{"id":"2693e899-360e-4bb1-a205-ac4f1b9e881e","arxiv_id":"2608.13157","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Varying assumed metallicity, C/O, S/N, and 1-bar temperature in thermochemical equilibrium models changes predicted mixing ratios and cloud decks of Uranus and Neptune by more than an order of magnitude, and thermal profiles by tens of kelvins.","lead":"This paper uses chemical equilibrium models to map how cloud levels and gas amounts on Uranus and Neptune change when the assumed composition and temperature are varied. It shows that plausible assumptions produce very different atmospheres, which matters for planning future observations of these planets.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Thermal profiles use a fixed deep H2O abundance q_int=0.25 regardless of metallicity, while the model's own deep H2O MMR ranges 0.005–0.238; the claimed composition-driven temperature differences are not actually implemented, undermining the reported cloud-deck ranges.","rationale":"The reader's weakest_assumption identifies the single-species moist adiabat with fixed q_int=0.25 as the load-bearing premise; my analysis agrees and sharpens it. The key problem is not just physical realism (other condensibles, cross terms) but an internal inconsistency: the thermal profile construction assumes a deep H2O abundance that differs by up to a factor of 50 from the model's own equilibrium prediction. Moreover, the model as described does not feed composition back into the thermal profile, contradicting the abstract's implication that composition drives the reported tens-of-kelvin differences. This does not invalidate the paper's qualitative conclusion that model assumptions matter, but it means the specific quantitative ranges and the stated cause of thermal-profile differences are not yet supported. The reader's CONDITIONAL verdict remains appropriate: the authors should provide the model outputs and rerun with consistent thermal profiles before the numbers are used as a reference framework.","tokens_in":18468,"tokens_out":7598,"duration_ms":70290,"concrete_test":"For the extreme metallicity cases (Z=1 and Z=80 Z⊙), recompute the thermal profile using the model's own self-consistent deep H2O mixing ratio as q_int (≈5×10−3 for 1 Z⊙, ≈0.238 for 80 Z⊙) instead of the fixed 0.25, and also include the heavy-element contribution to μ and c_p. Rerun the FastChem grid with these new P-T profiles and compare the resulting H2O, CH4, and H2S cloud deck pressures and deep MMRs against Table 2. If the cloud pressure ranges shift by more than a factor of ~2, or if the CH4 cloud appears/disappears qualitatively, the central sensitivity claim is not robust to the thermal-profile assumption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The sensitivity results depend on the P-T profiles input to FastChem, which are generated (Sec. 2.4, Eqs. 3–4, Appendix A) as single-species H2O moist/dry adiabats with a prescribed deep water abundance q_int=0.25 (Table A1). The dry phase is fixed to 85% H2 + 15% He, and μ and c_p include only H2O vapor. Therefore, for a fixed T1bar, the thermal profile is independent of metallicity, C/O, and S/N. Yet the FastChem equilibrium models produce deep H2O MMRs from 5.0×10−3 at 1 Z⊙ to 2.38×10−1 at 80 Z⊙ (Table 2). For low metallicity, the assumed deep water abundance is 50× larger than the model's own value, so the moist adiabat stays moist to pressures far deeper than consistent with the actual water abundance, shifting deep temperatures and all condensation levels. The abstract's claim that 'thermal profiles can differ by several tens of kelvins due to composition' is not implemented by the described model setup: only T1bar changes the profile, not the varied elemental abundances. The reported order-of-magnitude ranges for cloud deck altitudes are therefore partly artifacts of an inconsistent q_int rather than a robust response to composition.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses the FastChem chemical-equilibrium code to construct one-dimensional atmospheric models of Uranus and Neptune, varying metallicity (1–80 solar), C/O ratio (0.1–2.0), S/N ratio (0.19–1.6), and 1-bar temperature (66–86 K). It reports the resulting vertical mass mixing ratios and cloud decks of CH4, NH3, H2S, H2O, and NH4SH, and states that mixing ratios and cloud-deck altitudes vary by more than an order of magnitude while thermal profiles differ by tens of kelvins. The results are compared with the Hueso et al. (2020) model and with retrievals of H2S and CH4 abundances. The paper explicitly lists its equilibrium, single-vapor, rainout, and no-microphysics assumptions.","tokens_in":18795,"tokens_out":5673,"duration_ms":52995,"significance":"If the quantitative results were robust, the paper would provide a useful sensitivity envelope for ice-giant atmospheric structure that can be compared with ground-based and JWST observations and with future probe measurements. The use of an open-source equilibrium code, the broad parameter sweep, and the external comparisons against Hueso et al. (2020) and observed H2S/CH4 abundances are strengths. However, the central composition-dependence claim is weakened by an internal inconsistency in how the thermal profiles are constructed, so the quantitative cloud ranges and the claimed tens-of-kelvin composition effect need to be re-established.","major_comments":[{"comment":"The thermal profiles are constructed with a fixed deep H2O abundance q_int = 0.25, while the FastChem equilibrium models in Table 2 have deep H2O mass mixing ratios from 5.0e-3 (1 Zsun) to 2.38e-1 (80 Zsun). For the low-metallicity end, the prescribed deep water abundance is about a factor of 50 larger than the model's own deep abundance, so the moist adiabat remains active to pressures where the equilibrated atmosphere is actually dry; for the high-metallicity end the two values are closer. Because the same q_int is used for all metallicity, C/O, and S/N cases, the thermal profile is effectively independent of elemental composition except through the radiative-gradient term used in the convection-inhibition case of Sec. 4.4. The abstract's statement that thermal profiles differ by several tens of kelvins 'due to composition' is therefore not implemented by the model as described. The authors should iterate q_int until it matches the deep H2O MMR for each case, or explicitly restrict the composition-driven temperature claim to the convection-inhibition scenario.","section":"Sec. 2.4 and Appendix A, Eq. (4), Table A1"},{"comment":"The reported cloud-deck altitude ranges for metallicity, C/O, and S/N are computed on thermal profiles tied to q_int = 0.25 rather than to the actual equilibrium water abundance of the modelled atmosphere. Since the condensation levels of H2O, NH4SH, and, to a lesser extent, H2S respond to the deep temperature profile, the order-of-magnitude cloud-altitude variations in Table 2 are not a clean response to the varied elemental abundances. The central sensitivity conclusion must be re-evaluated with self-consistent thermal profiles, or the affected quantities must be shown to be insensitive to q_int over its plausible range.","section":"Sec. 3, Figs. 1-4, Table 2"}],"minor_comments":[{"comment":"The phrase 'due to composition' in the abstract should be reworded or removed until the thermal profiles are actually made composition-dependent, as discussed in Major Comment 1.","section":"Abstract"},{"comment":"The caption implies that each metallicity produces its own thermal profile, but with q_int fixed the orange profiles overplot; please state this explicitly and label the single profile when that is the case.","section":"Fig. 1 caption"},{"comment":"All cloud ranges use the arbitrary 10^-4 g/l cloud-top threshold; since the paper explicitly identifies this convention, a short sensitivity test at 10^-5 and 10^-3 g/l would let the reader judge how much of the reported order-of-magnitude variation comes from the threshold choice.","section":"Sec. 3, Table 2"},{"comment":"The phrase 'characterizationsteptoward' appears to be a missing space, and the manuscript should be proofread for similar spacing errors in the typeset text.","section":"Sec. 5"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the authors have been transparent about many limitations. The load-bearing issue is the fixed q_int=0.25 in the thermal profile construction, which makes the claimed composition-driven temperature differences and the associated cloud ranges internally inconsistent. This is fixable in a major revision, so I do not recommend rejection, but the central tables and figures need to be recomputed or explicitly defended against this issue."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis is a useful parameter scan, not a discovery, and the best way to read it is as a convenience reference for ice-giant modelers. The authors run FastChem over a plausible grid of metallicity, C/O, S/N, and 1-bar temperature and tabulate the resulting deep mixing ratios and cloud pressure ranges for CH4, NH3, H2S, H2O, and NH4SH. The comparison with Hueso et al. (2020) is pointed and fair, the limitations (equilibrium chemistry, single-species rainout, no microphysics) are mostly stated up front, and the paper does not oversell itself as a measurement of the real planets.\n\nThe soft spot is real and it is load-bearing. The thermal profiles are computed with a fixed deep H2O mass fraction q_int=0.25, independent of the metallicity used in the chemistry. But the FastChem models give deep H2O mixing ratios from about 0.005 at 1 Z_sun to 0.24 at 80 Z_sun. At low Z, the moist adiabat is therefore integrated with a water reservoir fifty times larger than the model's own water abundance. The profile stays moist far deeper than the actual water content justifies, which shifts the deep P-T structure and all condensation levels tied to it. In the main calculations the thermal profile does not actually respond to composition; only T1bar changes it. So the abstract's claim that thermal profiles can differ by tens of kelvins 'due to composition and 1-bar temperature' is not implemented as written. The composition part of that claim is unsupported. The cloud-deck ranges in Table 2 are partly artifacts of this inconsistency. This does not destroy the qualitative conclusion that equilibrium cloud structure is highly sensitive to assumed abundances, but the quantitative reference map needs to be recomputed with q_int consistent with the deep water abundance, or with a multi-species moist adiabat, before I would trust the specific numbers.\n\nMinor but worth fixing: the data are only available 'on reasonable request,' which is weak for a paper offering a reference framework; and the extreme-case bullets in Section 3 label the heavy-element-rich runs as '1 Z_sun' in two places, which looks like a copy-paste error.\n\nMy bottom line: this deserves peer review, not a desk rejection, and after revision it could be a citable reference for retrieval and mission-planning groups. I would not cite the current numbers. The honest read is that the paper is a sensitivity study with one internal inconsistency that partially feeds its main quantitative claims.","headline":"A useful equilibrium sensitivity map for ice giants, but the fixed q_int=0.25 decouples the thermal profiles from composition, so the headline temperature claim is not implemented and the cloud-deck numbers need rechecking.","tokens_in":19312,"tokens_out":5443,"would_cite":false,"duration_ms":50044,"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":"Plausible changes in metallicity, element ratios, and 1-bar temperature shift Uranus and Neptune cloud decks by more than an order of magnitude and thermal profiles by tens of kelvins.","keywords":["planets and satellites: atmospheres","planets and satellites: composition","planets and satellites: gaseous planets","planets and satellites: Uranus","planets and satellites: Neptune","thermochemical equilibrium","cloud condensation","moist adiabat"],"falsifier":"Recompute the full parameter grid with a multi-species moist adiabat that lets CH4, NH3, H2S, and H2O condense simultaneously, including the cross terms cited in the paper; if the predicted cloud deck altitudes and deep mixing ratios then vary by less than an order of magnitude across the same metallicity, C/O, S/N, and 1-bar temperature ranges, the paper's central claim of order-of-magnitude sensitivity would fail. A complementary check: an in-situ probe measuring the water abundance and the temperature profile down to 50 bar would show whether the H2O-only moist adiabat with $q_{int}=0.25$ matches the real lapse rate.","tokens_in":18290,"feed_emoji":"🪐","tokens_out":17406,"duration_ms":136017,"temperature":0.7,"pith_summary":"Uranus and Neptune are observed far less than Jupiter and Saturn, so atmospheric compositions and temperature-pressure profiles are usually inferred from thermochemical equilibrium models rather than measured directly. This paper tests how much those inferred structures change when the assumptions built into the models are moved across plausible ranges: metallicity from 1 to 80 times solar, C/O from 0.1 to 2.0, S/N from 0.19 to 1.6, and 1-bar temperature from 66 to 86 K. Using equilibrium chemistry with rainout condensation, the authors find that the predicted mixing ratios and cloud deck altitudes of CH4, NH3, H2S, H2O, and NH4SH can vary by more than an order of magnitude, and that the thermal profiles can differ by several tens of kelvins. The result matters because it shows that current equilibrium models do not single out one atmospheric structure for either planet; composition, thermal profile, and cloud formation must be treated as jointly degenerate until better observations or a dedicated mission break the degeneracy.","feed_headline":"Model assumptions shift ice giant cloud decks by an order of magnitude","feed_subtitle":"Predicted cloud deck altitudes swing by over an order of magnitude; thermal profiles shift by tens of kelvins.","key_machinery":"The argument runs on two coupled pieces. The first is FastChem, a chemical equilibrium code that solves the mass-action law and element conservation equations for a gas of roughly 500 species and iteratively selects the set of stable condensates, with a rainout approximation that removes condensed material from the overlying atmosphere and thus sets the cloud deck structure. The second is the single-vapor moist adiabat of Leconte et al. (2017), in which the lapse rate is dry where water is undersaturated and moist where water condenses, with a fixed deep water mixing ratio $q_{int}=0.25$; when the water mixing ratio reaches a critical value, convection is inhibited and the profile switches to a radiative gradient. The grid of metallicity, C/O, S/N, and 1-bar temperature inputs is scanned through this machinery, and the output mixing ratios, cloud decks, and thermal profiles are compared.","core_discovery":"The paper's central claim is that the vertical structure of Uranus and Neptune is strongly non-unique under thermochemical equilibrium with currently plausible inputs. Over the stated grid, the deep mixing ratio of CH4 spans $3.16\\times10^{-3}$ to $1.51\\times10^{-1}$, H2S spans $3.23\\times10^{-4}$ to $1.54\\times10^{-2}$, and H2O spans $5.00\\times10^{-3}$ to $2.38\\times10^{-1}$, including a factor-of-about-50 increase from the lowest to highest metallicity, and the altitude of cloud decks shifts by more than an order of magnitude: water cloud bases move from 510 to 180 bar, methane clouds appear between 0.2 and 0.6 bar only at high metallicity or low temperature, and the S/N ratio decides whether NH3 or H2S is the condensing species. Thermal profiles built on the water moist adiabat differ by several tens of kelvins as composition and 1-bar temperature vary. The authors conclude that the data do not yet select a single atmospheric state, and that the possible range extends from cold, heavy-element-rich, cloudy atmospheres to warmer, heavy-element-poor, nearly cloud-free ones.","pith_inferences":["If the same equilibrium sensitivity carries over to ice-giant exoplanets, then using a single observed tracer, such as a methane feature, to infer a sub-Neptune's metallicity is degenerate: several (Z, C/O, S/N, T1bar) combinations can reproduce the same tracer with very different deep compositions.","The tens-of-kelvin spread in deep thermal profiles is large enough to alter interior boundary conditions; Uranus and Neptune evolution models that start from an atmospheric profile should therefore propagate this spread into thermal-cooling and radius estimates, which the paper does not do.","A direct test of the paper's single-vapor lapse-rate assumption: recompute the grid with a multi-species moist adiabat including the cross terms cited in the paper. If the cloud-level shifts shrink below an order of magnitude, then much of the claimed sensitivity comes from the H2O-only thermal-profile approximation rather than from the composition grid.","The models assume a fixed factor-of-ten nitrogen depletion; if Uranus or Neptune instead retains nitrogen near solar proportions, NH3 and NH4SH cloud levels would shift, which microwave observations at tens of bars could check."],"forward_implications":["Inferred deep abundances of CH4, H2O, and H2S cannot be read off a single equilibrium model; the same observational constraints can be matched by different metallicity, C/O, S/N, and T1bar combinations, so bulk elemental ratios derived from equilibrium fits carry the model's parameter choice as an error bar.","High metallicity alone, from about 40 solar, puts a methane cloud above 1 bar, higher than the 1-2 bar methane cloud in earlier models, meaning a detected methane cloud altitude can discriminate between metallicity regimes even with the same species mix.","The S/N ratio is the switch for the upper cloud: with S/N below about 0.8, ammonia condenses between 3.6 and 8 bar, while with S/N above unity hydrogen sulfide condenses between 1.6 and 4 bar, matching the H2S detections reported for both planets.","Because only cold models, with T1bar around 66-70 K at 30 solar, produce significant methane condensation, observing a persistent methane cloud at 1-2 bar would require either a colder or more metal-rich atmosphere than the warm end of the grid, or an additional non-equilibrium process.","The authors conclude that distinguishing among these structures needs observations that can probe below the upper cloud deck, such as microwave mapping or an in-situ probe, since remote visible and near-infrared soundings stop within a few bars."],"supporting_citations":[{"why":"Describes the FastChem equilibrium solver that computes the gas-phase composition for the paper's model grid.","marker":"Stock et al. (2018)"},{"why":"Adds condensation and rainout to FastChem, which sets the cloud deck altitudes and depletes the gas phase.","marker":"Kitzmann et al. (2024)"},{"why":"Provides the moist-adiabat formulation used to build thermal profiles from the 1-bar temperature.","marker":"Leconte et al. (2017)"},{"why":"Shows the cross terms in the moist adiabat when several vapors condense, the limitation the paper acknowledges for its H2O-only lapse rate.","marker":"Li et al. (2018)"},{"why":"The equilibrium model the paper compares against for cloud bases, mixing ratios, and a shared deep water abundance value.","marker":"Hueso et al. (2020)"},{"why":"Supplies the solar abundances used to scale metallicity and element ratios in the models.","marker":"Asplund et al. (2009)"},{"why":"Observational detection of H2S in Uranus' upper atmosphere that motivates the high-S/N regime and the H2S cloud comparison.","marker":"Irwin et al. (2018)"},{"why":"Extends the H2S detection and abundance constraints to Neptune, anchoring the model comparison to observations.","marker":"Irwin et al. (2019)"}],"fun_headline_variants":["Ice giant cloud decks swing by 10x with model inputs","Uranus and Neptune structures shift tenfold with assumptions","Cloud altitudes vary by order of magnitude in ice giant models","Model assumptions alter ice giant cloud bases by 10x","Thermal profiles shift by tens of kelvins in ice giant models"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the whole temperature-pressure profile can be generated from a single condensing vapor, water, with a fixed deep water mixing ratio $q_{int}=0.25$; if methane, ammonia, or hydrogen sulfide condensing at other levels changes the lapse rate, the predicted cloud decks would shift.","fun_headline_variants_meta":{"raw":{"variants":["Ice giant cloud decks swing by 10x with model inputs","Uranus and Neptune structures shift tenfold with assumptions","Cloud altitudes vary by order of magnitude in ice giant models","Model assumptions alter ice giant cloud bases by 10x","Thermal profiles shift by tens of kelvins in ice giant models"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000267,"raw_usage":{"total_tokens":1679,"prompt_tokens":1074,"completion_tokens":605,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":690,"completion_tokens_details":{"reasoning_tokens":520}},"tokens_in":690,"tokens_out":605,"duration_ms":5712,"temperature":1.0,"reasoning_tokens":520,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:31:29.990711+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the full parameter grid with a multi-species moist adiabat that lets CH4, NH3, H2S, and H2O condense simultaneously, including the cross terms cited in the paper; if the predicted cloud deck altitudes and deep mixing ratios then vary by less than an order of magnitude across the same metallicity, C/O, S/N, and 1-bar temperature ranges, the paper's central claim of order-of-magnitude sensitivity would fail. A complementary check: an in-situ probe measuring the water abundance and the temperature profile down to 50 bar would show whether the H2O-only moist adiabat with $q_{int}=0.25$ matches the real lapse rate.","supporting_citations":[],"review_version":1}