Pith. sign in

REVIEW 2 major objections 9 minor 20 references

Astrometric Calibration of the 4-m International Liquid Mirror Telescope Observations

T0 review · 2 major / 9 minor · reviewed 2026-07-12 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A TDI-specific astrometric pipeline delivers ~100 mas precision (70–80 mas for moderate stars) on four years of ILMT zenith scans, tied to Gaia and stable enough for public release.

desk verdict Solid, multi-year-validated TDI astrometry pipeline for ILMT that delivers ~100 mas residuals against Gaia and public data products; the modeling choice is pragmatic and empirically supported, not a load-bearing flaw. read the letter →

arxiv 2607.02662 v1 pith:3CLS7SJT submitted 2026-07-02 astro-ph.IM

Vibhore Negi , Naveen Dukiya , Gokul Singh Mehra , Bhavya Ailawadhi , Monalisa Dubey , Sara Filali , Paul Hickson , Brajesh Kumar
show 7 more authors
Priyanshi Kumari Sapna Mishra Kuntal Misra Bikram Pradhan Kumar Pranshu Jean Surdej Sarvesh Kumar Yadav
This is my paper
classification astro-ph.IM
keywords zenithtelescopesastrometrysurveysliquidmirrortelescopetimedelayintegrationGaiaDR3WCScalibration
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper presents an instrument-specific pipeline that converts pixel coordinates of the 4-m International Liquid Mirror Telescope’s long, drifting TDI frames into sky coordinates. Because a liquid-mirror telescope cannot track, stars drift continuously across a narrow strip; residual optical distortions remain even after a TDI corrector. The authors therefore fit a simple linear map plus a single quadratic term in the cross-scan direction, using only two small end-chunks of each frame and tying the solution to Gaia DR3. Applied to 347 nights spanning four observing cycles, the method yields typical positional residuals of about 100 milliarcseconds, improving to 70–80 mas for moderately bright stars, with no large systematic trends and little dependence on filter. The calibrated images have been released, establishing a practical foundation for nightly monitoring of moving and variable objects.

What carries the argument

The seven-parameter transformation (linear terms in both axes plus one quadratic term only in the DEC/cross-scan direction) that maps detector pixels to epoch-corrected Gaia coordinates and is encoded as a FITS WCS header with SIP distortion coefficients.

What would settle it

Measure residuals of stars lying strictly between the two calibration chunks against Gaia; if those mid-frame residuals systematically exceed the quoted 100 mas or show coherent curvature, the global-model claim fails.

Watch

Extended reading notes

Core claim

A linear-plus-quadratic transformation fitted to two end-chunks of each ILMT TDI frame, referenced to Gaia DR3, delivers a typical full-frame astrometric precision of ~100 mas (reaching 70–80 mas near G = 16.5–18.5) that remains stable across 347 nights and three optical bands.

Load-bearing premise

That a single global model fitted only on stars inside two small end-chunks fully describes residual distortions across the entire long TDI frame.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 9 minor

Summary. The manuscript presents an astrometric calibration pipeline for 4-m ILMT TDI images. It obtains an initial plate solution on two 2048×2048 end-chunks with astrometry.net, cross-matches to a low-PM/low-parallax Gaia DR3 calibrator catalog, propagates coordinates to the observation epoch, and fits a seven-parameter near-linear model (Eqs. 1–2) with a single quadratic term in the DEC (cross-scan) direction. The solution is encoded as a FITS WCS with CAR projection plus second-order SIP terms. Applied to ~347 nights from cycles 1–4 (2022–2025), the pipeline yields typical residuals of ~100 mas versus Gaia DR3 (~70–80 mas at G~16.5–18.5), with band-independent performance, moderate seeing dependence, multi-year stability, and a factor ~3–4 improvement over full-frame astrometry.net. Calibrated products are publicly released.

Significance. ILMT is a distinctive zenith-pointing liquid-mirror TDI facility; a documented, instrument-specific astrometric solution tied to Gaia and validated over hundreds of nights is a necessary foundation for its time-domain science (transients, variability, Solar System and Galactic monitoring). Strengths include: (i) a physically motivated low-order model matched to TDI-corrector residuals rather than a generic plate solve; (ii) quantitative residual statistics (MAD, Rayleigh) versus magnitude, filter, and seeing; (iii) direct spatial residual comparison with astrometry.net showing removal of large-scale curvature; (iv) multi-year coverage and public data release. These make the work a useful methods and data paper for the survey community and a template for other long-strip TDI systems.

major comments (2)
  1. Section 4 (epoch conversion): Gaia coordinates are said to be converted to the observation epoch with precession, nutation, and aberration only. Proper motions are not mentioned, despite a Gaia DR3 reference epoch near J2016.0 and a calibrator cut µ < 20 mas/yr. Over ~6–9 years that cut still allows up to ~120–180 mas of uncorrected motion—comparable to or larger than the quoted 70–100 mas residual floor. Please state explicitly whether µα, µδ (and optionally parallax) are applied when building the fit and when evaluating residuals; if not, quantify the residual PM contribution (e.g., with a tighter µ subsample or by re-fitting with PM applied) so that the claimed full-frame precision is not partly an epoch-mismatch scatter term.
  2. Section 4, Eqs. (1)–(2) and the two-chunk procedure: The global 7-parameter model is fitted only to calibrators in the two end 2048×2048 chunks, then applied across the full 4096×36864 frame. Full-frame residual histograms (Fig. 3) and the example residual maps (Fig. 7 top) already argue against large mid-scan systematics, but a short quantitative check would make the central claim more robust: e.g., residual RMS (or MAD) in RA/DEC as a function of along-scan pixel for a representative set of frames, or a three-chunk test (beginning/middle/end). This need not change the model; it would document that the quoted ~100 mas is not optimistic relative to the fit regions.
minor comments (9)
  1. Section 3.1: parallax is written as “δ < 10 mas,” which collides with the usual DEC symbol. Prefer ϖ or π.
  2. Section 5.2 heading: “corelation” → “correlation.”
  3. Section 2: server names “ic1” vs “icc1” appear inconsistently; clarify if these are distinct machines.
  4. Title line in the draft has a spurious space (“T elescope”); fix in production.
  5. WCS header excerpt (Section 4.1): RADESYS='FK5' with EQUINOX set to the observation epoch, while Gaia is ICRS. A one-sentence note on the intended celestial frame (and any FK5/ICRS distinction at the ~100 mas level) would help downstream users.
  6. Figure 4 uses a single night (2024-02-07, i-band). A sentence stating that other nights/bands behave similarly (or a brief multi-night check) would strengthen the magnitude-dependence claim.
  7. Section 4: report typical numbers of accepted Gaia calibrators per frame (and after the 2-arcsec companion cut) so readers can judge the leverage of the 7-parameter fit.
  8. Figure 3 caption: the Rayleigh scale relation is fine; also state whether the plotted sample is all detections or restricted in magnitude/SNR, for consistency with Fig. 4 and Fig. 8.
  9. Abstract/Introduction: “~120 sq. degrees” vs later “115 sq. deg” (Section 3)—align the survey-area figure.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: empirical pipeline residuals against Gaia are ordinary external accuracy, not a tautology or forced prediction.

full rationale

The paper constructs a 7-parameter pixel-to-sky transformation (linear terms plus one quadratic DEC term, Eqs. 1–2) by least-squares fit to Gaia DR3 calibrators extracted from two end-chunks of each TDI frame, then reports the resulting full-frame positional residuals against the same Gaia catalog (median ~0.12 arcsec RA / ~0.11 arcsec DEC; ~70–80 mas at G~16.5–18.5). This is the standard definition of external astrometric accuracy for a calibrated survey pipeline; the residuals are not a prediction independent of the fit, nor are they forced by construction beyond ordinary least-squares. An independent baseline (full-frame astrometry.net) is shown to leave larger, spatially coherent residuals (Figs. 7–8), confirming that the improvement is empirical rather than definitional. Self-citations (to prior ILMT/TDI-corrector papers) supply instrument context and justify the near-linear model form but are not load-bearing uniqueness theorems or ansatzes that close a logical loop. No equation reduces to its own input; the multi-year stability claim is simply the same residual statistic aggregated over 347 nights. The derivation chain is therefore self-contained against an external reference frame.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central performance claim is an empirical measurement against an external catalog. It rests on standard domain assumptions (Gaia as reference, optical behavior of the TDI corrector) plus a small set of modeling choices (polynomial order, chunk geometry, matching radius) that are fitted or fixed per frame. No new physical entities are postulated.

free parameters (3)
  • per-frame transformation coefficients f1,f2,f3,g1,g2,g3,g4
    Seven free parameters of Eqs. 1–2 are solved by non-linear least squares (LMFIT) for every TDI frame; the quoted residual statistics are the post-fit scatter of that model.
  • quadratic polynomial order (A_ORDER=B_ORDER=2) and zero higher SIP terms
    The decision to stop at second order and to set all other SIP coefficients to zero is a modeling choice that directly determines residual floor; it is not derived from first principles in the paper.
  • cross-match radius 2 arcsec and companion rejection radius 2 arcsec
    Fixed thresholds used to build the calibrator sample; they control which Gaia stars enter the fit and therefore the reported precision.
assumptions (4)
  • domain assumption Gaia DR3 positions (after µ<20 mas/yr, ϖ<10 mas, D<2 cuts) constitute an unbiased external reference free of systematics at the ~10–20 mas level relevant to ILMT.
    Invoked throughout Sections 3.1 and 4 as the sole absolute reference; the entire residual analysis is defined relative to this catalog.
  • domain assumption After the TDI optical corrector, stellar trajectories are nearly linear, so residual field distortion is adequately described by a single quadratic term in the DEC (cross-scan) direction.
    Stated in Section 4 and used to justify the functional form of Eqs. 1–2; supported by prior ILMT optical papers but not re-derived here.
  • standard math A cylindrical (CAR) projection plus SIP polynomials can encode the pixel-to-sky map without introducing additional rotation or higher-order terms beyond those fitted.
    Section 4.1 maps the fitted coefficients onto FITS WCS keywords under this assumption.
  • ad hoc to paper Two 2048×2048 end-chunks supply a calibrator sample that is representative of the full 3.3° scan length.
    Methodological choice in Section 4; if the distortion field varies non-monotonically along the scan, the global fit under-corrects the middle of the frame.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Astrometric Calibration of the 4-m International Liquid Mirror Telescope Observations." pith.science (2026). https://pith.science/paper/3CLS7SJT

@misc{pith2026260702662,
  author       = {Pith},
  title        = {Pith review of: Astrometric Calibration of the 4-m International Liquid Mirror Telescope Observations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3CLS7SJT}},
  note         = {Machine review of arXiv:2607.02662}
}
read the original abstract

The 4-m International Liquid Mirror Telescope (ILMT) is a dedicated time domain survey telescope that continuously scans the zenithal sky over the Indian Himalayas in the g', r' and i' optical bands. Its unique capability to repeatedly image the same strip of sky every night makes it a highly useful instrument for the photometric and astrometric studies of Solar System, Galactic and extragalactic objects. We present a robust astrometric calibration pipeline developed for the ILMT data obtained in the time delay integration (TDI) mode. The pipeline uses a linear transformation model from pixel to world coordinates, with a second order correction for the asymmetric optical distortions introduced by the telescope's optical corrector, and ties the astrometric solution to the Gaia DR3 reference frame. The pipeline is integrated to the routine ILMT data processing workflow. Using data from the first four observing cycles (2022-2025), we present the first assessment of the astrometric performance of the pipeline based on positional residuals of sources cross-matched with Gaia DR3. The pipeline achieves a typical astrometric precision of ~100 milliarcseconds (mas), reaching ~70-80 mas for moderately bright sources (G~16.5-18.5). These results, based on 347 nights of data, demonstrate the stability and reliability of ILMT astrometry over multi-year timescales. The astrometrically calibrated data from these four observing cycles have been made publicly available to the astronomical community. This work establishes a validated framework for precision astrometry with zenith-pointing TDI surveys and provides a foundation for future time-domain studies with ILMT, including variability characterization, transient localization, and long-term positional monitoring.

Figures

Figures reproduced from arXiv: 2607.02662 by the authors.

Figure 1
Figure 1. An example raw (left) and clean (right) TDI image obtained with the ILMT on 2025-02-02. The first 4096 pixels along the RA direction (Y pixels) in the raw image, are not fully exposed due to TDI ramping, and hence are removed while cleaning. g ′ , r ′ , and i ′ bands, respectively, with an average of ∼20 TDI exposures collected each night. For this work, we have used the commissioning phase data from the ILMT to dev… view at source ↗
Figure 2
Figure 2. Sky distribution (in J2000) of all the ILMT detections in cycles 1-4. The apparent sinusoidal footprint is a consequence of the observation epoch to J2000 precession transformation, while the strip width is determined by the 22.3′ ILMT field of view. The map is generated using hexagonal bins with a gridsize of 360, and the color scale shows log of the number of sources per bin. High-density vertical structures trace… view at source ↗
Figure 3
Figure 3. Astrometric offsets from our custom astrometry pipeline for the data obtained with the ILMT during the cycles 1-4, compared with Gaia positions. The left plot shows the distribution of the offsets in the RA, whereas the center plot shows the offsets in DEC. The right most subplot shows the distribution in the total angular separation between our astrometric coordinates and the Gaia coordinates. The Rayleigh scale pa… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Astrometric scatter (σr) as a function of Gaia G magnitude for our fine astrometric solution. The scatter in each magnitude bin is estimated using a robust MAD-based statistic and is computed for all the stars detected with ILMT on 2024-02-07, in 26 individual frames (…
Figure 5
Figure 5. Figure 5: The three subplots show the distribution in the total angular separation between our astrometric coordinates and the Gaia coordinates, for the three ILMT bands g ′ , r ′ and i ′ , respectively. 1 2 3 4 5 FWHM (arcsec) 0.1 0.2 0.3 0.4 0.5 0.6 M e dia n total offs et r (…
Figure 6
Figure 6. Figure 6: Median astrometric offset as a function of at￾mospheric seeing, for the full cycle-3 and cycle-4 data. The median total astrometric residual is shown as a function of the image FWHM. Blue circles show the median total resid￾ual per FWHM bin, computed from the median of…
Figure 7
Figure 7. Figure 7: Astrometric offsets in one of the TDI frames observed with ILMT on 2024-02-07. The top panel shows the offsets in RA (/DEC) as a function of pixels along RA (/DEC), from our fine astrometry. The pixels along DEC (/RA) are color coded to show variation along the DEC (/R…
Figure 8
Figure 8. Figure 8: Two-dimensional residual distributions of astrometric offsets with respect to Gaia DR3 for all the objects (17 < imag < 19) detected with ILMT on 2024-02-07, in 26 individual full TDI frames, observed in i−band. The left panel shows the initial astrometric solution fro…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

20 extracted references · 12 canonical work pages

  1. [1]

    2024, Bulletin de la Societe Royale des Sciences de Liege, 93, 820, doi: 10.25518/0037-9565.11892 Astropy Collaboration, Robitaille, T

    Ailawadhi, B., Akhunov, T., Borra, E., et al. 2024, Bulletin de la Societe Royale des Sciences de Liege, 93, 820, doi: 10.25518/0037-9565.11892 Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558, A33, doi: 10.1051/0004-6361/201322068

    work page doi:10.25518/0037-9565.11892 2024

  2. [2]

    R., & Greisen, E

    Calabretta, M. R., & Greisen, E. W. 2002, A&A, 395, 1077, doi: 10.1051/0004-6361:20021327 11

    doi:10.1051/0004-6361:20021327 2002

  3. [3]

    C., Magnier, E

    Chambers, K. C., Magnier, E. A., Metcalfe, N., et al. 2016, arXiv e-prints, arXiv:1612.05560, doi: 10.48550/arXiv.1612.05560

    arXiv doi:10.48550/arxiv.1612.05560 2016

  4. [4]

    2022, Journal of Astronomical Instrumentation, 11, 2240001, doi: 10.1142/S2251171722400013 Gaia Collaboration, Prusti, T., de Bruijne, J

    Dukiya, N., Misra, K., Pradhan, B., et al. 2022, Journal of Astronomical Instrumentation, 11, 2240001, doi: 10.1142/S2251171722400013 Gaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al. 2016, A&A, 595, A1, doi: 10.1051/0004-6361/201629272 Gaia Collaboration, Vallenari, A., Brown, A. G. A., et al. 2023, A&A, 674, A1, doi: 10.1051/0004-6361/202243940

    work page doi:10.1142/s2251171722400013 2022

  5. [5]

    W., & Calabretta, M

    Greisen, E. W., & Calabretta, M. R. 2002, A&A, 395, 1061, doi: 10.1051/0004-6361:20021326

    doi:10.1051/0004-6361:20021326 2002

  6. [6]

    Hall, P., & Mackay, C. D. 1984, MNRAS, 210, 979, doi: 10.1093/mnras/210.4.979

    work page doi:10.1093/mnras/210.4.979 1984

  7. [7]

    2019, in LPI Contributions, Vol

    Hickson, P. 2019, in LPI Contributions, Vol. 2109, First International Orbital Debris Conference, 6066

    2019

  8. [8]

    Hickson, P., & Richardson, E. H. 1998, PASP, 110, 1081, doi: 10.1086/316230

    work page doi:10.1086/316230 1998

Show all 20 references
  1. [9]

    2022, Journal of Astrophysics and Astronomy, 43, 10, doi: 10.1007/s12036-021-09795-3

    Kumar, B., Negi, V., Ailawadhi, B., et al. 2022, Journal of Astrophysics and Astronomy, 43, 10, doi: 10.1007/s12036-021-09795-3

    2022 doi

  2. [10]

    W., Mierle, K., Blanton, M., & Roweis, S

    Lang, D., Hogg, D. W., Mierle, K., Blanton, M., & Roweis, S. 2010, AJ, 139, 1782, doi: 10.1088/0004-6256/139/5/1782

    2010 doi

  3. [11]

    T., Angel, J

    McGraw, J. T., Angel, J. R. P., & Sargent, T. A. 1980, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 264, Conference on Applications of Digital Image Processing to Astronomy, ed. D. A. Elliott, 20–28, doi: 10.1117/12.959777

    1980 doi

  4. [12]

    2024, Bulletin de la Societe Royale des Sciences de Liege, 93, 872, doi: 10.25518/0037-9565.11908

    Misra, K., Ailawadhi, B., Akhunov, K., et al. 2024, Bulletin de la Societe Royale des Sciences de Liege, 93, 872, doi: 10.25518/0037-9565.11908

    2024 doi

  5. [13]

    2024, Bulletin de la Societe Royale des Sciences de Liege, 93, 863, doi: 10.25518/0037-9565.11904

    Negi, V., Ailawadhi, B., Akhunov, T., et al. 2024, Bulletin de la Societe Royale des Sciences de Liege, 93, 863, doi: 10.25518/0037-9565.11904

    2024 doi

  6. [14]

    B., et al

    Newville, M., Stensitzki, T., Allen, D. B., et al. 2016, Lmfit: Non-Linear Least-Square Minimization and Curve-Fitting for Python,, Astrophysics Source Code Library, record ascl:1606.014 http://ascl.net/1606.014

    2016

  7. [15]

    2025, MNRAS, 538, 133, doi: 10.1093/mnras/staf206

    Pranshu, K., Misra, K., Ailawadhi, B., et al. 2025, MNRAS, 538, 133, doi: 10.1093/mnras/staf206

    2025 doi

  8. [16]

    L., Moshir, M., Li, J., et al

    Shupe, D. L., Moshir, M., Li, J., et al. 2005, in Astronomical Society of the Pacific Conference Series, Vol. 347, Astronomical Data Analysis Software and Systems XIV, ed. P. Shopbell, M. Britton, & R. Ebert, 491

    2005

  9. [17]

    2018, Bulletin de la Societe Royale des Sciences de Liege, 87, 68

    Surdej, J., Hickson, P., Borra, H., et al. 2018, Bulletin de la Societe Royale des Sciences de Liege, 87, 68

    2018

  10. [18]

    2024, Bulletin de la Societe Royale des Sciences de Liege, 93, 1054, doi: 10.25518/0037-9565.11961

    Surdej, J., Ailawadhi, B., Akhunov, T., et al. 2024, Bulletin de la Societe Royale des Sciences de Liege, 93, 1054, doi: 10.25518/0037-9565.11961

    2024 doi

  11. [19]

    2025, A&A, 694, A80, doi: 10.1051/0004-6361/202452667

    Surdej, J., Hickson, P., Misra, K., et al. 2025, A&A, 694, A80, doi: 10.1051/0004-6361/202452667

    2025 doi

  12. [20]

    2002, A&A, 388, 712, doi: 10.1051/0004-6361:20020516

    Vangeyte, B., Manfroid, J., & Surdej, J. 2002, A&A, 388, 712, doi: 10.1051/0004-6361:20020516

    2002 doi

Pith tools

Reviewed July 12, 2026 · model on record in the stance chip above.