-
Family definition/membership and the use of proper vs. osculating elements are not specified clearly enough to support family-level dispersion inferences (Sec. 2.1–2.2, Sec. 3.1–3.2). It is unclear (i) what catalog/version and algorithm produced the family assignments, (ii) whether SemimajorAxis_AU/Eccentricity/Inclination are proper or osculating elements, and (iii) whether interlopers/contamination were mitigated. Using osculating elements (or mixed definitions) can artificially inflate within-family dispersions and introduce epoch-dependent noise, directly impacting $\mathrm{Std}_{\mathrm{SemimajorAxis\_AU}}$ and related conclusions.
Recommendation: In Sec. 2.1–2.2 and Sec. 3.1, explicitly document: (a) the family catalog used (name, reference, version/date), family identifiers, and membership definition; (b) whether orbital elements are proper (preferred for family work) or osculating, including the source (e.g., AstDyS/proper-element service) and the exact variables used (proper $a$, $e$, $\sin i$ vs $i$). If osculating elements were used, either redo the primary dispersion metrics with proper elements or provide a quantitative justification/bias estimate (e.g., compare $\mathrm{Std}(a)$ computed from osculating vs proper for a subset of families). Add a brief contamination/interloper discussion: if taxonomic/albedo filtering is not possible, at minimum test robustness by removing obvious outliers in $(a,e,i)$ space or by using published family “core” memberships where available.
-
The current proxy for Yarkovsky-driven spreading—unconditioned within-family dispersion of semimajor axis (e.g., $\mathrm{Std}_{\mathrm{SemimajorAxis\_AU}}$)—is physically coarse and conflates multiple processes (initial ejection field, resonance dynamics, truncation by family-identification boundaries, and size-dependent Yarkovsky V-shapes) (Sec. 2.4.1, Sec. 3.2–3.4, Sec. 3.5). This weakens the causal interpretation of any correlation between orbital dispersion and spin-state diversity.
Recommendation: Strengthen the orbital-spreading metric in Sec. 2.4.1 and re-evaluate key results in Sec. 3.3–3.4 using at least one more physically grounded alternative: (a) a V-shape slope metric ($|a-a_c|$ vs $1/D$ or $H$) for each family where sizes/$H$ are available; (b) size-conditioned dispersion (e.g., $\mathrm{Std}(a)$ computed within narrow diameter bins, then aggregated); and/or (c) $\mathrm{IQR}(a)$ and robust scale estimators with explicit controls for family central $a$. Demonstrate that $\mathrm{Std}(a)$ tracks (correlates with) the V-shape/Yarkovsky metric for the same families, or explicitly limit the claim to “orbital width” rather than “integrated Yarkovsky drift.” Flag/exclude families strongly affected by nearby major mean-motion/secular resonances (or include resonance-proximity indicators as covariates) and report whether the spin-diversity coefficient remains stable (Sec. 3.4.3, Sec. 3.5).
-
Filtering on sparse spin/obliquity/age availability likely induces strong selection effects and sampling-variance artifacts (“collider bias”), and the manuscript currently treats family-level spin dispersion (e.g., $\mathrm{IQR}_{\mathrm{SpinPeriod\_hr}}$) as comparably measured across families despite widely varying $N_{\mathrm{spin\_members}}$ and heterogeneous observing strategies (Sec. 2.2–2.3, Sec. 3.1–3.2, Sec. 3.5). With small $n$ (e.g., 5 spin periods), IQR estimates are noisy and can correlate with the number and heterogeneity of observed targets rather than the true family distribution.
Recommendation: In Sec. 3.1–3.2, report for each of the 50 families: $N_{\mathrm{members\_total}}$, $N_{\mathrm{spin\_members}}$, $N_{\mathrm{obliq\_members}}$, and (where possible) completeness fractions (e.g., $N_{\mathrm{spin\_members}} / N_{\mathrm{members\_total}}$). Add a sensitivity suite in Sec. 3.3–3.4.3: (a) rerun key correlations/regressions at stricter thresholds (e.g., $N_{\mathrm{spin\_members}} \geq 20$ and/or $\geq20\%$ completeness) and compare coefficients; (b) include $N_{\mathrm{spin\_members}}$ (and possibly $N_{\mathrm{obliq\_members}}$) as explicit covariates; (c) perform within-family bootstrap/resampling of the observed spin sample to estimate uncertainty on $\mathrm{IQR}_{\mathrm{SpinPeriod\_hr}}$ and propagate it into an errors-in-variables or weighted regression. Summarize the domain of validity in Sec. 3.5/5 (e.g., “large, well-observed families with adequate spin sampling”).
-
Family\_Age\_Gyr is central to the analysis but its provenance, construction, and uncertainties are under-specified and not propagated (Sec. 2.4.4, Sec. 3.2–3.4). The text implies member-level ages aggregated to a family-level age when “varied significantly,” which is conceptually unclear for family ages and risks introducing non-reproducible branching logic and biased regression coefficients.
Recommendation: In Sec. 2.4.4, provide a reproducible age model: list the age source(s) with references/versioning; clarify whether ages are truly per-family (preferred) or per-object; and define an explicit, deterministic aggregation rule if multiple literature ages exist per family (e.g., median of published estimates, with a recorded spread). Add an uncertainty field per family (quoted literature uncertainty or inter-estimate spread) and propagate it in Sec. 3.3–3.4 via (a) Monte Carlo perturbation of ages within uncertainties, and/or (b) down-weighting/removing families with large age uncertainty. Report how often the key $\mathrm{IQR}_{\mathrm{SpinPeriod\_hr}}$ coefficient remains positive/significant under these perturbations (Sec. 3.5).
-
Statistical robustness is not yet demonstrated at a level commensurate with the small sample size (50 families) and the number of explored correlations/models (Sec. 2.5, Sec. 3.3–3.4). Key gaps include: incomplete multiple-testing control for the correlation matrix; limited influence diagnostics; and limited reporting of effect-size uncertainty (confidence intervals) for the main regression coefficient interpreted as the coupling signature.
Recommendation: In Sec. 2.5 and Sec. 3.3–3.4, explicitly enumerate (a) the set of confirmatory hypotheses (pre-specified) versus exploratory tests; (b) the number of correlation pairs examined; and (c) the regression models compared. Apply a concrete multiple-testing procedure (Holm or FDR) to the correlation results used for claims (Table 1 and any supporting text) and report adjusted $p$-values. For the main regression (Eq. (2) and any alternatives), add: bootstrap or robust-SE confidence intervals; leverage/Cook’s distance and a leave-one-family-out (LOO) plot of the $\mathrm{IQR}_{\mathrm{SpinPeriod\_hr}}$ coefficient; and a robust fit (Huber RLM / Theil–Sen) to show sign/magnitude stability. Include a compact diagnostics figure/table (residuals, heteroskedasticity test, influential points) for the primary model (Sec. 3.4.3).
-
The interpretation of a $\mathrm{Std}_{\mathrm{SemimajorAxis\_AU}}$–$\mathrm{IQR}_{\mathrm{SpinPeriod\_hr}}$ association as evidence for YORP$\rightarrow$Yarkovsky coupling is vulnerable to confounding by unmodeled family/environment properties: belt region (heliocentric distance), thermal/compositional class, resonance proximity, family identification quality/contamination, and observational targeting (Sec. 3.4.3, Sec. 3.5, Sec. 5). These factors can jointly affect both measured orbital width and measured spin diversity without requiring dynamical coupling.
Recommendation: Temper causal language throughout (Abstract, Sec. 1, Sec. 3.4.3, Sec. 5) to emphasize “consistent with” rather than “evidence for” coupling unless robustness checks support stronger phrasing. Then, in Sec. 3.4.3, add controls/proxies for key confounders: family central semimajor axis (or belt region category: inner/mid/outer), inclination/eccentricity regime, resonance-proximity flags, and (where available) taxonomic/albedo indicators. Refit the main model with these added covariates and report whether the spin-dispersion term remains stable. If taxonomy is unavailable broadly, include at least NEOWISE albedo where present or restrict to families with consistent albedo/type as a robustness subset.
-
Spin-state characterization is concentrated on a single metric ($\mathrm{IQR}_{\mathrm{SpinPeriod\_hr}}$), while obliquity—more directly tied to the sign/magnitude of Yarkovsky drift—is underdeveloped and likely too sparse for stable inference at current thresholds (Sec. 2.4.2, Sec. 3.2, Sec. 3.4.2). In addition, spin-period dispersion in linear hours is sensitive to long-period tails; a log-period dispersion may be more appropriate and comparable across families.
Recommendation: In Sec. 3.3–3.4.2, present a systematic metric comparison: repeat the key regression(s) using (a) $\mathrm{Std}_{\mathrm{Log\_SpinPeriod}}$ (with a precise definition; see minor issues), (b) $\mathrm{IQR}(\log P)$, and (c) any feasible obliquity metrics. For obliquity, either (i) explicitly limit/withdraw obliquity-based conclusions due to small $n$ and treat results as exploratory, or (ii) restrict to a higher $N_{\mathrm{obliq\_members}}$ threshold and use metrics better matched to Yarkovsky expectations (e.g., fraction near $0^\circ/180^\circ$, bimodality indicators, or prograde/retrograde split if sense can be inferred). Report null results clearly as constraints rather than omitting them.
-
The paper reports fitted slopes (e.g., age–dispersion and spin-IQR coefficients) but does not quantitatively connect them to plausible Yarkovsky/YORP magnitudes across the sampled size and heliocentric-distance range, nor does it benchmark against alternative mechanisms (Sec. 3.4.1–3.5, Sec. 5). This makes it difficult to assess whether the estimated effect sizes are physically reasonable.
Recommendation: In Sec. 3.5, add an order-of-magnitude physical check: compute representative Yarkovsky drift rates (as a function of $D$, $a$, and obliquity assumptions) and compare integrated drift over $0.1$–$3$ Gyr to the observed family-level dispersion scale. Then interpret the regression coefficient for $\mathrm{IQR}_{\mathrm{SpinPeriod\_hr}}$ (or log-spin metric) in terms of plausible YORP-driven spin/obliquity diversification and its impact on the distribution of drift rates. Explicitly discuss alternative drivers (collisional reorientation, resonance-driven spreading, family truncation/contamination) and include at least one simple exclusion/control test (e.g., remove resonance-adjacent families) to show these are unlikely to fully explain the pattern.
-
Data provenance and catalog-level systematics are not described in sufficient detail to evaluate biases and ensure reproducibility (Sec. 2.1–2.2, Sec. 2.4.4, Sec. 2.6). The manuscript references internal filenames (e.g., asteroid\_age.csv) without clear mapping to published sources/versions, and it is unclear which spin/obliquity catalogs (and methods) dominate the final sample.
Recommendation: Add a concise provenance table in Sec. 2.1–2.2 and/or Sec. 2.6 listing each key quantity (family membership, proper elements, diameters/albedos, spin periods, obliquities, ages) with: source catalog/survey, literature reference, version/access date, approximate coverage, and known selection effects (e.g., brightness/lightcurve amplitude targeting for spins; shape-modeling campaign biases for obliquities). Provide a Data/Code Availability statement (end of Sec. 2.6 or after Sec. 5) with a repository link (or planned archive) containing scripts and the derived family-level table used for Tables/Figures.