{"id":"185d71b2-9cb5-474d-a3ab-b43ca748d5a8","arxiv_id":"2607.21236","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Fast and slow CMEs both expand faster laterally than radially, but they differ in how expansion speed relates to angular width at 10 solar radii.","lead":"This study reconstructs 14 coronal mass ejections in 3D and finds they expand roughly 1.7 times faster sideways than radially, and that fast and slow events do not follow the same rules. The results argue for treating fast and slow CMEs as separate populations in space-weather and CME-evolution statistics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The lateral expansion speeds rest on an unreproducible R_lat definition: Appendix Eq. (g)-(h) yields values incompatible with §2.3, so V_lat/V_rad and the fast/slow correlations may change with the correct formula.","rationale":"Reading in good faith, the paper has real strengths: multi-viewpoint GCS fitting is a standard and appropriate tool, the comparison with the full cone model is informative, and the broad result that lateral expansion exceeds radial expansion is consistent with earlier work. The reader's conditional verdict is reasonable. However, the most load-bearing part of the paper is the derived lateral expansion speed V_lat, which feeds both the headline asymmetry ratio and the novel fast/slow correlation. The manuscript gives a precise mathematical definition of R_lat in Appendix A/Table 3, but that definition appears to be numerically incompatible with the lateral dimensions reported in §2.3 for the same events. If the published formula were actually used, the lateral widths would be roughly a factor of two smaller, and the derivative—hence V_lat and the asymmetry ratio—would differ. If instead a different formula was used, it is not documented. This is not a mere style issue; it is an internal consistency problem in the central measurement. The reader's weakest-assumption identification about GCS geometry is related but broader; the concrete inconsistency I identify is more specific and testable. It does not automatically prove the conclusions wrong, but it makes the central numbers not reproducible as written, which is a correctness risk that must be resolved before the 'fundamental difference' conclusion can be accepted. The n=7 sample issue compounds this fragility. My recommendation is to keep the reader's conditional verdict: the paper should be accepted only after the R_lat formula is clarified and the correlations are rechecked, or rejected if the recheck changes the sign of the fast-CME result.","tokens_in":26917,"tokens_out":16410,"duration_ms":175270,"concrete_test":"Recompute R_lat for all 14 events at 10 R_sun using Eq. (g)-(h) of Table 3 with the fitted α, κ, and h_f, and compare against the 2R_lat ranges quoted in §2.3. Then re-derive V_lat from this formula and re-run the Figure 6 correlations between V_lat and face-on angular width for fast and slow CMEs. If the recomputed widths do not match the quoted ranges, the paper must state the actual R_lat formula used; if the sign or magnitude of the fast-CME correlation changes, the 'fundamentally different evolutions' claim is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central asymmetry ratio V_lat/V_rad (≈1.7) and the opposite-sign fast/slow correlations at 10 R_sun all depend on R_lat, but the manuscript provides two incompatible definitions. Appendix A and Table 3 define R_lat as the maximum of BP_x (Eqs. g-h). Inserting the 2012 Mar 07 GCS parameters from Table 1 (α=30°, κ=0.5, h_f=10 R_sun) into Eq. (g)-(h) gives R_lat ≈ 2.45 R_sun, i.e., 2R_lat ≈ 4.9 R_sun. Yet §2.3 states that at ~10 R_sun the fast CME population has 2R_lat ≈ 10–14 R_sun. The plotted V_lat therefore cannot be the time derivative of the Appendix BP_x maximum; some other formula, possibly based on h_f sin(α+δ), must have been used without being stated. Because V_lat is a numerical derivative of R_lat, an ambiguity in R_lat directly changes the reported asymmetry ratio and can even alter the sign of the fast-CME correlation. Additionally, the fast-CME correlation is based on n=7 events with no significance or leave-one-out testing, making it fragile even if the R_lat definition were unambiguous. The conclusion that fast and slow CMEs follow fundamentally different evolutions should be conditioned on resolving this calculation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript analyzes 14 CMEs (7 fast, 7 slow) using GCS reconstructions from STEREO/SOHO and derives time-dependent radial and lateral expansion speeds, face-on and edge-on angular widths, and correlations at ~10 R_sun. Its main claims are that lateral (in-plane) expansion systematically exceeds radial expansion by an average factor of ~1.7 in both populations; that the full ice-cream cone model's expansion speed tracks lateral rather than radial expansion; and that fast and slow CMEs show opposite correlations between lateral expansion speed and face-on angular width, implying fundamentally different evolutionary pathways. The paper is well structured, makes explicit use of multi-viewpoint GCS fitting, and includes a considered discussion of limitations.","tokens_in":27412,"tokens_out":14648,"duration_ms":152535,"significance":"If the findings are robust, they strengthen earlier evidence for asymmetric CME expansion and provide useful quantitative input for space-weather modeling, particularly in cautioning against fixed cone conversion factors and single-population treatments of fast and slow CMEs. The paper carefully compares its ratios with earlier empirical relations and is transparent about many limitations, including manual GCS fitting subjectivity and small sample size. However, the headline fast/slow distinction rests on seven manually fitted events per population with no significance testing, and the lateral-extent formula at the center of the asymmetry calculation is not reproduced by the equations given in the Appendix. These issues must be resolved before the conclusions can be considered reliable.","major_comments":[{"comment":"Appendix A and Table 3 define the lateral half-dimension as R_lat = max_β BP_x (Eqs. g-h). In Eq. (g), h is the conical-leg height OD, and Eq. (b) gives h_f = (b+ρ)/(1−κ). For the 2012 Mar 07 event (Table 1: h_f=10 R_sun, α=30°, κ=0.5), this relation gives h ≈ 2.89 R_sun; solving Eq. (A1) for the maximizing β gives BP_x,max ≈ 1.64 R_sun, i.e., 2R_lat ≈ 3.3 R_sun. This is far below the 10–14 R_sun range quoted for fast CMEs at ~10 R_sun in §2.3. The quoted range is approximately reproduced if h in Eq. (g) is replaced by h_f, but that reading conflicts with the Appendix's statement that h is the conical-leg height. Since R_lat is the input to V_lat = dR_lat/dt (Table 3, Eq. m), the asymmetry ratio and the correlations in Fig. 6 depend directly on this choice. The authors must reconcile Eq. (g) with §2.3, state the exact formula used for the reported V_lat, and recompute the affected number","section":"Appendix A, Table 3 (g)-(h), §2.3"},{"comment":"The main new result — fast CMEs show a negative correlation between lateral expansion speed and face-on angular width, while slow CMEs show a positive one — is based on seven points per population. No significance test, confidence interval, bootstrap, or leave-one-out result is reported, and the thresholds used to classify correlations as strong/moderate/weak are arbitrary. With n=7, a single influential event (e.g., 2012 Mar 07, also the most extreme in leading-edge speed) can determine the sign. The paper should report p-values or equivalent, a non-parametric alternative such as Spearman's ρ, and a leave-one-out or jackknife sensitivity check. The adopted GCS parameter uncertainties (5% height, ±0.05 κ, ±5–10° α) should also be propagated into the correlation analysis. Without this, the claim that fast and slow CMEs evolve fundamentally differently is not statistically supported.","section":"§3, Fig. 6"},{"comment":"The asymmetry ratio V_lat/V_rad is not a direct observable; it is a derived property of the assumed GCS hollow-croissant geometry. Appendix A explicitly notes that one lateral dimension equals the radial dimension by construction, and the reported 'lateral' direction is the in-plane width of this particular shell. If real CMEs deviate from the croissant shape, the average ratio of ~1.7 and the correlation signs in Fig. 6 could change. The paper should state this conditionality in the conclusions and, ideally, test sensitivity to the shape assumption, for example by comparing with an independent cone-shell or elliptical parameterization, or by quoting the model-induced scatter.","section":"§2.3, Appendix A, §5"},{"comment":"The full cone model expansion speed is computed using ω = α+δ from the same GCS fit and V_LE from the same GCS height-time series. The agreement of V_exp_fullcone with V_lat is therefore in part a consistency check between two derived quantities of one model, not an independent validation. The statements in §4.2 and §5 that the full cone model 'primarily reflects lateral expansion' should be qualified as a consequence of the GCS geometry and the adopted mapping, unless an independent cone-model fit is performed.","section":"§2.4, Eqs. (1)-(2)"}],"minor_comments":[{"comment":"Please specify how the representative values at 10 R_sun are obtained: interpolation between measurements, evaluation of the moving-box fit, or the nearest observed height? The phrase 'around 10 R_sun' is ambiguous, especially since the last tracked heights range from 10.0 to 23.2 R_sun.","section":"§3"},{"comment":"The column headers '3D half face-on width [α+δ]' and '3D half edge-on width [δ]' are clear in context but should explicitly state that the values are half widths in degrees. Also add units to the speed columns for readability.","section":"Table 2"},{"comment":"The caption says 'The dashed line represents the error bars,' but the figure does not show how these error bars were computed. Please define the uncertainty source and the propagation method in the caption.","section":"Fig. 6"},{"comment":"The phrase 'fundamentally different' appears in the conclusions and abstract. Given the small sample and the issues above, a more measured wording such as 'consistent with different evolutionary behaviors, pending a larger sample' would better match the evidence presented.","section":"§4.4"}],"recommendation":"major_revision","confidential_remarks":"The inconsistency between Table 3 Eq. (g) and the quoted 2R_lat values in §2.3 is the most serious technical issue; if the authors can clarify the formula and recompute the relevant V_lat values, the paper may become publishable. The opposite-sign fast/slow correlation is interesting but currently lacks statistical support; I would encourage requiring significance testing or at least a leave-one-out analysis. The authors should also be asked to provide the GCS parameter time series as supplementary material so that the fits are reproducible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Anjali et al. do something useful: they apply the GCS model to 14 well-observed CMEs, split them by speed, and compare lateral vs radial expansion speeds and angular-width evolution. The result that lateral expansion exceeds radial (average ratio ~1.7) is consistent with Cremades et al. 2020 and other small-sample work, so the main asymmetry claim is credible. The new bits are the fast-slow split, the opposite-sign correlations at 10 Rsun, and the explicit demonstration that the full-cone model tracks lateral rather than radial expansion. Those are worth having, and the paper is honest about its small sample and user-dependent GCS fitting.\n\nThe central problem is the definition of R_lat. The paper says in §2.3 that at ~10 Rsun, fast CMEs have 2R_lat ≈ 10–14 Rsun. But plugging the 2012 Mar 07 parameters from Table 1 into Eq (g)–(h) of Appendix A gives 2R_lat ≈ 3–5 Rsun, depending on whether you use the leg height h or the leading-edge height h_f. If you use h_f, you get close to 10–14, but Table 3 defines h as the leg height. The authors need to state explicitly which variable goes into Eq (g) and show that the resulting R_lat matches the numbers plotted. Without that, V_lat and the V_lat/V_rad ratio are not reproducible, and the sign of the fast-CME correlation could change with the correct formula.\n\nTwo further soft spots. First, the fast-CME negative correlation is based on seven manually fitted events, with no significance test, no confidence intervals, and uncertainties imported from other studies. That is a weak foundation for the phrase 'fundamentally different evolutions.' Second, the asymmetry is partly inherited from the GCS geometry (one in-plane lateral dimension equals the radial dimension by construction), and the full-cone comparison is an internal-consistency check using the same GCS face-on width. Neither is fatal, but the paper should frame them as such.\n\nWho this is for: people who work on CME expansion statistics and cone-model conversion factors. It deserves a serious referee: the sample, the multi-viewpoint fitting, and the comparison are executed carefully, and the problems are fixable with a clarified formula and minimal statistics (bootstrap or leave-one-out on the correlations). I would not cite it until the R_lat ambiguity is resolved.","headline":"A careful GCS study of fast vs slow CME expansion, but the headline fast/slow correlation is fragile and the lateral-size formula does not reproduce the paper's own numbers.","tokens_in":27774,"tokens_out":5854,"would_cite":false,"duration_ms":59525,"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":"Coronal mass ejections expand sideways (laterally) faster than they expand outward (radially), by about 1.7x, and fast and slow events show opposite width–expansion correlations—so they cannot be treated as one population.","keywords":["coronal mass ejections","CME expansion","asymmetric expansion","Graduated Cylindrical Shell model","angular widths","fast and slow CMEs","stereoscopic reconstruction","space weather"],"falsifier":"Reconstruct the same 14 CMEs with an independent 3D method that does not impose the GCS equal-radial/perpendicular-lateral constraint (e.g., forward-modeling with free elliptical cross-sections or stereoscopic triangulation of the flanks), and check whether lateral-in-plane expansion still exceeds radial expansion by ~1.7 and whether the fast-CME negative width–expansion correlation survives; alternatively, track fast CMEs below 1.5 solar radii to capture the peak expansion phase and test whether the negative correlation is an artifact of missing the low corona.","tokens_in":26881,"feed_emoji":"☀️","tokens_out":7867,"duration_ms":72557,"temperature":0.7,"pith_summary":"Using three-view coronagraph reconstructions of 14 coronal mass ejections, the paper tries to show that CMEs expand asymmetrically: their in-plane lateral (sideways) expansion speed systematically exceeds their radial (outward) expansion speed, by an average factor of about 1.7, in both fast and slow events. It then argues that fast and slow CMEs follow fundamentally different evolutionary paths, because at 10 solar radii the two groups show opposite correlations between expansion speed and angular width: wider slow CMEs are the ones expanding fastest, while faster-expanding fast CMEs tend to be narrower. The authors further claim that the long-used full ice-cream cone model's expansion speed actually tracks the lateral, not the radial, expansion, so fixed-cone conversion factors misrepresent CME evolution. If true, statistical treatments that mix all CMEs into one population, and forecasts that extrapolate coronagraphic widths and speeds with constant cone factors, need reworking—which matters for predicting CME arrival times and impact at Earth.","feed_headline":"CMEs expand sideways 1.7x faster than outward—fast or slow","feed_subtitle":"Cone-model forecasts miss the radial component; fast and slow CMEs can't share one width-speed rule.","key_machinery":"The load-bearing tool is the Graduated Cylindrical Shell (GCS) model, which fits each CME as a hollow 'croissant' made of two conical legs and a torus-shaped front, specified by six parameters including the leg half-angle α and aspect ratio κ. From α and κ the paper computes two angular widths—face-on (≈2(α+δ), along the torus axis) and edge-on (≈2δ, poloidal)—and two size measures: the radial flux-rope radius R_rad = κ/(1+κ) h_f and the in-plane lateral half-extent R_lat. The key geometric identity is that the croissant's circular cross-sections make the radial dimension equal to the lateral dimension perpendicular to the propagation plane, so the only genuinely independent lateral expansio","core_discovery":"On the paper's own terms, the central discovery is that asymmetric expansion is a persistent property of CMEs in the coronagraphic height range (~2–20 solar radii): for all 14 events the lateral expansion speed measured in the plane of propagation is higher than the radial expansion speed along the propagation direction, with a mean ratio V_lat/V_rad ≈ 1.7 at the final tracked height. The second principal finding is a fast/slow dichotomy at 10 solar radii: slow CMEs show strong positive correlations between expansion speeds and the corresponding angular widths (lateral with face-on, radial with edge-on), whereas fast CMEs show a negative correlation between lateral expansion speed and face-o","pith_inferences":["The reported asymmetry may be partly baked into the model: GCS forces the perpendicular lateral dimension to equal the radial dimension, so the 'asymmetry' examined is between two directions that the model does not constrain independently. A natural test is to fit the same events with a model that relaxes circular cross-sections and see whether V_lat/V_rad ≈ 1.7 survives.","A testable prediction follows from the authors' 'still expanding' interpretation: fast CMEs' face-on widths should continue to grow beyond 10 solar radii and their lateral expansion speeds decline, so a heliospheric-imager sample extending to 20–50 solar radii should show the fast-CME negative correlation weaken or reverse.","The fast/slow classification at ~2–5 solar radii divides a continuum; with only seven events per bin and hand-fitted GCS parameters, the opposite correlation signs could be driven by the two or three most extreme fast events rather than a true dichotomy, so a larger sample with reported significance levels is needed.","If the asymmetry is real, MHD simulations of CMEs launched with elliptical rather than circular cross-sections should reproduce the observed ratio and the fast/slow differences, providing an independent check."],"forward_implications":["If CMEs expand asymmetrically, empirical relations like V_rad = 0.88 V_exp and fixed f(ω) cone factors misidentify which dimension they measure; the paper shows the cone-model expansion speed is essentially the lateral speed, so radial speeds inferred from cone fits are likely overestimated.","Fast and slow CMEs should be analyzed separately; combining them can produce misleading positive width–speed correlations (as the paper notes earlier studies found) that hide opposing behaviors.","CME parameters at 10 solar radii are not a universal snapshot: slow CMEs are near the end of their expansion phase there, while many fast CMEs are still expanding, so arrival-time and impact-width predictions need height-dependent, direction-dependent expansion.","Time-dependent, deprojected 3D parameters (GCS or equivalent) should replace constant angular-width cone approximations in CME kinematic and heliospheric models.","Observing fast CMEs from lower coronal heights (~1.5 solar radii and below) is necessary to capture their peak expansion phase, which is missed in the current sample and likely drives the negative correlation."],"fun_headline_variants":["CMEs expand sideways 1.7x faster than outward","CME expansion is asymmetric: lateral outruns radial","Fast and slow CMEs expand differently—study","Lateral CME expansion beats radial by 1.7x","Cone model fails: CMEs expand asymmetrically"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The conclusion assumes that the GCS hollow-croissant geometry—with its circular cross-sections tying the perpendicular lateral width to the radial width—is the true shape of real CMEs; if actual CMEs depart from that shape, the measured asymmetry and the opposite fast/slow correlations could be artifacts of the fitting model.","fun_headline_variants_meta":{"raw":{"variants":["CMEs expand sideways 1.7x faster than outward","CME expansion is asymmetric: lateral outruns radial","Fast and slow CMEs expand differently—study","Lateral CME expansion beats radial by 1.7x","Cone model fails: CMEs expand asymmetrically"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000269,"raw_usage":{"total_tokens":1501,"prompt_tokens":833,"completion_tokens":668,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":577,"completion_tokens_details":{"reasoning_tokens":595}},"tokens_in":577,"tokens_out":668,"duration_ms":6294,"temperature":1.0,"reasoning_tokens":595,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T08:03:03.486542+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Reconstruct the same 14 CMEs with an independent 3D method that does not impose the GCS equal-radial/perpendicular-lateral constraint (e.g., forward-modeling with free elliptical cross-sections or stereoscopic triangulation of the flanks), and check whether lateral-in-plane expansion still exceeds radial expansion by ~1.7 and whether the fast-CME negative width–expansion correlation survives; alternatively, track fast CMEs below 1.5 solar radii to capture the peak expansion phase and test whether the negative correlation is an artifact of missing the low corona.","supporting_citations":[],"review_version":1}