{"id":"d2d2be66-320c-4d34-b24e-2dd1dea1defc","arxiv_id":"2412.01361","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Two-dimensional hydrodynamic simulations show that tidally assisted escape of vapor from the lunar magma ocean can explain the Moon's sodium and potassium depletion, and predict a leading/trailing surface dichotomy.","lead":"This paper uses 2D hydrodynamic simulations to show that tidal forces from the early Earth could strip sodium and potassium vapor from the molten Moon's surface, forming a circum-Earth disk and depleting the Moon in these elements. It predicts a measurable difference in volatile re-accretion between the Moon's leading and trailing sides.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 1800-2000 K window and ~10^3 yr stop time are computed from constant-composition, constant-temperature magma-ocean source runs (Section 2.3); depletion feedback is neglected and could shift both quantities.","rationale":"The reader and I converge on the same load-bearing point: the source term in the hydrodynamic runs is held constant while the composition-evolution calculation in Section 4 uses those fixed fluxes. This is not a numerical bug; it is a modeling choice stated in Section 2.3, but it is precisely where the central quantitative claim is least secure. I credit the paper for what is independently supported: the FARGOCA modification is described and the code is available; the 65 simulations include explicit resolution and boundary convergence checks; the VAPOROCK-based vapor composition agrees with LAVATMOS and Ito et al. (2015); and the reaccretion-fraction and leading/trailing dichotomy results are novel and plausible. However, those strengths establish the mechanism's existence, not the precise temperature-time window. The paper itself acknowledges additional unquantified effects (condensation, 3D meridional flows, tidal distortion, adiabatic-index uncertainty), but those mostly affect efficiency in a direction that is not always clear. The constant-reservoir assumption has a definite, in-principle sign: it makes early loss too fast and stop times too short. Because the observed Na/K ratios span wide ranges, the model may survive a self-consistent treatment, but the headline numbers cannot be claimed until the coupled depletion calculation is done. Hence the reader's CONDITIONAL verdict remains appropriate; I would not change it.","tokens_in":27182,"tokens_out":6017,"duration_ms":59502,"concrete_test":"Run the same 1-year FARGOCA setup with magma-ocean Na and K contents depleted by factors 0.5 and 0.2 relative to BSE, recomputing partial pressures, mean molecular mass, and adiabatic index with VAPOROCK, and compare the net loss fluxes to a linear scaling of Fig. 6. Then integrate the coupled dM_Na/dt and dM_K/dt equations along the tidal tracks of Fig. 7, with and without convective surface replenishment, and recompute Fig. 9. If the required stop time increases by more than about 1.5-2x, or the best-fit temperature shifts by more than about 100 K, the constant-reservoir assumption is load-bearing and the claimed 1800-2000 K / 10^3 yr match is conditional on it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's quantitative claim is the matching of lunar Na/K depletions via a ~1800-2000 K magma ocean with volatile loss ceasing after ~10^3 yr (Figs. 8-10). This is derived from 1-year FARGOCA runs in which the magma-ocean source is held at fixed surface density, vapor composition, and temperature. Section 2.3 states 'we treat these parameters as constants throughout the simulations', and Table 1 fixes vapor atomic fractions at BSE values. VAPOROCK partial pressures of Na and K depend on their activities in the melt, so as the ocean degasses the source strength should fall. The convective-replenishment argument in Section 5 (~20-day overturn vs ~0.4-day atmospheric recycle) only keeps the surface equilibrated with the ocean; it does not replenish the whole-ocean inventory. Over the 10^2-10^4 yr loss timescale the constant-source integration therefore overestimates early loss and underestimates the required stop time. For a target final fraction f=0.3, replacing linear depletion by first-order (activity-proportional) depletion changes the required duration by ln(1/f)/(1-f) ~ 1.7, and the steep temperature dependence of saturation pressure means this can move the best-fit 1800-2000 K window. The hydrodynamic flow itself could also change because the source surface density enters the boundary condition; a simple flux rescaling may not capture changes in reaccretion fraction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes that tidally-assisted hydrodynamic escape from the proto-Moon's magma ocean, operating while the Moon is still close to the Earth, can explain the lunar depletion of Na and K relative to the bulk silicate Earth (BSE). The authors implement a magma-ocean vapor source in the 2D time-dependent FARGOCA code and run 65 inviscid, one-year simulations spanning surface temperatures of 1400-2200 K and Earth-Moon distances of 3-9 Earth radii. Key outputs are the ejection flux, the fraction of vapor re-accreted by the Moon (18% at 3 RE, rising to roughly 99% at 9 RE), the fraction lost to Earth or the outer boundary, and the longitude distribution of re-accretion. Combining the net loss fluxes with a constant-Q tidal migration model, the paper finds that matching the observed Na and K depletions requires volatile loss to cease on timescales of roughly 10^2-10^4 years, and it argues that a stagnant lid or anorthite flotation crust (formed after about 10^3 years in an accompanying thermal model) provides the stopping mechanism, provided the magma ocean surface temperature was 1800-2000 K. A final prediction is that for aM > 3.5 RE, re-accreted volatiles preferentially land on the trailing side, with the strength of the dichotomy depending on the Earth's k2/Q ratio.","tokens_in":27454,"tokens_out":30038,"duration_ms":249840,"significance":"If the quantitative claims survive scrutiny, the paper would establish a formation-model-independent mechanism for lunar volatile depletion and turn the leading/trailing reaccretion asymmetry into a falsifiable observable. The paper has real strengths: reported convergence tests for grid resolution, disk-edge placement, and the Moon's smoothing shell (Section 2.2); a forward-modeling approach in which the observed Na/K abundances enter only as a comparison interval, not as fitted constants; an open code and linked animations; and unusually transparent caveats (Section 6.3) covering condensation, tidal distortion, the 2D approximation, and the sensitivity to the adiabatic index. It also makes a genuine methodological advance over the 1D steady-state treatment, most notably by quantifying re-accretion and demonstrating that gas pressure displaces the Lagrange points away from the Earth-Moon line. The qualitative architecture of the mechanism (efficient escape near 3 RE, increasing re-accretion with distance, trailing-side reaccretion) is more robust than the temperature window, because it depends on the flow geometry rather than on the source strength.","major_comments":[{"comment":"The constant-composition, constant-temperature source is load-bearing for the quantitative claims and is not tested for robustness. Section 2.3 states that the vapor parameters are held constant for the whole simulation, and Section 4 integrates these constant fluxes to produce the depletion curves (Fig. 8) and the inferred stop times (Fig. 9). Yet Section 6.3 shows that doubling the Na and K mass fractions in the melt increases the lunar surface density by a factor of 1.4, i.e., the source strength tracks the melt concentration. Since the convective overturn time of the magma ocean (~20 days, Section 5) is far shorter than the 10^3-10^4-year loss timescales quoted in Fig. 9, the ocean should behave as a well-mixed reservoir whose Na and K activities (and hence partial pressures and total source density) decline as material is removed; the ~0.4-day atmospheric recycle time keeps the surface equilibrated with the ocean but does not replenish the whole-ocean inventory. For a target remaining fraction f ~ 0.3, switching from linear depletion (constant flux) to first-order depletion lengthens the required loss duration by ln(1/f)/(1-f) ~ 1.7 at fixed initial flux, and the differential depletion of Na versus K changes the simultaneous-match condition in Figs. 9-10. The statement in Section 6.3 that an initial excess or deficit of volatiles 'might be balanced by corresponding changes in depletion rates' addresses the initial abundance, not the time-dependent decline. A self-consistent calculation, or a bracketing estimate using the established flux-concentration proportionality together with a few hydrodynamic runs at reduced concentrations to verify that the re-accretion fraction is unchanged, is needed before the 1800-2000 K window and the 10^3-year stop time can be claimed.","section":"§2.3, §4 (Figs. 8-10), §6.3"},{"comment":"The adiabatic-index sensitivity statement is internally hard to reconcile. Section 6.3 reports that a 10% reduction in gamma increases the net loss flux by an order of magnitude, and that reducing the flux by the same amount requires a 60% increase in gamma, yet the same paragraph concludes that within 0.9gamma-1.6gamma the volatile loss timescale remains of the same order of magnitude. A factor-10 flux change means a factor-10 change in the derived loss timescale, so the quoted bracket spans roughly two orders of magnitude in the stop time, not one. Because the vertically integrated 2D treatment should use a lower effective gamma than the 3D VAPOROCK value (as the authors themselves note), the high-flux end of this bracket is the physically relevant direction, and it would shorten the required stop time and shift the matching temperature relative to the published values. Please report the flux-versus-gamma measurements and recompute the Fig. 9-10 matching at the bounding values; the assertion that the conclusions are unaffected is not supported by the quoted sensitivity.","section":"§6.3 (adiabatic-index sensitivity)"},{"comment":"The comparison between the hydrodynamic constraint and the thermal model is made with the time-weighted potential temperature (bar-T_p = 1844-1991 K, Section 5), but the vapor source in the hydrodynamic model is controlled by the surface temperature T_M through the Table 1 vapor pressures. In the thermal model these temperatures differ: Eq. (A.6) balances the convective flux F(T_p - T_surf) against the radiative flux sigma(T_surf^4 - T_eq^4), so T_surf < T_p, and Fig. 11a plots the two separately. If the early surface temperature is systematically below T_p by a few hundred kelvin, the effective temperature of the degassing surface at the time of lid formation may fall below the 1800-2000 K window, and the claimed correspondence would not hold. Please report T_surf(t) explicitly and justify the equivalence between the constant T_M of the hydrodynamic runs and the time-dependent thermal model by comparing the window with a time-weighted surface temperature (or by quantifying the offset introduced by using bar-T_p).","section":"§5, Appendix A (Eqs. A.5-A.7)"},{"comment":"The text in Section 5 argues that because about 50 atmospheric recycle cycles occur per magma-ocean overturn, the surface remains equilibrated with the ocean, and this is used to justify the constant-composition source. This argument establishes surface-atmosphere equilibrium, but the loss-integration in Section 4 removes material from the whole ocean inventory over 10^3-10^4 years, so the relevant quantity is the ocean-wide depletion, not the surface equilibration. The paper should either (a) demonstrate that the magma ocean interior is replenished on a timescale comparable to the loss timescale (which the 20-day overturn does not provide against 10^3-year integration), or (b) treat the ocean-composition decline explicitly; without this, the linear-depletion curves in Fig. 8 overestimate the late-time loss rate. This is the same load-bearing point as Major Comment 1, but the Section 5 formulation invites a misreading as a resolution of the feedback problem, so it should be clarified on its own.","section":"§5 (convective replenishment argument)"}],"minor_comments":[{"comment":"Please check the placement of T_M in Eq. (10): combining Eqs. (6)-(7) gives Sigma_M = P_2D / [c_v (gamma-1) T_M], so T_M should appear in the denominator; the current typesetting is ambiguous.","section":"Eq. (10)"},{"comment":"In Table A.2, the entries labeled T_liq = 1400 + 149.5 p and T_sol = 1977 + 64.1 p give T_sol > T_liq at p = 0, and the mu_l and mu_s entries (10^21 described as 'solid phase viscosity' and 0.1 as 'liquid phase viscosity') are correspondingly mislabeled; the text around Fig. A.12 (crystallization beginning at T_p = 1977 K) is consistent only if the 1977 + 64.1 p curve is the liquidus. Please relabel the entries so that subscripts match physical phases.","section":"Table A.2"},{"comment":"The quantitative match is to the elemental Na and K abundances of Visscher and Fegley (2013); the paper defers the K and Zn isotopic constraints without a calculation. An explicit sentence in the conclusions stating that the reported match applies to elemental abundances, and that isotopic fractionation in the hydrodynamic escape remains an open question, would prevent readers from inferring that the model also reproduces the isotopic enrichments cited as motivation in the Introduction.","section":"§6.1, §7"},{"comment":"The keyword list ('Moon, surface, Satellites, composition, Satellites, surfaces') contains a duplicated 'Satellites' and appears to be a formatting artifact; please use standard index terms without repetition.","section":"Keywords"},{"comment":"The time-weighted average potential temperatures are quoted to 0.01 K (1990.77 K and 1844.11 K); given the 200 K grid spacing of the hydrodynamic runs and the simplicity of the thermal model, one or two significant figures would better reflect the actual precision.","section":"§5"},{"comment":"The abstract's 'less than 30% of material being re-accreted' is valid only for aM less than or similar to 3.5 RE (Fig. 6b); adding the distance qualifier would improve precision.","section":"Abstract"},{"comment":"The sentence 'for aM = 9 RE, only about 1% of the ejected material does not return to the satellite' is easily misread as '1% is re-accreted'; please rephrase to state that about 99% of the ejected material is re-accreted at that distance.","section":"§3.1"},{"comment":"The caveat that the 2D approximation becomes marginal at large aM (disk scale height exceeding 2 lunar radii at 8 RE) is acknowledged; a single 3D test run at a distant orbit would provide a quantitative bound on the re-accretion error rather than a purely qualitative caveat.","section":"§6.3"}],"recommendation":"major_revision","confidential_remarks":"This is a well-executed follow-up to Charnoz et al. (2021) and fits EPSL's scope, bridging hydrodynamics and lunar geochemistry. My main concern is that the headline quantitative result (1800-2000 K window, stop time of order 10^3 years) may be quoted elsewhere before it is tested against depletion feedback and gamma calibration; I would not accept before those two points are addressed. The leading/trailing dichotomy prediction is the most durable contribution and is likely robust to the source-strength simplification. The authors' self-citation pattern is standard for a continuing research line and raises no concern for me."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a real step forward from Charnoz et al. (2021): the first 2D time-dependent hydrodynamic treatment with a resolved lunar surface, and it produces genuinely new results — reaccretion fractions below 30% inside ~3.5 R_E, a much smoother distance falloff of net loss flux, and a predicted leading/trailing reaccretion asymmetry beyond ~3.5 R_E. The numerical work looks solid: resolution and domain tests show converged fluxes below 10%, and the authors are unusually candid about what the model omits. I trust the hydrodynamic results themselves.\n\nThe soft spot is the one the stress-test flags, and the authors flag it too in Section 2.3: the vapor source is held at constant composition and temperature over the 1-year runs. That is fine for computing flow topology and reaccretion fractions. It is not fine for the headline claim — the 1800–2000 K window and ~10^3 yr stop time — because ocean degassing should lower the Na and K partial pressures as loss proceeds. The convective overturn argument in Section 5 only keeps the surface equilibrated with the whole-ocean inventory; it does not replenish that inventory. Over 10^2–10^4 yr, a constant source overestimates early loss and underestimates the required stop time. The paper's own sensitivity test in Section 6.3 shows the loss flux scales with volatile content, so time-dependent depletion could shift the inferred temperature window. This is an acknowledged simplification, not an oversight, but it is load-bearing.\n\nTwo smaller caveats. The leading/trailing dichotomy appears at distances where the 2D approximation is weakest — the disk scale height exceeds two Moon radii at 8 R_E, and the authors note meridional flows could raise reaccretion. And the rebuttal of Dauphas et al. (2022) is plausible but brief; the domain-of-validity argument against Hertz–Knudsen evaporation is fair, but the isotopic fractionation side is not fully engaged.\n\nFor lunar geochemists and early-Moon dynamicists, the flow calculations and the dichotomy prediction are worth taking seriously; the geochemical matching is conditional pending a self-consistent treatment of ocean depletion. It deserves a serious referee, and I would want the depletion feedback addressed before accepting the quantitative conclusion.","headline":"A solid, transparent upgrade to the tidally assisted escape scenario with new predictions, but the headline 1800–2000 K / ~10^3 yr matching is computed with a constant-composition vapor source and should be treated as conditional until depletion feedback is included.","tokens_in":28061,"tokens_out":3350,"would_cite":true,"duration_ms":27875,"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":"Tides stripped the young Moon of its sodium and potassium","keywords":["Moon","magma ocean","hydrodynamic escape","volatile depletion","sodium","potassium","tidal evolution","circumterrestrial disk"],"falsifier":"A single coupled calculation that lets the magma ocean's Na and K abundances decline and its surface temperature drop during escape would settle it: if the measured depletions take longer than about 1000 years or require temperatures outside 1800–2000 K, the central claim fails.","tokens_in":1701,"feed_emoji":"🌙","tokens_out":6252,"duration_ms":89347,"temperature":0.7,"pith_summary":"The paper argues that the Moon's depletion of sodium and potassium was caused by hydrodynamic escape from a magma ocean, aided by Earth's tides, in a way that does not depend on a particular Moon-formation model. The simulations match the measured lunar Na and K depletions if the magma ocean surface was about 1800–2000 K and volatile loss stopped within about 1000 years when a stagnant lid or anorthite crust formed. It also predicts a leading/trailing asymmetry in volatile reaccretion that could be checked.","feed_headline":"Tides stripped the young Moon of its sodium and potassium","feed_subtitle":"2D simulations match lunar Na and K depletions for 1800-2000 K magma oceans, capped by a crust in about 1000 years.","key_machinery":"The central mechanism is tidally assisted hydrodynamic escape: near Earth, tides lower the energy needed for gas to leave the Moon's Roche lobe. The paper uses the FARGOCA 2D hydrodynamic code with the magma ocean as a gas source, a constant-Q tidal migration model, and a stagnant-lid thermal evolution model to turn surface temperature and orbital distance into net loss fluxes and a shutoff timescale.","core_discovery":"Using 2D time-dependent hydrodynamic simulations, the authors find that vapor released from a molten proto-Moon near the Roche limit forms a circum-Earth disk through spiral arms at L1 and L2, with less than 30% of the vapor reaccreted at distances up to 3.5 Earth radii. The net loss fluxes, combined with tidal migration, can reproduce lunar Na and K abundances for magma ocean temperatures of 1800–2000 K. Escape must be shut off by a conductive lid or anorthite crust within about 1000 years, otherwise the Moon would be fully depleted.","pith_inferences":["The same escape mechanism should apply to other moderately volatile elements like zinc and rubidium, though the paper only computes Na and K.","A coupled model with a depleting magma ocean could shift the inferred temperature window or stop time; this is an editorial extension.","The predicted dichotomy could be tested by remote sensing or sample analysis if it survived later reorientation and impact gardening."],"forward_implications":["Volatile depletion can be explained without invoking a specific giant-impact or disk scenario.","The magma ocean surface temperature is constrained to 1800–2000 K, consistent with lunar chromium isotope estimates.","A lid must form within about 1000 years to prevent complete loss of Na and K.","For Moon-Earth distances above about 3.5 Earth radii, volatiles are preferentially reaccreted on the trailing side, predicting a hemispheric dichotomy."],"supporting_citations":[{"why":"Proposed tidally assisted hydrodynamic escape and gave the 1D steady-state model this paper extends.","marker":"Charnoz et al. (2021)"},{"why":"Supplies the VAPOROCK code used for vapor composition and Na/K partial pressures.","marker":"Wolf et al. (2023)"},{"why":"Provides the measured lunar Na and K abundance ranges used as the target.","marker":"Visscher and Fegley (2013)"},{"why":"Gives the 10^3-year anorthite crust timescale the paper uses for lid shutoff.","marker":"Elkins-Tanton et al. (2011)"},{"why":"The alternative kinetic-evaporation model the paper argues against.","marker":"Dauphas et al. (2022)"},{"why":"Supports magma ocean temperatures around 1600-1800 K from chromium isotopes.","marker":"Sossi et al. (2018)"}],"fun_headline_variants":["Tidal escape explains Moon's sodium and potassium loss","2D simulations show tidal winds devolatilized the proto-Moon","Tidal escape stripped Moon's K and Na in about a millennium","Magma ocean temperature pinned by Moon's sodium and potassium loss","Proto-Moon degassing: tidal winds set K and Na depletions"],"cache_read_input_tokens":30080,"weakest_assumption_plain":"The key premise is that the magma ocean is an infinite reservoir: the simulations keep the vapor source density, temperature, and composition constant, so the loss flux does not decrease as Na and K are removed; a depleting and cooling magma ocean could escape more slowly and require a different stop time or temperature window.","fun_headline_variants_meta":{"raw":{"variants":["Tidal escape explains Moon's sodium and potassium loss","2D simulations show tidal winds devolatilized the proto-Moon","Tidal escape stripped Moon's K and Na in about a millennium","Magma ocean temperature pinned by Moon's sodium and potassium loss","Proto-Moon degassing: tidal winds set K and Na depletions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000729,"raw_usage":{"total_tokens":3307,"prompt_tokens":1028,"completion_tokens":2279,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":644,"completion_tokens_details":{"reasoning_tokens":2189}},"tokens_in":644,"tokens_out":2279,"duration_ms":16822,"temperature":1.0,"reasoning_tokens":2189,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T04:25:58.171166+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A single coupled calculation that lets the magma ocean's Na and K abundances decline and its surface temperature drop during escape would settle it: if the measured depletions take longer than about 1000 years or require temperatures outside 1800–2000 K, the central claim fails.","supporting_citations":[{"cited_title":", author Moynier, F","cited_arxiv_id":null,"evidence_quote":"Supports magma ocean temperatures around 1600-1800 K from chromium isotopes."}],"review_version":1}