{"id":"6878c685-efd9-440b-8d47-035e35b7b412","arxiv_id":"2411.16951","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Using 40-year VLBI time series, the authors find that the preferred direction of astrometric position scatter in many ICRF3 quasars aligns within 30 degrees of their radio jet position angle.","lead":"Astronomers measured how the apparent positions of 265 bright quasars wander over 40 years of VLBI observations and found that many wander preferentially along the same direction as their radio jets. If confirmed, this ties quasar position noise to jet physics and could improve how the celestial reference frame is built.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unmodeled VLBI source structure may create jet-aligned astrometric excursions, so the reported alignment may be partly a technical artifact rather than intrinsic core shifts.","rationale":"The reader's conditionality is well placed, and the isotropy concern is real. I would sharpen it: the concrete mechanism that can manufacture the headline alignment is VLBI source structure, whose position-angle signature is aligned with the very jet directions used in the comparison. Since C_xy only contains formal single-session covariances, structure-induced delay errors remain in the standardized residuals and are exactly what PCA will extract. This makes the Section 5 comparison potentially circular at the systematic level, independent of the missing significance test. The paper's internal examples illustrate the ambiguity: B1038+52A is an elongated/dual source, and 0607-157 aligns with the inner 8 GHz jet but not the extended jet, which could be either a physical core-shift signature or a scale-dependent structure artifact. The method is competently described and reproducible, and the bootstrap uncertainties are useful, so the result is best treated as conditional pending a structure-sensitive control. I therefore do not move the reader's verdict; the concern strengthens the conditionality rather than overturning it.","tokens_in":9399,"tokens_out":7829,"duration_ms":87825,"concrete_test":"Apply the same PCA and alignment comparison to a subset of the 265 sources with compact ICRF3 structure indices (e.g., structure index 0 or no resolved mas-scale jet components in the Plavin catalog beyond the core), and compare the alignment fraction with the full 127-source sample. If the ±30° alignment excess persists for compact, unresolved sources, the source-structure artifact is not the main driver; if it drops toward the ~31% triangular expectation, the reported 'nearly one-half' alignment is contaminated by structure-induced astrometric noise. A complementary check: recompute ϑast from a global VLBI solution that applies source-structure corrections to the group delays and see whether the histogram peak near 0 survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3, Eq. 1 whitens residuals using only the per-session formal covariance C_xy. That covariance does not include VLBI source-structure errors: for a resolved core-jet or dual-component source, the measured group delay and apparent position depend on array geometry and the source brightness distribution. These structure-induced offsets survive the standardization, and the PCA in Eqs. 3-4 then picks up any residual anisotropy. Because the Plavin et al. jet position angles describe the same brightness distribution, an excess of small |ϑast − ϑjet| can arise from technical structure effects with no intrinsic centroid wander. The paper's own example B1038+52A is a ~1.5 mas 'dual source' whose elongation lies along the jet; smooth evolution of two component flux densities would mimic a 'coherent transition northward' over 12 yr without any physical core shift. The Section 4 isotropy assumption addresses declination-dependent systematics, but not this jet-aligned, source-dependent term. The central physical claim—that roughly half of well-observed ICRF quasars show intrinsic, jet-related position wander—is therefore not yet cleanly separated from a known VLBI systematic. Missing a structure-index control is the load-bearing gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes single-epoch VLBI position measurements of 265 ICRF3 quasars with at least 200 sessions. For each source, residuals from the maximum-likelihood mean are standardized with the per-session formal covariance (Eq. 1), and the empirical covariance of the standardized residuals is diagonalized to obtain an elongation ratio ε and a preferred position angle ϑ (Eqs. 3–4), with bootstrap-based uncertainties. The resulting ϑ values are compared with independent radio jet position angles from Plavin et al. (2022). The authors report that 49% of the 265 sources have ε > 1.3 and that, among 127 sources with σϑ < 10°, nearly half show astrometric excursions aligned with jet directions within ±30°. They also report correlations between misalignment and small viewing angles and between misalignment and small ε.","tokens_in":9581,"tokens_out":2196,"duration_ms":23768,"significance":"If the central claim holds, it is an important result: a substantial fraction of well-observed ICRF quasars show intrinsic, jet-related astrometric wander, with direct implications for the stability of the celestial reference frame and for VLBI geodetic applications. The analysis is commendably non-parametric, uses an independent external catalog for jet directions, and provides bootstrap-based uncertainties for individual ϑ values; there are no fitted parameters that directly target the alignment fraction. The main weaknesses are that the headline alignment fraction lacks a formal significance test and that the interpretation of ϑ as a purely physical direction depends on an unvalidated assumption of isotropic technical noise, which is especially concerning given known VLBI source-structure effects.","major_comments":[{"comment":"The headline claim that 'nearly one-half' of the 127 sources are aligned within ±30° is stated without a formal significance test or confidence interval. The histogram comparison to a triangular null distribution is visual; a quantitative assessment is needed, for example a bootstrap or permutation p-value for the excess in the central bin, or a confidence interval on the aligned fraction. Without this, the central quantitative claim is not statistically supported.","section":"§5, Fig. 3"},{"comment":"The standardization in Eq. (1) removes only the per-session formal covariance C_xy, which does not include VLBI source-structure errors. For a resolved core-jet or dual-component source, the measured group delay and apparent position depend on array geometry and the source brightness distribution, so structure-induced offsets survive the standardization and can produce anisotropic u distributions. Because the Plavin et al. jet angles describe the same brightness distribution, an excess of small |ϑast − ϑjet| can arise from technical structure effects without any intrinsic centroid wander. A control using a source-structure index, or a comparison restricted to compact, unresolved sources, is needed to separate these effects.","section":"§3, Eq. (1) and §5"},{"comment":"The assumption that 'the cosmic error is assumed to be isotropic' is load-bearing for interpreting ϑ as the physical direction of intrinsic excursions, but it is not validated. If unmodeled VLBI systematics (declination-dependent errors, network-geometry effects, or time-correlated station errors) are anisotropic, the estimated ϑast can be biased and the apparent alignment with jets could be partly artificial. The authors should test this assumption, for example by comparing ϑast determinations from independent subsets of sessions or by injecting simulated isotropic cosmic errors into real covariance structures and checking the recovery of ϑ.","section":"§4"},{"comment":"The sample is restricted to sources with at least 200 sessions, and the comparison to jet directions is further restricted to σϑ < 10° (127 sources). The paper does not discuss how these selection cuts might affect the alignment fraction, for example if sources with many sessions are preferentially those exhibiting detectable variability or if the σϑ cut correlates with source compactness. A sensitivity analysis or a discussion of selection effects is needed to support the generality of the claim.","section":"§2 and §5"}],"minor_comments":[{"comment":"The bootstrap description says subsets are selected 'without repetition,' which is subsampling rather than standard bootstrap resampling; the text should clarify this choice and justify the use of half-size subsets.","section":"§3"},{"comment":"The sentence 'The greatest elongation ϵ = 5.43 is found for IVS B1038+52A' is followed by a description of a 'dual source,' which is more naturally interpreted as resolved structure; this connection to the structure-systematics issue should be acknowledged explicitly, since the example actually illustrates the concern raised in the major comments.","section":"§4"},{"comment":"The statement 'Counting all cases in excess of the expected distribution within the interval [−30°, +30°]' is ambiguous: it is unclear whether the fraction quoted refers to the number of sources in that bin minus the expected triangular count, or the total number of sources in that bin. Please define the fraction precisely.","section":"§5"},{"comment":"The word 'stricture' in 'possible hierarchical stricture' appears to be a typo for 'structure'.","section":"§6"},{"comment":"The author name 'Moór' is typeset as 'Mo´ or' in the text and reference list; please fix the encoding.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is suitable in scope for the journal, and the empirical methodology is a reasonable approach to an important problem. The main risk is that the physical interpretation depends on separating intrinsic centroid wander from VLBI source-structure systematics; the current manuscript does not include a structure-index control, so the central claim is not yet fully supported. A revision that adds a significance test for the alignment fraction and a structure-based control, or at least a frank quantitative discussion of the expected contamination, would be needed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper does something genuinely new. It applies PCA to VLBI residuals standardized by per-session formal covariances and derives, for 265 ICRF3 quasars, a preferred direction of astrometric scatter with bootstrap uncertainties. The comparison with the independent Plavin et al. (2022) jet catalog is an external check, not a fit, and the per-source parameter table is a useful resource for ICRF stability work.\n\nThe headline claim—nearly half of sources with σϑ<10° have excursions aligned with their jets within ±30°—is suggestive but not yet secure. The histogram of PA differences is compared to an isotropic triangular distribution by eye; there is no permutation test or error bar on the fraction. Threshold choices (≥200 sessions, σϑ<10°) could bias the rate, and the paper doesn't explore that sensitivity.\n\nThe more serious issue is the Section 4 isotropy assumption. Standardizing by the formal covariance removes the expected declination–R.A. anisotropy, but it does not remove systematic errors from unmodeled source structure. For a resolved core-jet or dual-component source, the VLBI position offset depends on array geometry and the brightness distribution, and those errors have a preferred direction—the jet direction. The paper's own example B1038+52A is a dual source elongated along the jet; its astrometric wander could be driven by smooth flux-density evolution of the two components rather than a physical core shift. If this mechanism contributes broadly, the alignment is partly technical, because the same brightness distribution produces both the jet angle and the astrometric scatter direction. The paper lacks a control sample of compact, unresolved sources or any structure-index comparison to separate these contributions.\n\nThe core empirical result probably survives: even if the wander is due to source structure rather than opacity-driven core shifts, the alignment matters for ICRF stability and Earth orientation. But the interpretation as jet-related core shifts is not cleanly supported yet.\n\nThis deserves a serious referee. The method is novel, the data are valuable, and the problems are addressable: add a permutation test, check threshold sensitivity, and include a structure control. I would not cite it for the strong physical interpretation until those are in.","headline":"A novel PCA-based measurement of preferred astrometric excursion directions for 265 ICRF quasars, with a suggestive jet-alignment signal that needs a structure control before the physical interpretation holds.","tokens_in":10197,"tokens_out":4554,"would_cite":true,"duration_ms":45927,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Nearly half of well-observed ICRF quasars show astrometric excursions aligned with their radio jet directions within ±30°.","keywords":["astrometry","VLBI","ICRF3","radio-loud quasars","radio jets","position angle","core shift","bootstrapping"],"falsifier":"Simulate isotropic, heavy-tailed time series with the same session timings and formal covariances as the real data, run them through the identical standardization and angle-estimation pipeline, and compare the resulting ϑast against the same Plavin et al. jet catalog. If the simulated data yield a similar excess of alignments within ±30° after subtracting the triangular null distribution, the claimed alignment is a statistical artifact; if not, the alignment reflects real source physics.","tokens_in":9153,"feed_emoji":"📡","tokens_out":8369,"duration_ms":82045,"temperature":0.7,"pith_summary":"The paper argues that the apparent positional jitter of ICRF reference quasars, measured by decades of VLBI sessions, is not purely noise: for a large fraction of well-observed sources the jitter is elongated along the same direction as the radio jet. Working with 265 sources with at least 200 single-epoch position measurements, the authors standardize each measurement by its formal covariance to remove the known north-south bias of VLBI astrometry, then measure the angle of maximal scatter of the residuals. After comparing with an independent catalog of jet directions, they find that nearly half of the 127 sources with the most secure angles have astrometric excursions aligned with the jet within ±30°. If correct, this implies a physical link between time-variable apparent positions and jet-related core shifts, with practical consequences for how reference frame and Earth-orientation solutions treat quasar noise.","feed_headline":"Nearly half of quasars wander along their radio jets","feed_subtitle":"VLBI data over 40 years show astrometric jitter aligned with jet structures, pointing to core shifts.","key_machinery":"The key machinery is the standardized offset vector u = $C^{{-1/2}}$ [x,y]^T, which transforms each single-epoch VLBI position residual into a coordinate system where the formal covariance is the identity matrix. The paper then computes the empirical covariance matrix of the u-values for each source, extracts its eigenvectors and eigenvalues, and defines the elongation ε = ε1/ε2 and the position angle ϑ of the major axis from the largest eigenvector. This standardization is what separates the well-understood technical elongation of VLBI errors in the declination direction from any remaining physical anisotropy; bootstrapped subsets of half the data provide robust uncertainties for ε and ϑ, and only sources with σϑ < 10° are used in the jet-alignment comparison.","core_discovery":"The paper's central claim is that the excess astrometric scatter of ICRF3 quasars is not directionless: after removing the known anisotropic component of VLBI measurement noise, a large fraction of well-observed sources show residual wander that is elongated along the same position angle as their radio jets. Based on 265 sources with at least 200 single-epoch VLBI measurements and using the principal axis of the empirical covariance of standardized residuals, the paper reports that 49% of the sample has elongation exceeding 1.3 and that, among the 127 sources with robustly determined angles (σϑ < 10°), nearly half are aligned with jet directions from an independent catalog (Plavin et al. 2022) within ±30°, after subtracting the expected random alignment. The authors interpret this as evidence that jet-related core shifts contribute substantially to apparent source instability, while noting that alignment is not universal and that some sources are confidently misaligned.","pith_inferences":["A likely extension is to apply the same standardized-eigenanalysis statistic to Gaia optical astrometric time series, which would test whether the anisotropic wander is common across wavelength regimes or specific to VLBI core emission.","The paper's negative correlation between alignment and viewing angle predicts that blazars viewed almost along the jet should show more isotropic astrometric scatter; this is testable with existing VLBI and MOJAVE samples.","The counting method subtracts a triangular null distribution to infer the aligned fraction, so the 'nearly half' figure is a population statement; a per-source significance test might identify a smaller set of individual sources with convincing alignment, which would be a more direct basis for source-dependent downweighting.","If a large fraction of defining ICRF sources have jet-aligned wander, the frame itself may carry a slowly time-varying, direction-dependent distortion that is not captured by formal errors; geodetic solutions might need to include source-specific, direction-dependent noise in the covariance model."],"forward_implications":["If the alignment is real, the dominant source of astrometric wander for many ICRF3 quasars is physical (jet-related core shifting) rather than random technical noise, implying that VLBI source noise cannot be treated as isotropic in reference frame work.","Earth-orientation and UT1–UTC monitoring solutions, which are most sensitive to the R.A. component of source positions, should downweight ICRF sources with large elongation ε and position angles near 90°, as the paper explicitly recommends.","Single-band, time-variable positions can serve as a probe of the core-jet geometry on scales of a few milliarcseconds, complementing multi-wavelength offset studies that established the core-jet model.","Because the alignment is partial, the sample should be split into aligned and misaligned sources to search for differences in variability, jet structure, or spectral properties, which could reveal additional mechanisms affecting apparent positions."],"supporting_citations":[{"why":"provides the usn2023a diurnal VLBI time series and the maximum-likelihood framework for computing mean positions and standardized residuals, the exact data and foundation this analysis builds on.","marker":"Cigan et al. 2024"},{"why":"supplies the independent catalog of radio jet position angles against which the astrometric excursion angles are compared.","marker":"Plavin et al. 2022"},{"why":"quantifies core-shift magnitudes in VLBI sources, the physical mechanism invoked to explain why position excursions align with jet directions.","marker":"Pushkarev et al. 2012"},{"why":"establishes the core-jet model for systematic position offsets across radio bands and optical, which this paper extends to time-variable positions within one band.","marker":"Lambert et al. 2021"},{"why":"earlier detection of preferred directions of single-epoch position offsets, mostly along the technical south-north axis, which motivates the covariance standardization used here.","marker":"Gattano & Charlot 2021"}],"fun_headline_variants":["Quasar jitter aligns with radio jets","40 years of VLBI show quasar wobble tracks jets","Nearly half of quasar motion follows jet direction","Astrometric drift in quasars points along jets","Quasar position excursions mirror jet orientations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"After removing the known north-south bias of VLBI errors by standardizing each measurement, the paper assumes the leftover 'cosmic' scatter is directionally isotropic, so any remaining elongation must point along a real physical direction of the source rather than an unmodeled technical systematic.","fun_headline_variants_meta":{"raw":{"variants":["Quasar jitter aligns with radio jets","40 years of VLBI show quasar wobble tracks jets","Nearly half of quasar motion follows jet direction","Astrometric drift in quasars points along jets","Quasar position excursions mirror jet orientations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0002,"raw_usage":{"total_tokens":1368,"prompt_tokens":933,"completion_tokens":435,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":549,"completion_tokens_details":{"reasoning_tokens":363}},"tokens_in":549,"tokens_out":435,"duration_ms":4777,"temperature":1.0,"reasoning_tokens":363,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:42:08.299397+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate isotropic, heavy-tailed time series with the same session timings and formal covariances as the real data, run them through the identical standardization and angle-estimation pipeline, and compare the resulting ϑast against the same Plavin et al. jet catalog. If the simulated data yield a similar excess of alignments within ±30° after subtracting the triangular null distribution, the claimed alignment is a statistical artifact; if not, the alignment reflects real source physics.","supporting_citations":[{"cited_title":"Metrics of Astrometric Variability in the International Celestial Reference Frame: I. Statistical analysis and selection of the most variable sources","cited_arxiv_id":"2408.01373","evidence_quote":"provides the usn2023a diurnal VLBI time series and the maximum-likelihood framework for computing mean positions and standardized residuals, the exact data and foundation this analysis builds on."},{"cited_title":"F., et al","cited_arxiv_id":null,"evidence_quote":"establishes the core-jet model for systematic position offsets across radio bands and optical, which this paper extends to time-variable positions within one band."}],"review_version":1}