-
Overgeneralization from a single snapshot: the manuscript frequently uses definitive/causal language (e.g., “definitive,” “primary engine,” “direct causal link,” “dictates the entire filamentary architecture”) while the analysis is based on one instantaneous snapshot of an intrinsically intermittent convective/turbulent system (Abstract; Sec. 1; Sec. 3; Sec. 4). Without temporal context, it is unclear how representative the reported filamentarity, extreme intermittency (very large kurtosis), Q-structure conclusions, and footprint contrasts are.
Recommendation: Recalibrate claims throughout (Abstract; Sec. 1; Sec. 3; Sec. 4) to explicitly frame the work as a detailed case study of one snapshot (“in this model snapshot…”, “suggests…”, “is consistent with…”). If feasible, add a minimal multi-snapshot robustness check (even 3–5 snapshots bracketing the analyzed time): show time variability of (i) a mass-transfer proxy (e.g., flux through an L1 control surface or across $\Phi = \Phi_{\rm L1}$), (ii) key PDF moments (mean/variance/kurtosis of $|{\bf j}|$) in global and L1 regions (Sec. 3.2), and (iii) footprint area/fraction and Table 3 mean ratios (Sec. 3.4). At minimum, report where the snapshot sits in the simulation timeline (time since RLOF onset; fractions of orbital/convective turnover times) and discuss representativeness explicitly in Sec. 4.
-
Insufficient documentation of the simulation and gravity/potential consistency limits reproducibility and interpretation. Sec. 2.1 mentions Athena++ but omits key physical/numerical parameters (M1, M2, q, separation/period, rotation state, donor radius/luminosity/T_{\rm eff}, EOS, opacities, radiation closure/transport, resolution, domain extents, boundary conditions, companion treatment/sink, runtime). Moreover, the analysis builds an effective Roche potential (Sec. 2.1–2.2), but it is unclear whether this potential matches the gravity actually used in the simulation (e.g., point-mass vs self-gravity/monopole/multipole, indirect terms, softening). In an extended, non-spherical donor, “outside Roche lobe” can be sensitive to these choices.
Recommendation: Expand Sec. 2.1 with a concise but complete simulation summary: system parameters ($M_1$, $M_2$, $q$, $a$, $P$, $\Omega$), donor properties, rotation/corotation assumptions, grid ($n_r$, $n_\theta$, $n_\phi$), extents, boundary conditions, EOS, radiation method (e.g., FLD/M1; comoving vs lab-frame moments), opacity treatment, gravitational softening/sink/accretion prescription, and evolution time to the analyzed snapshot. Clarify explicitly what gravitational potential the simulation evolves under, and use that same potential field in the Roche-mask analysis if available. State how $\Phi_{\rm L1}$ and the L1 location are found numerically (e.g., 1D search along the line of centers vs saddle search), and report $x_{\mathrm{L1}}$ and $\Phi_{\rm L1}$ (Sec. 2.1–2.2). Add a short sensitivity test showing how the “outside Roche lobe” mask fraction changes under modest perturbations of $\Phi_{\rm L1}$ (or equivalent threshold).
-
The “source fingerprinting” method (Sec. 2.5) is central to the paper’s main claim (convective upwellings preferentially supply the stream), but the algorithmic details and robustness are not sufficient to support causal/exclusive phrasing. It is unclear whether trajectories are true streamline integrations through the 3D field or a single back-projection; the seeding layer “just outside the Roche lobe” is not precisely defined; and failure cases (non-intersection with the surface, leaving domain) and interpolation/step-size errors are not discussed. In a strongly accelerating, curved, compressible flow near L1, a simplistic back-projection can misidentify origins.
Recommendation: Substantially expand Sec. 2.5 with operational details: (i) define seeding cells (exact radial range/thickness relative to $\Phi = \Phi_{\rm L1}$ and any $v_r > 0$ or $|{\bf j}|$ thresholds), (ii) specify whether you integrate $d{\bf x}/ds = -{\bf v}({\bf x})$ (and which integrator, step size/adaptive control, interpolation scheme), or do a one-step projection; (iii) list termination criteria and how you treat paths that do not hit the photospheric surface or exit the domain; (iv) report the traced mass fraction (or number of stream cells) that successfully maps to the surface. Add robustness tests: vary seeding-layer thickness and photospheric radius ($\pm 5$–$10\%$) and show Table 3 ratios/footprint morphology are stable (Sec. 3.4). Replace “direct causal link”/“exclusive” with “strong association consistent with…” unless additional validation is provided (e.g., forward–backward consistency checks, or time-integrated tracer particles/passive scalars if available). Quantify overlap with objectively defined upwellings (e.g., $v_r$ percentile threshold) via contingency tables/odds ratios or KS tests on $v_r$ and $F_{r,{\rm rad}}$ distributions (Sec. 3.4).
-
Force/work decomposition is described (Sec. 2.4) and invoked to argue that convection plus gas/radiation pressure lift material over the barrier (Sec. 3.4; Sec. 4), but the manuscript does not present the quantitative outcomes needed to support relative-importance claims (e.g., gas vs radiation vs gravity; where positive work is done). As written, the central “driving engine” conclusion is not adequately substantiated.
Recommendation: In Sec. 3.4, add quantitative conditional statistics for forces and power densities: PDFs and/or mean/median ratios of $|{\bf F}_{\rm gas}|/|{\bf F}_{\rm grav}|$, $|{\bf F}_{\rm rad}|/|{\bf F}_{\rm grav}|$ and $W_{\rm gas}$, $W_{\rm rad}$, $W_{\rm grav}$ in (a) fingerprint surface cells, (b) non-fingerprint surface cells on the facing hemisphere, and (c) the L1 stream mask. Consider also reporting signed work (how often $W_{\rm rad} > 0$, $W_{\rm gas} > 0$) to demonstrate systematic acceleration vs random fluctuations. Clarify that ${\bf F}_{\rm rad} = -\nabla\cdot{\bf P}_{\rm r}$ is computed consistently with the simulation’s radiation closure/frame (Sec. 2.4). Use these results to refine (and if necessary soften) claims in Sec. 4 about which force actually dominates the launching/acceleration in this model.
-
Roche potential equation and reference-frame specification are currently ambiguous/problematic, undermining the Roche-mask and $\Phi_{\rm L1}$ definitions (Sec. 2.1–2.2). The Roche-potential expression appears dimensionally inconsistent as written (companion term formatting) and the centrifugal term is not explicitly defined with respect to origin/axis; if donor-centered coordinates are used, indirect terms/offsets must be handled consistently.
Recommendation: Rewrite the Roche potential in a fully parenthesized, standard form with explicit coordinate dependence and clearly defined origin and rotation axis (barycentric vs donor-centered rotating frame). State whether indirect terms are included and ensure consistency with how $\Phi_{\rm L1}$ is computed and how the simulation’s rotating frame is defined (Sec. 2.1). After rewriting, verify that the “outside Roche lobe” criterion ($\Phi_R > \Phi_{\rm L1}$) matches the stated sign convention and yields the intended geometry (Sec. 2.2).
-
Q-criterion-based claims (“no stable, long-lived vortices,” “developed turbulence,” “shear-dominated”) are based on a single snapshot and a largely threshold-based, qualitative connected-component analysis (Sec. 2.3; Sec. 3.3; Fig. 4). For compressible, stratified, shearing flows, $Q$ alone can be subtle to interpret; results may depend strongly on $Q_{\rm thresh}$, resolution, derivative noise, and whether structures are identified in true 3D connectivity or inferred from slices.
Recommendation: In Sec. 2.3 and Sec. 3.3, (i) report $Q$ distribution statistics (percentiles or mean/$\sigma$) and reconcile any inconsistent descriptions of thresholding; (ii) perform a threshold-sensitivity scan (e.g., vary $Q_{\rm thresh}$ over $1$–$2$ dex or in units of $\sigma_Q$) and show how the number/volume of connected positive-$Q$ structures changes; (iii) state clearly the 3D connectivity definition (e.g., 6/18/26-neighbor equivalent), minimum voxel count, any smoothing/filtering, and masking (Sec. 3.3). If feasible, add one complementary diagnostic (e.g., $|\boldsymbol{\omega}|$, enstrophy, or $\lambda_2$) to support the “shear-dominated” interpretation. Temper conclusions to “no large coherent vortices above resolution/threshold limits in this snapshot” unless temporal persistence is demonstrated across multiple outputs.
-
Mass-flux intermittency statistics may be distorted by sampling choices and numerical floors. The extremely large kurtosis (Sec. 3.2) can reflect genuine intermittency, but can also be amplified by inclusion of near-vacuum/background regions, mixing inflow/outflow, cell-volume weighting choices, and pile-ups at density/pressure floors (Table 1). The “global vs L1 region” comparison is hard to interpret because the L1 region geometry is not precisely defined (Sec. 2.2; Sec. 3.2; Table 2).
Recommendation: In Sec. 2.2 and Fig. 2 caption, explicitly state weighting (cell-count vs volume vs mass), binning, whether ghost zones are excluded, and how floors are treated (masking/marking; report fraction of cells at floors). Provide restricted PDFs (or at least key moments) for physically relevant subsets: $v_r > 0$ outflow, outside-Roche mask, and/or within a donor-neighborhood radius, to show intermittency is tied to launching/stream rather than ambient low-density volume. Precisely define the L1 “localized spherical region” (center coordinates and radius) used in Table 2 and demonstrate modest robustness to that radius choice.