) + f_p(X,Y)$ — a fourth-order polynomial in instrumental magnitude, a second-order polynomial in Gaia $\mathrm{BP}-\mathrm{RP}$ color, and a flat-field term split into large-scale and small-scale components — is enough to calibrate digitized photographic plates onto the JKC system. For each plate, the closest JKC passband is chosen empirically from the slope of the color residuals rather than from plate or filter metadata, which is missing for 18% of plates (plate type) and 54% (filter). Calibration succeeds for 15,680 of 15,696 plates, reaching a typical precision of 0.11–0.23 mag depending on observatory, with the best plates at 0.05–0.11 mag. The authors further flag variable-source candidates by the deviation of calibrated magnitudes from standard magnitudes, and validate the results against the color–magnitude diagram of the globular cluster M4 and against Landolt standard stars.
Load-bearing premise
The whole calibration rests on using a star's present-day Gaia $\mathrm{BP}-\mathrm{RP}$ color as a proxy for its color at the moment the plate was exposed, decades earlier; a star that changed color substantially over that time would be fitted with the wrong color term and receive a systematically biased calibrated magnitude.
Editorial extensions
If this is right
- The released catalog of 33,282,558 unique sources with 229,271,500 independent observations gives a roughly 100-year photometric record, letting astronomers measure brightness changes that no single modern survey can cover.
- The empirical band-selection step recovers the correct photometric band even when plate type or filter is unrecorded, so the method does not require intact metadata.
- The calibrated data directly record historical outbursts — the paper demonstrates a 0.4 mag brightening of an M31 supergiant on 1985 plates — making the plates usable for event archaeology.
- Roughly 2.8 million 3-sigma variable candidates, with collected light curves, populate the expected regions of the HR diagram, including supergiants, white dwarfs, and pre-main-sequence stars.
- The per-plate model is generic and, the authors argue, can be applied to other digitized plate archives beyond the Chinese collection.
Reading between the lines
- The model should extend to the several thousand multiple-exposure plates already digitized but not calibrated here, roughly enlarging the usable archive without new scanning; the main obstacle would be the more complex point-spread structure of overlapping exposures.
- Merging these historical epochs with modern time-domain surveys, for example Gaia epoch photometry, could convert the single-epoch 3-sigma flags into confirmed long-term variables, since a genuine variable should scatter consistently across many plates rather than on one.
- The modern-color proxy could be hardened by iterative exclusion: fit each plate's color term only on stars flagged as photometrically stable in modern surveys, then re-run the variable search on the residuals; this would test whether the 2.8 million candidates are contaminated by color-changing stars.
- Because the typical 0.11–0.23 mag per-epoch precision is comparable to the amplitudes of classical variables such as RR Lyrae stars, the catalog is better suited to population-level variability statistics and rare outbursts than to detecting low-amplitude variations from single plates.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript presents a photometric calibration pipeline for 15,680 digitized single-exposure Chinese photographic plates, covering 1901-1999 from five observatories. For each plate, standard stars are drawn from the BEST database of synthetic JKC photometry based on corrected Gaia DR3 XP spectra, and the calibration models the magnitude response as a fourth-order polynomial in instrumental magnitude, a second-order color term in BP-RP, and large- and small-scale position-dependent flat-field terms (Eqs. 2-4). The closest JKC band is selected empirically per plate, false detections are removed, and the zero-point is iterated. The authors report typical per-plate precision of 0.11-0.23 mag, release catalogs containing 33,282,558 unique sources with 229,271,500 independent observations, and identify roughly 2.8 million 3-sigma variable candidates. External checks include an M4 color-magnitude diagram and a cross-match with Landolt standard stars.
Significance. If the precision claims hold, this is a valuable data release: it provides century-scale photometry from a large, previously underused archive, is publicly available at the China National Astronomical Data Center, and the per-plate calibration approach could be adapted to other digitized plate collections. The use of synthetic standards from corrected Gaia XP spectra is a modern and scalable strategy, and the paper is commendably explicit about the need to calibrate each plate independently. However, the central quantitative claim — the typical 0.11-0.23 mag precision — is currently supported only by in-sample residuals computed against the same standard stars used in the fit, and the external checks are qualitative or limited in scope. That issue must be addressed before the precision numbers and variable-candidate counts can be used reliably by the community.
major comments (3)
- [Section 4; Section 3.4, Eqs. (2)-(4)] The quoted per-plate precision (0.15/0.23/0.17/0.11/0.19 mag typical, with best cases in the Abstract and Section 4) is estimated "by comparing with the Gaia standard stars," but these are the same BEST/JKC standard stars used to fit the zero-point model in Section 3.4. The model is flexible: Eqs. (2)-(4) contain 18 global coefficients (five magnitude, three color, ten large-scale flat-field) plus a locally fitted first-order flat-field term per star, and Section 3.3 states that plates with as few as 10 standard stars are calibrated. In-sample residuals can therefore be systematically smaller than the error on independent objects, especially on low-N plates, so the reported values are calibration residuals rather than validated accuracies. The Landolt-star comparison in Section 4 gives sigma about 0.20 mag, but the text does not state that those Landolt stars were excluded from the BEST standards used in the fit, and it covers mainly NAOC and PMO plates. The M4 color-magnitude check in Figure 7 is qualitative and uses only two YNAO plates. I request an out-of-sample test, such as per-plate k-fold cross-validation or a reserved independent standard-star subset, with the resulting precision reported separately from the fitting residuals; the variable-candidate counts in Section 5 should then be re-derived using the out-of-sample scatter.
- [Section 3.3-3.4] The interaction between model complexity and sample size is not addressed. Plates with fewer than 30 standard stars skip the false-star removal step, and plates with 10-30 standard stars are calibrated with the same 18-parameter global model. The manuscript does not report how many plates fall in the 10-30 range, nor the precision distribution for that subsample. If in-sample precision is quoted globally, a small number of low-N plates with overfit solutions could pass the quality selection and contaminate both the aggregate precision statement and the variable-candidate statistics. Please report the N_std distribution and either reduce the model order for low-N plates or validate them with a holdout test.
- [Section 3.4, footnote] The assumption that a star's modern BP-RP color is a valid proxy for its historical color at the epoch of each plate is acknowledged in the footnote but not quantified. This assumption is not merely cosmetic for the variable-source component of the paper: a source whose color changed by a few tenths of a magnitude will have a color-term bias proportional to the color derivative of f_c, and if the color change is large the bias can be comparable to the quoted precision. The authors state that the impact is "not significant" because the color term is relatively flat, but no demonstration or tolerance analysis is given. I ask for a quantitative test, for example by simulating plausible color changes and propagating them through the fitted f_c, or by checking a sample of stars with known long-term color variations.
minor comments (4)
- [Throughout] There are several typographical and formatting errors: "U BV RI" in Section 2.2, "a approach" and "phometric" in the Conclusions, "V ariable" in the Section 5 heading, and inconsistent "YNO" versus "YNAO" labels in Figure 6. These should be corrected in the revised version.
- [Section 4, Figure 6] The term "plate grade" is used in the middle panels of Figure 6 and in the text ("plates of larger grade (lower quality)"), but the grading system is not defined. Please specify what the grades are and where they come from.
- [Table 2 and Section 2.1] The catalog columns are not fully consistent with the text: Table 2 lists "OUT_FLAG" and "XP_FLAG" while the text refers to "FLAGS" and "Syn_FLAG." Please unify the nomenclature and describe how FLAGS flags affect the photometric calibration.
- [Section 5] The variability candidate selection is described as comparing calibrated magnitudes with standard magnitudes in a sliding window, but no details are given about the window width (beyond 1 mag), the minimum number of stars per window, or how the 3-sigma threshold is computed for windows with few stars. Adding these details would improve reproducibility.
Circularity Check
Reported per-plate photometric precision is an in-sample residual computed from the same standard stars used to fit each plate's zero-point model, so the headline precision values are partially forced by the fit.
-
fitted input called prediction
[Section 3.4 (zero-point model, Eqs. 1-4) and Section 4 (precision estimate)]
"By comparing with the Gaia standard stars, the final photometric precision was estimated for each plate."
The 'Gaia standard stars' are the same BEST/JKC standards described in Section 2.2 and matched in Section 3.1 to construct the calibration sample used in Section 3.4 to fit the zero-point model (Eqs. 1-4: 4th-order magnitude term, 2nd-order color term, 10-parameter large-scale flat-field polynomial, and per-source 3-parameter local flat-field polynomial). The quoted per-plate precision is therefore the RMS of the residuals of this same fitted model on its training sample, not an independent out-of-sample accuracy measure. With as few as 10 standard stars per plate (Section 3.3) and highly flexible spatial terms, the in-sample RMS can be substantially smaller than true external error.
full rationale
The calibration pipeline is mostly self-contained: the standard-star reference frame is externally anchored to Landolt-standardized JKC photometry synthesized from corrected Gaia XP spectra, and the per-plate band selection, false-star rejection, and zero-point computation are data-driven. No load-bearing uniqueness theorem is imported from the authors' prior work, and the self-citations to the BEST/Xiao/Huang method papers provide reproducible methodology rather than circular justification. The one significant circularity is the central quantitative claim: the 'final photometric precision' of 0.15/0.23/0.17/0.11/0.19 mag is computed by comparing calibrated magnitudes with the very same BEST standard stars used to fit each plate's zero-point model. Because the model is flexible (fourth-order magnitude, second-order color, 10-parameter flat field, plus a per-source local polynomial), training-set residuals can be substantially smaller than out-of-sample errors, especially on plates with only 10-30 standards. The external checks (Landolt sigma about 0.20 mag; M4 CMD morphology) provide partial independent support but are not per-plate quantitative confirmations of the claimed best-case values. This makes the precision estimate partially forced by construction, yielding a score of 6.
Assumptions & free parameters
free parameters (7)
- Magnitude term coefficients a_0 to a_4 =
per-plate, not tabulated
- Color term coefficients b_0 to b_2 =
per-plate, not tabulated
- Large-scale flat-field coefficients p_ij =
per-plate, not tabulated
- Small-scale flat-field coefficients q_0 to q_2 =
per-plate, not tabulated
- Neighborhood radius a and threshold N_TS_std =
per-plate, not tabulated
- std mag cut and inst mag cut =
per-plate
- Bin width and percentage thresholds for cut determination =
empirically varied per plate
assumptions (4)
- domain assumption Gaia DR3 XP spectra, after the corrections of Huang et al. (2024), can accurately synthesize JKC magnitudes for standard stars.
- domain assumption A star's BP-RP color is effectively constant between the plate epoch (1901-1999) and the Gaia epoch.
- ad hoc to paper The plate response can be described by a fourth-order polynomial in magnitude, a second-order polynomial in color, and a third-order 2D polynomial in position.
- standard math Cross-matching standard stars with a 1 arcsec radius yields correctly matched counterparts.
Cite this review
Pith. "Pith review of Digitization of Astronomical Photographic Plate of China: Photometric Calibration of Single-exposure Plates and Release of Photometric Catalogs." pith.science (2026). https://pith.science/paper/CNNZNWZK
@misc{pith2026250710270,
author = {Pith},
title = {Pith review of: Digitization of Astronomical Photographic Plate of China: Photometric Calibration of Single-exposure Plates and Release of Photometric Catalogs},
year = {2026},
howpublished = {\url{https://pith.science/paper/CNNZNWZK}},
note = {Machine review of arXiv:2507.10270}
}
read the original abstract
The Chinese Plate-Digitizing Project has digitized a total number of about 30,000 astronomical plates from 11 telescopes of five observatories (SHAO, NAOC, PMO, YNAO, and QDO) in China, spanning nearly 100 years of observations. In this work, we present a photometric calibration method to calibrate about 15,000 single-exposure plates to the JKC photometric system. Using standard stars constructed from the BEST database, we have identified and corrected various systematic effects (including the magnitude term, color term, and flat-field term) to a high precision. The final photometric precision is typically 0.15, 0.23, 0.17, 0.11, 0.19 mag for plates collected in SHAO, NAOC, PMO, YNAO, and QDO, respectively, with best cases reaching 0.07, 0.08, 0.06, 0.05, and 0.11 mag, respectively. Candidates of variable sources are also identified. The catalogs are publicly available at the China National Astronomical Data Center. Such a dataset will provide a valuable resource for conducting long-term, temporal-scale astronomical research. Our calibration method can also be applied to other digitized astronomical plates in general.
Figures
Figures from the paper (10 more)
Reference graph
Works this paper leans on
-
[1]
Angione, R. J. & Smith, H. J. 1985, AJ, 90,
work page 1985
-
[13]
2018, Astronomische Nachrichten, Astrophysics with digitized astronomical plate archives, 339, 5,
doi:10.3847/1538-4365/ad18b1 Hudec, R. 2018, Astronomische Nachrichten, Astrophysics with digitized astronomical plate archives, 339, 5,
-
[14]
Figure 11.Same to Figure 4 but for the plate BJ8404DA3924001
doi:10.3847/1538-4365/ac9dfa 13 APPENDIX Figure 10.Same to Figure 4 but for the plate SH6401V64087001. Figure 11.Same to Figure 4 but for the plate BJ8404DA3924001. 14Ma et al. Figure 12.Same to Figure 4 but for the plate ZT8409SN3433001. Figure 13.Same to Figure 4 but for the plate YN8105175001
-
[26]
doi:10.3847/1538-4365/adaf27 Montegriffo, P., De Angeli, F., Andrae, R., et al. 2023, A&A, 674, A3. doi:10.1051/0004-6361/202243880 Perlbarg, A.-C., Desmars, J., Robert, V., et al. 2023, A&A, 680, A41. doi:10.1051/0004-6361/202347100 Robert, V., Desmars, J., Lainey, V., et al. 2021, A&A, The NAROO digitization center. Overview and scientific program, 652,...
-
[28]
doi:10.1088/1674-4527/17/3/28 Zhang, R. & Yuan, H. 2023, ApJS, 264,
-
[41]
doi:10.3847/1538-4365/ad24fa Xiao, K., Yuan, H., Huang, B., et al. 2023, arXiv:2307.13238. doi:10.48550/arXiv.2307.13238 Xiao, K., Yuan, H., Huang, B., et al. 2023, ApJS, 268,
work page Pith review arXiv doi:10.48550/arxiv.2307.13238 2023
-
[44]
doi:10.3847/1538-4365/ac3df6 Yacob, A. M., Berdnikov, L. N., Pastukhova, E. N., et al. 2022, MNRAS, 516,
-
[53]
doi:10.3847/1538-4365/acee73 Xiao, K., Yuan, H., L´ opez-Sanjuan, C., et al. 2023, ApJS, 269,
Show all 25 references
-
[58]
Xu, S., Yuan, H., Niu, Z., et al
doi:10.3847/1538-4365/ad0645 12Ma et al. Xu, S., Yuan, H., Niu, Z., et al. 2022, ApJS, 258,
2022 doi
- [69]
-
[88]
2003, Astronomical Data Analysis Software and Systems XII, Digital Access to Aero- and Astrophotographic Archives, 295,
doi:10.1088/0004-6256/146/4/88 de Cuyper, J.-P., van Dessel, E., Pauwels, T., et al. 2003, Astronomical Data Analysis Software and Systems XII, Digital Access to Aero- and Astrophotographic Archives, 295,
2003 doi
-
[93]
J., De Cuyper, J.-P., et al
Do, A., Shappee, B. J., De Cuyper, J.-P., et al. 2021, MNRAS, 508,
2021
-
[101]
C., MacGillivray, H
Hambly, N. C., MacGillivray, H. T., Read, M. A., et al. 2001, MNRAS, The SuperCOSMOS Sky Survey - I. Introduction and description, 326, 4,
2001
-
[140]
doi:10.1086/664083 Clem, J. L. & Landolt, A. U. 2013, AJ, Faint UBVRI Standard Star Fields, 146, 4,
2013 doi
- [393]
-
[408]
1990, A&A, 235, 174 Hudec, R
doi:10.1002/asna.201813515 Hudec, R., Peresty, R., & Motch, C. 1990, A&A, 235, 174 Hudec, R. 2019, Astronomische Nachrichten, Astronomical photographic data archives: Recent status, 340, 7,
1990 doi
-
[690]
2007, Progress in Astronomy, 25, 1 Lang, D., Hogg, D
doi:10.1002/asna.201913676 Jin, W.-J., Tang, Z.-H., Wang, S.-H., et al. 2007, Progress in Astronomy, 25, 1 Lang, D., Hogg, D. W., Mierle, K., et al. 2010, AJ, 139,
2007 doi
-
[889]
2024, ApJS, 271,
doi:10.1086/167059 Huang, B., Yuan, H., Xiang, M., et al. 2024, ApJS, 271,
2024 doi
-
[1279]
W., Hurley, K., et al
doi:10.1111/j.1365-2966.2001.04660.x Hartmann, D., Pogge, R. W., Hurley, K., et al. 1989, ApJ, 336,
2001
-
[1782]
2025, ApJS, The S-PLUS Ultra-short Survey: Photometric Recalibration with the Best Star Database, 277, 1,
doi:10.1088/0004-6256/139/5/1782 Li, X., Xiao, K., Huang, Y., et al. 2025, ApJS, The S-PLUS Ultra-short Survey: Photometric Recalibration with the Best Star Database, 277, 1,
2025 doi
- [1914]
-
[2095]
C., Dourneau, G., et al
doi:10.1093/mnras/stac2167 Yan, D., Qiao, R. C., Dourneau, G., et al. 2016, MNRAS, 457,
2016 doi
- [2474]
-
[2900]
2017, Research in Astronomy and Astrophysics, 17,
doi:10.1093/mnras/stw120 Yu, Y., Zhao, J.-H., Tang, Z.-H., et al. 2017, Research in Astronomy and Astrophysics, 17,
2017 doi
-
[3649]
2024, A&A, 687, A165
doi:10.1093/mnras/stab2660 Enke, H., Tuvikene, T., Groote, D., et al. 2024, A&A, 687, A165. doi:10.1051/0004-6361/202348793 Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1. doi:10.1051/0004-6361/201833051 Gaia Collaboration, Montegriffo, P., Bella...
2024 doi
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.