A posteriori correction of Gaia CRF proper motions using conditional probability and zero-parallax prior
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.
Vera: I'm Vera.
Jocelyn: And I'm Jocelyn, here with Subrahmanyan.
Vera: We're looking at the paper: "A posteriori correction of Gaia CRF proper motions using conditional probability and zero-parallax prior".
Jocelyn: Here's the gist of the paper. Precision of proper motions and positions given in the Gaia Celestial Reference Frame (CRF) catalog can be statistically improved introducing conditional probability based on the prior knowledge of zero parallax for these extremely distant sources <ref:2608.16767#pg2>.
Vera: The method developed in this study produces statistically conditioned proper-motion estimates under the assumed per-source Gaussian covariance model <ref:2608.16767#pg3>.
Jocelyn: It is demonstrated that thus corrected proper motion vectors of 1.2 million CRF objects in Gaia DR3 show a tighter distribution around zero and a general reduction in the rate of outliers by approximately 15% <ref:2608.16767#pg4>.
Vera: A general vector spherical harmonic fit to degree 4 is produced with the corrected data for a filtered subset of CRF quasars with redshifts between 1 and 1.5 <ref:2608.16767#pg4>.
Paper summary: Vera: So we’re looking at this paper now, "A posteriori correction of Gaia CRF proper motions using conditional probability and zero-parallax prior." Essentially, they're taking the precision of those proper motions and positions from the Gaia Celestial Reference Frame catalog and trying to make it statistically better.
Jocelyn: They are introducing a conditional probability approach based on the prior knowledge that these extremely distant sources have zero parallax. That’s their main idea for improving things >
Subrahmanyan: But what does this actually mean for the measurements? Are we talking about a little tweak, or something bigger in terms of cosmic structure?
Vera: The paper claims that this method produces statistically conditioned proper motion estimates assuming a per-source Gaussian covariance model. They show that when you apply this correction to one point two million Gaia DR3 objects, the proper motion vectors get much tighter around zero and they see a general reduction in the rate of outliers by approximately fifteen percent <ref:2608.16767#pg4> > <ref:2608.16767#pg1,a general reduction in the rate of outliers by approximately 15>
Jocelyn: That reduction in outliers is interesting, but they also use this corrected data for a general vector spherical harmonic fit to degree four on a filtered subset of CRF quasars with redshifts between one and one point five <ref:2608.16767#pg4> > <ref:2608.16767#pg1,a filtered subset of CRF quasars with redshifts between 1 and 1>
Subrahmanyan: A vector spherical harmonic fit of degree four sounds like they are trying to model some kind of systematic distortion or perhaps an unknown component in the global parallax field itself, which is something we don't fully constrain yet <ref:2608.16767#pg4> > <ref:2608.16767#pg1>
Vera: And this application, predicated on that unknown stochastic component of the global parallax field, shows a three sigma change in the estimated magnitude of the Galactocentric acceleration of the solar system barycenter <ref:2608.16767#pg4> >
Jocelyn: A three sigma change in that acceleration estimate is significant because it quantifies how sensitive our fundamental determination is to this proposed conditioning of proper motions >
Subrahmanyan: So, if this correction helps us pin down the motion of our solar system relative to the galaxy, what does that tell us about the overall dynamics or perhaps large-scale structure formation?
Vera: The paper also mentions that they used a null test where parallaxes were randomly permuted to confirm that the improvement seen with true parallaxes isn't just an artifact of the algebraic procedure <ref:2608.16767#pg5> >
Jocelyn: And they did check how sensitive the low-degree VSH terms are to different parallax models, finding that those coefficients shift by one or two of their formal errors when using an SSH-based zero-point model <ref:2608.16767#pg9> >
Subrahmanyan: If the fit parameters are this sensitive to the assumed zero point, does that suggest we really have a strong prior on the global parallax field, or is there still a lot of uncertainty in that unknown component?
Vera: The sensitivity suggests that the sky distribution of proper motion updates mostly follows the peculiar distribution of astrometric correlations on the sky <ref:2608.16767#pg9> >
Jocelyn: And they also propagated instrumental bias through those conditional updates in Equation one to get their corrected proper motion field <ref:2608.16767#pg9> > <ref:2608.16767#pg1>
Subrahmanyan: So, looking at the bigger picture, what is the real significance here for cosmology? Is this just cleaning up the data or does it help constrain models of dark matter or dark energy?
Vera: The paper suggests that the prime purpose of Gaia proper motions is to determine and subtract the rigid spin of the entire observed sample, which they say is arbitrary in an unconstrained global solution <ref:2608.16767#pg5> >
Jocelyn: That means any inferred frame spin might be sensitive to how we observe quasar parallaxes and that sky-dependent parallax proper motion correlation pattern <ref:2608.16767#pg5> >
Subrahmanyan: So, what’s the final word on this work, "A posteriori correction of Gaia CRF proper motions using conditional probability and zero-parallax prior"? What is the main thing we should remember about it?
Vera: The main point is that you can probabilistically improve the precision of Gaia CRF proper motions by using that strong prior of zero parallax for those extremely distant sources <ref:2608.16767#pg5> >
Jocelyn: They quantify this improvement as a fifteen percent reduction in statistical outliers, for example <ref:2608.16767#pg4> > <ref:2608.16767#pg1>
Subrahmanyan: Even though most of the zero-parallax corrections are numerically small, some of those updates are comparable in absolute value to the formal errors they were already given for proper motions <ref:2608.16767#pg5> >
Vera: And they stress that no external astrometric information is needed, but everything hinges on that adopted zero-parallax prior and the assumed per-source Gaussian covariance model <ref:2608.16767#pg4> >
Jocelyn: So, it’s a conditional improvement where the result really depends on what you assume about those distant sources >
Conclusion: Vera: So we’re wrapping up this look at "A posteriori correction of Gaia CRF proper motions using conditional probability and zero-parallax prior." Basically, they took the raw proper motion data from Gaia and tried to make it more accurate by assuming those very distant sources have zero parallax.
Jocelyn: That's the core idea, right? They used that prior knowledge—zero parallax for distant things—to create a conditional probability model to improve the precision of those measurements. It’s about using what we *know* to fix what we *don't* know about the data.
Subrahmanyan: So if this works, it means we can get a tighter grip on the motion of these distant objects, which should help us map out how matter moves in the universe on large scales. It’s about getting a clearer picture of cosmic flow.
Vera: Exactly. They show that applying this correction to a million Gaia DR3 objects reduces statistical outliers by about fifteen percent, which is a solid number for data cleaning. But they also have to be careful; the correction depends entirely on assuming that per-source Gaussian covariance model holds true for those measurements.
Jocelyn: And they test that assumption by running a null test where they randomly permute the parallaxes, just to make sure the improvement isn't some algebraic trick, which is a good check. That kind of validation is important for building trust in the final numbers.
Subrahmanyan: The implication here is that this correction helps us quantify uncertainty in fundamental measurements, like how fast things are moving relative to each other in our galaxy. If these corrections are accurate, it impacts how we interpret the acceleration of the solar system barycenter.
Vera: That's right. The paper isn't just about cleaning up data; it’s about showing that even small priors can have measurable impacts on larger cosmological inferences, like that three sigma change they found in galactocentric acceleration.
Jocelyn: It really highlights how much the underlying assumptions—like what we assume about those distant sources—can shift our understanding of these motions. That opens up a new avenue for investigation into the structure of the local universe.
Subrahmanyan: So, while this method doesn't require any new external astrometric information, it proves that existing data can be refined using probabilistic methods to get tighter constraints on things like cosmic acceleration. This is exactly what we need to push our structural models further.
U.S. Naval Observatory
astro-ph.GA
Submitted: 2026-08-17
Updated: 2026-10-29
Comments: Published in PASP 138:104509, October 2026, DOI 10.1088/1538-3873/aea7da
Journal ref: 2026 PASP 138 104509
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 79/100
The gist: Precision of proper motions and positions given in the Gaia Celestial Reference Frame (CRF) catalog can be statistically improved introducing conditional probability based on the prior knowledge of…
Key concepts
- Conditional Probability Paradigm
- This statistical method uses prior knowledge—specifically assuming zero parallax for very distant objects—to adjust the measured proper motion. It mathematically reverses correlations based on this prior assumption to produce a statistically conditioned estimate of the true proper motion vector.
- Zero Parallax Prior
- The core assumption is that extremely distant sources have a zero parallax. This prior knowledge is used in the statistical model to correct the measured proper motions, effectively using this known condition to refine the estimates for these specific objects.
- Proper Motion Vector Correction
- The correction formula mathematically adjusts the measured proper motion ($μ$) by subtracting an expected value derived from parallax ($ψ$). This process is designed to reverse the sign of correlation coefficients, leading to a more accurate vector estimate for the objects.
- Galactocentric Acceleration
- This physical application uses the corrected proper motions to determine and subtract the rigid spin of observed samples. The study quantified how this correction impacts estimates of the solar system barycenter's galactocentric acceleration, showing sensitivity to parallax conditioning.
Terminology
Summary
Precision of proper motions and positions given in the Gaia Celestial Reference Frame (CRF) catalog can be statistically improved introducing conditional probability based on the prior knowledge of zero parallax for these extremely distant sources <ref:2608. p. 016767#pg2>. The method developed in this study produces statistically conditioned proper-motion estimates under the assumed per-source Gaussian covariance model <ref:2608. p. 016767#pg3>. It is demonstrated that thus corrected proper motion vectors of 1. p. 02 million CRF objects in Gaia DR3 show a tighter distribution around zero and a general reduction in the rate of outliers by approximately 15% <ref:2608. p. 016767#pg4>. A general vector spherical harmonic fit to degree 4 is produced with the corrected data for a filtered subset of CRF quasars with redshifts between 1 and 1. p. 05 <ref:2608. p. 016767#pg4>. This application, predicated on the unknown stochastic component of the global parallax field, shows a 3-sigma change in the estimated magnitude of the Galactocentric acceleration of the solar system barycenter, quantifying the sensitivity of this fundamental determination to the proposed conditioning of proper motions <ref:2608. p. 016767#pg4>.
Statistical Methodology
The core statistical improvement involves employing a conditional probability paradigm based on prior knowledge, specifically assuming zero parallax for extremely distant sources <ref:2608.16767#pg3>. The corrected estimate for the proper motion vector is given by the equation µˆ ≡ µ − E(µϖ) = µ − ϖ / σϖ [r01 σ1, r02 σ2] T, where µ is the measured proper motion value for a given object and ϖ is the measured parallax <ref:2608. p. 116767#pg3>. This correction reverses the sign of the correlation coefficients, which is intuitively clear from general geometric considerations <ref:2608. p. 116767#pg3>. The conditional dispersion of the proper motion vector is given by a formula that incorporates the parallax-corrected covariance matrix, which is positive semidefinite in the Loewner sense <ref:2608. p. 116767#pg3>.
Validation and Performance
The method was validated using Gaia data for 1.2 million CRF3 objects with known spectroscopic or Machine Learning-predicted redshifts <ref:2608. p. 216767#pg4>. The empirical histogram of the corrected proper motions is much closer to the expected χ(2) curve, and the rate of outliers in the tail is reduced by 15% when compared to the original distribution <ref:2608. p. 216767#pg4>. Specifically, for standardized magnitudes, the fraction of outliers above 3σ was reduced from 0. p. 30223 to 0.0190 <ref:2608. p. 016767#pg5>. A null test where parallaxes were randomly permuted confirmed that the improvement seen with true parallaxes is not an artifact of the algebraic procedure <ref:2608. p. 316767#pg5>.
Physical Interpretation and Application
The corrected proper motions are used to determine and subtract the rigid spin of the entire observed sample, which is arbitrary in an unconstrained global solution <ref:2608.16767#pg5>. The inferred frame spin may be sensitive to the observed distribution of quasar parallaxes and to the sky-dependent parallax–proper-motion correlation pattern <ref:2608. p. 6 p. 616767#pg5>. Furthermore, a pilot investigation showed that the corrected amplitude of the galactocentric dipole is 5. p. 459 ± 0.73 µas yr−1, which is closer to the Gaia-determined value representing a 3σ change <ref:2608. p. 516767#pg7>.
Sensitivity to Prior Models
The sensitivity of the low-degree VSH terms to the adopted parallax zero-point was demonstrated by comparing results using different parallax models <ref:2608.16767#pg5>. The low-degree coefficients shift by 1–2 of their formal errors when using an SSH-based zero-point model, indicating that the VSH fit is moderately sensitive to the choice of zero-point model <ref:2608. p. 616767#pg9>. This sensitivity suggests that the sky distribution of proper motion updates mostly follows the peculiar distribution of astrometric correlations on the sky <ref:2608. p. 616767#pg9>. The instrumental bias is then propagated via the conditional updates in Eq. p. 7 1 to the corrected proper motion field <ref:2608. p. 516767#pg9>.
Conclusion
Precision of Gaia CRF proper motions can be probabilistically improved using the strong prior of zero parallax for these extremely distant sources <ref:2608.16767#pg5>. The improvement can be quantified as a 15% reduction of statistical outliers, for example <ref:2608. p. 616767#pg4>. Although the majority of zero-parallax corrections is numerically quite small, some of the updates are comparable in absolute value to the given formal errors of proper motions <ref:2608. p. 616767#pg5>. No external astrometric information is required, although the result remains conditional on the adopted zero-parallax prior and the assumed per-source Gaussian covariance model <ref:2608. p. 616767#pg4>. The prime purpose of Gaia proper motions is to determine and subtract the rigid spin of the entire observed sample, which is arbitrary in an unconstrained global solution <ref:2608. p. 616767#pg5>. The inferred frame spin may be sensitive to the observed distribution of quasar parallaxes and to the sky-dependent parallax–proper-motion correlation pattern <ref:2608.16767#pg5>.
References
Butkevich, A. G., Klioner, S. p. 9 A., Lindegren, L. p. 1, Hobbs, D. p. 9, & van Leeuwen, F. p. 9 2017, A&A, 603, A45 <ref:2608. p. 916767#pg2>
Ding, Y., Liao, S., Wu, Q., Qi, Z., & Tang, Z. p. 9 2024, A&A, 691, A81 <ref:2608. p. 916767#pg4>
Gaia Collaboration et al. 2016, A&A, 595, A1 <ref:2608. p. 916767#pg5>
Gaia Collaboration et al., 2021, A&A, 649, A9 <ref:2608. p. 916767#pg5>
Gaia Collaboration et al., 2022, A&A, 667, A148 <ref:2608. p. 916767#pg5>
Holl, B., & Lindegren, L. p. 1 2012, A&A, 543, A14 <ref:2608. p. 916767#pg3>
Kopeikin, S. M., & Makarov, V. p. 0 V. 2006, AJ, 131, 1471 <ref:2608. p. 916767#pg6>
Kovalevsky, J. 2003, A&A, 404, 743 <ref:2608. p. 916767#pg6>
Lindegren et al., 2021, A&A, 649, A4 <ref:2608. p. 916767#pg6>
Makarov, V. 2025, Nature Astronomy, 9, 1396 <ref:2608. p. 916767#pg6>
Makarov & Berghea, C. T., 2026, Gaia parallax bias via spherical harmonics: A Python tool and discussion of possible causes <ref:2608.16767#pg6>
Makarov et al., 2012, AJ, 144, 22 <ref:2608. p. 916767#pg6>
Makarov & Secrest, N. J., 2023, AJ, 166, 8 <ref:2608. p. 916767#pg6>
Makarov & Secrest, N. J., 2022, ApJ, 933, 28 <ref:2608. p. 916767#pg6>
Mannucci et al., 2022, Nature Astronomy, 6, 1185 <ref:2608. p. 916767#pg6>
Orechovesky, J. R. 1996, Single Source Error Ellipse Combination,, Master Thesis <ref:2608. p. 9
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improve Gaia proper motion estimation by introducing conditional probability based on zero parallax prior, which produces
statistically conditioned proper-motion estimates under the assumed per-source Gaussian covariance model.
This allows AI systems to generate more robust and less biased proper motion vectors for distant sources in the Gaia catalog. -
Enhance the determination of local acceleration of free fall by utilizing corrected CRF proper motions to produce a
3-sigma change in the estimated magnitude of the Galactocentric acceleration of the solar system barycenter.
This provides a more accurate measure of secular aberration drift compared to methods relying solely on original data. -
Refine astrometric solutions for global catalog fitting by implementing corrections derived from Eq. 1, which calculates
the likeliest value of true proper motion µˆ given the measured parallax value ϖ is not equal to the measured µ but to a shifted value by the conditional mean.
This improves the precision of overall astrometric solutions in Gaia DR3. -
Reduce statistical outliers in proper motion distributions by applying correction methods, as demonstrated by a
general reduction in the rate of outliers by approximately 15%
for 1.2 million CRF objects. This leads to cleaner datasets where extreme measurements are less likely to skew subsequent analyses of galactic rotation or spin determination. -
Optimize VSH fitting for proper motion fields by using corrected vectors, as the updated fit shows that
the corrected amplitude of the galactocentric dipole is 5.59 plus 0.73 µas yr-1, which is closer to the Gaia-determined value representing a 3-sigma change.
This makes low-degree VSH estimates more reliable for determining large-scale proper motion patterns on the celestial sphere.
Sources
- Gaia parallax bias via spherical harmonics: A Python tool and discussion of possible causes
- Position drift with Gaia
- Optical-morphology-based assessment of astrometric quality in Gaia-CRF3 quasars
Related papers
- The Pristine survey XXVIII. Journey to the Galactic outskirts: Mapping the outer halo red giant stars down to the very metal-poor end
- Origins of Extreme Emission-Line Ratios in z > 3 Galaxies: Insights from the Lumen Model
- The Kinematically Hot, Extremely Metal-Poor C-19 Stellar Stream in DESI DR2
- UV slopes of Starforming Galaxies in Strong Lensing fields at the Epoch of Reionization with JWST
- Evolution of Binaries Under Stochastic Perturbations
- The hydrodynamical response of cold circumgalactic clouds to quasar radiation