{"id":"e72ee78a-cf8f-4ec8-80c7-7a49b4221d5f","arxiv_id":"2506.00091","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"AGNI is an open-source, energy-conserving radiative-convective model for lava planet atmospheres, built for coupling with interior evolution simulations.","lead":"AGNI is a new open-source computer model that calculates the temperature and radiation in the atmospheres of lava planets, rocky worlds with molten surfaces. It is designed to run fast enough to be coupled with simulations of how these planets cool and evolve over billions of years.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No in-paper evidence that AGNI's Newton-Raphson solver actually conserves total energy flux; the key claim rests entirely on a prior paper.","rationale":"The reader identified the absence of in-paper validation and the delegation of solver convergence to prior work as the weakest assumption. I agree, and I have sharpened the concern to the specific property that the central claim needs: conservation of total energy flux (not just radiative flux) at every level after Newton-Raphson convergence. This property is not demonstrated anywhere in the manuscript, and the treatment of the convective/condensing branch is not described, so the solver's behavior in the optically thick, steam-dominated regimes central to the paper's stated applications remains unverified. The concern does not amount to an accusation of incorrectness; it is a request for evidence of robustness and convergence. Because the reader already issued a CONDITIONAL verdict, and because this concern supports that verdict without being severe enough to demand rejection, I recommend keeping the verdict unchanged. A focused diagnostic test on total energy flux balance with and without condensation would settle whether the concern lands.","tokens_in":6266,"tokens_out":5759,"duration_ms":84660,"concrete_test":"Add a diagnostic to AGNI that, after convergence, writes the total energy flux F_tot(level) = F_rad(level) + F_conv(level) + F_lat(level) at every model level. Run the published pure-H2O tutorial case (1 bar, 10 levels) with and without condensation enabled, and assert that max|F_tot(i) - F_tot(j)| / |F_tot| is below the solver's stated tolerance. Then repeat for a steam-rich, optically thick case (e.g., 10 bar, T_surf = 2500 K). If the imbalance exceeds tolerance in either run, or if the solver fails to converge in the 10-bar case, the energy-conservation claim is not established for the intended domain.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that AGNI provides an energy-conserving numerical solution for the temperature and radiation environment of lava-planet atmospheres. For this claim to hold, the Newton-Raphson iteration must converge, to a stated tolerance, to a state in which the total energy flux (radiative plus convective plus latent-heat transport) is constant across all layers. The manuscript nowhere quantifies this tolerance, shows a residual or flux-balance diagnostic, or compares a converged profile against an independent solution. Instead the entire method is delegated to Nicholls, Pierrehumbert, et al. (2024), and the runtime figures are given without grid size or hardware context. The specific risk is that the solver's convergence domain may not cover the intended parameter space. In particular, the paper advertises convection, condensation, and sensible heat as accounted for, but gives no indication of how the Jacobian or line search treats the discontinuity between radiative equilibrium and a convective/condensing branch. In the optically thick, steam-rich regime relevant to magma-ocean atmospheres, a Newton-Raphson scheme can converge to a local minimum of the residual rather than to true flux balance, or can fail to converge at all, if the derivative of the convective/latent flux is not consistently included. Since the paper's headline applications (convective shutdown, coupled magma-ocean evolution) occur in exactly that regime, the unverified convergence and energy conservation are load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper describes AGNI, a Julia code for computing the temperature and radiation environment of rocky exoplanet/lava-planet atmospheres. AGNI uses the SOCRATES correlated-k and two-stream radiative transfer code, a Newton-Raphson solver to enforce energy conservation at each atmospheric level, and representations of convection, condensation, and sensible heat transport. It supports real-gas equations of state, can run standalone via a command-line interface or as a library, and is designed for coupling into the PROTEUS interior-atmosphere evolution framework. The paper is a JOSS-style software description; it advertises the code's capabilities, compares it with HELIOS, GENESIS, and Exo_k, and lists future developments, but it does not present new scientific results or in-paper validation experiments.","tokens_in":6502,"tokens_out":4149,"duration_ms":54586,"significance":"If AGNI performs as claimed, it fills a useful niche: an open-source, CPU-based radiative-convective model for lava-planet atmospheres that can be coupled to an interior evolution code, with a real-gas equation of state. The fact that AGNI has already been used in several published applications (Hammond et al. 2025; Nicholls, Pierrehumbert, et al. 2024; Nicholls, Guimond, et al. 2025; Nicholls, Lichtenberg, et al. 2025) is a concrete strength and indicates practical utility. However, the central technical claims in this manuscript — energy-conserving solutions, solver convergence, and runtime performance — are not supported by in-paper evidence; they are delegated to a prior paper. The significance of the contribution is therefore real but contingent on the validation appearing in the cited work.","major_comments":[{"comment":"The statement that AGNI 'uses the Newton-Raphson method to conserve energy fluxes through each level of the column to a required tolerence' is the central methodological claim of the paper, but no tolerance value, no residual or flux-balance diagnostic, and no comparison against an independent solution are provided in the manuscript. Because the paper does not demonstrate that the converged state actually has constant total energy flux (radiative plus convective plus latent-heat transport) across the column, the advertised 'energy-conserving numerical solution' is unverified here. Please add a quantitative convergence diagnostic (e.g., net flux as a function of pressure for a representative case, together with the convergence tolerance and the final residual) or cite the specific figure/table in Nicholls, Pierrehumbert, et al. 2024 where this is shown and quote the tolerance.","section":"Comparison with other codes"},{"comment":"The paper states that convection, condensation, and sensible heat transport are accounted for, but it gives no details on how the Newton-Raphson iteration treats the discontinuity between the radiative-equilibrium branch and the convective/condensing branch, nor how the Jacobian or line search handles the derivative of convective and latent-heat fluxes. In optically thick, steam-rich conditions relevant to magma-ocean atmospheres, a Newton-Raphson scheme can fail to converge or can converge to a local minimum of the residual rather than to true flux balance. Since the headline applications (convective shutdown, coupled magma-ocean evolution) occur in exactly that regime, the manuscript should describe the residual equations and solver safeguards, or point to the specific section of Nicholls, Pierrehumbert, et al. 2024 that demonstrates convergence in this regime.","section":"Comparison with other codes"},{"comment":"The runtime figures ('a typical runtime ... is 3 minutes', 'less than 1 minute' with a good guess, 'approximately 30 ms' per radiative transfer calculation) are given without specifying the grid size (number of atmospheric levels), the spectral configuration (number of k-coefficients or g-points), the convergence tolerance, or the hardware used. Since speed is a central advertised advantage and the basis for the claim that Gyr-scale coupled evolution is feasible, these numbers need context to be meaningful. Please state the configuration and hardware for each timing, and report the number of levels and spectral points.","section":"Comparison with other codes"}],"minor_comments":[{"comment":"Typos should be corrected: 'tolerence' -> 'tolerance', 'obtaind' -> 'obtained', 'were are' -> 'we are', 'quantites' -> 'quantities', 'publically' -> 'publicly'.","section":"Summary"},{"comment":"The sentence 'Convection, condensation, and sensible heat transport are also accounted for' is vague; please specify the parameterizations used (e.g., moist adiabat, mass-flux scheme, latent heat release) or cite the section of the prior methods paper that describes them.","section":"Comparison with other codes"},{"comment":"The comparative claim that 'No other models of lava planet atmospheres implement a real-gas equation of state' is a strong negative claim with no citation or detailed comparison. It should either be softened or supported by a survey of the relevant literature.","section":"Comparison with other codes"},{"comment":"The reference list formatting is inconsistent: the entry 'Seidler, Fabian L., Sossi, Paolo A., & Grimm, Simon L.' uses given-name-first ordering whereas other entries use surname-first ordering. Please standardize.","section":"References"},{"comment":"The 'Future developments' list implicitly acknowledges current limitations (no aerosols/hazes, no multi-column dynamics, no Ledoux criterion, no full-spectrum convection, no parallelization). Consider adding a short 'Limitations' statement clarifying which of these are missing in the current release and whether they affect any of the cited applications.","section":"Future developments"}],"recommendation":"major_revision","confidential_remarks":"This is a JOSS software paper, so the validation bar is different from a research article, but the explicit claim of energy conservation is load-bearing and needs either an in-paper benchmark or a precise pointer to the prior paper with the tolerance and residual diagnostics. The heavy reliance on self-citations is understandable in a software methods paper, but the novelty claims ('only open source code integrated into PROTEUS', 'only real-gas EOS') should be substantiated or softened. No concerns about scientific misconduct or inappropriate citation behavior beyond what is noted above."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"AGNI is a real, open-source Julia implementation of a radiative-convective model for lava-planet atmospheres, and the paper's main claims are about the software's existence, its integration with PROTEUS, and its method. The genuinely new pieces are the Newton-Raphson energy-balance solver, the real-gas equation of state, and the coupling into a broader interior-evolution framework. The code ships with a CLI, library interface, tutorials, and documentation; that is reproducible evidence and should be treated as the core artifact. The paper also compares itself clearly to HELIOS, GENESIS, and Exo_k, and the niche it fills is real.\n\nThe soft spots are exactly where the reader and stress-test note point. The paper asserts energy conservation 'to a required tolerance' but never states the tolerance, shows a residual, or plots a converged flux profile. The runtime numbers (3 minutes cold, <1 minute warm, 30 ms per radiative transfer) come without grid size, number of levels, or hardware. The uniqueness claims—'only open source code integrated into PROTEUS,' 'no other models implement a real-gas EOS'—are strong and only supported by the limited survey in the comparison section. The Newton-Raphson method, including how it handles the convective/condensing branch, is delegated to Nicholls, Pierrehumbert, et al. (2024). That is a legitimate division of labor in a JOSS paper, but it means the present paper alone does not demonstrate convergence properties. The stress-test worry about Newton-Raphson getting stuck in a local residual minimum in the optically thick, steam-rich regime is plausible but speculative; there is no evidence in this paper that it happens, and the authors list convective inhibition and full-spectrum convection as future work, so they are aware of the edges. I would not treat this as a load-bearing flaw for the software paper, provided the prior paper actually shows the diagnostics.\n\nBottom line: this is a useful software announcement for people modeling magma-ocean atmospheres, and the open-source artifact deserves a serious referee. The main requests should be a validation or benchmark figure, an explicit convergence tolerance, and hardware/grid context for the benchmark times. I would accept it for peer review, with a recommendation that those additions go in.","headline":"A real open-source tool that fills a niche; the paper is thin on in-paper validation, but the code and prior method paper carry the load.","tokens_in":7035,"tokens_out":2835,"would_cite":true,"duration_ms":35222,"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":"AGNI gives lava planets a fast open-source atmosphere model.","keywords":["lava planets","magma ocean atmospheres","radiative-convective model","exoplanet atmospheres","energy conservation","Newton-Raphson solver","real-gas equation of state","climate modeling"],"falsifier":"Take a benchmark lava-planet case, such as a pure-steam or sulfur-rich atmosphere, run AGNI to convergence, and compare the resulting temperature profile and emergent spectrum to a line-by-line radiative-convective calculation or to a well-tested independent model; if the profiles disagree by more than the stated energy-conservation tolerance, or if the Newton-Raphson solver fails to converge from a poor initial guess for some gas mixtures, the paper's central utility claim is weakened. A direct internal check would be to verify that the net energy flux residual is below tolerance at every level in a published AGNI output.","tokens_in":6052,"feed_emoji":"🌋","tokens_out":8061,"duration_ms":93508,"temperature":0.7,"pith_summary":"The paper presents AGNI, a Julia program that solves for the temperature and radiation environment in the atmospheres of rocky exoplanets, with lava worlds (molten-surface planets) as the main target. The authors' central claim is that AGNI is the only open-source code currently designed to couple with a comprehensive interior-atmosphere evolution framework, and the only lava-planet atmosphere model using a real-gas equation of state. Its Newton-Raphson solution method is argued to make energy-conserving solutions faster and more scalable than the relaxation or time-stepping approaches used by most other models. A sympathetic reading is that this fills a practical gap for simulating magma-ocean planets over geological timescales.","feed_headline":"AGNI computes lava-planet atmospheres in minutes","feed_subtitle":"Open-source code is the first lava-world atmosphere model with a real-gas equation of state.","key_machinery":"The central mechanism is the Newton-Raphson solver acting on the residuals between the net upward and downward energy fluxes at each atmospheric level: the solver adjusts layer temperatures until the residual at every level falls below the required tolerance, coupling the radiative transfer calculation to the atmospheric structure in an energy-conserving way. This is what distinguishes AGNI from most atmosphere models, which typically use relaxation toward equilibrium or explicit time-stepping. The radiative fluxes themselves rely on the correlated-k and two-stream approximations, and the gas mixture is described with a real-gas equation of state, which allows exploration of the diverse, poorly constrained outgassed compositions expected on lava worlds.","core_discovery":"AGNI is a numerical model that obtains a radiative-convective equilibrium temperature structure for a one-dimensional atmospheric column by requiring energy fluxes to balance at every level simultaneously. It uses a Newton-Raphson optimizer to solve directly for the layer temperatures that conserve energy to a set tolerance, rather than iterating with a fixed-point scheme. Radiative fluxes are computed under the correlated-k and two-stream approximations using a well-established Fortran radiative transfer code, with convection, condensation, and sensible heat transport also represented. The paper reports that a standalone run takes about three minutes from a poor initial guess and under one minute when seeded with a good guess from a coupled evolution simulation, making the code practical for exploring wide parameter spaces and for long-timescale interior-atmosphere coupling.","pith_inferences":["Because the paper shows no internal validation against observations, analytic solutions, or established codes, the most decisive test will come from independent intercomparisons, such as comparing AGNI's emission spectra to line-by-line radiative transfer for a SiO- or SO2-rich rock-vapour atmosphere.","The Newton-Raphson formulation may transfer beyond lava worlds: any exoplanet atmosphere simulation that can supply a good initial guess from a previous timestep could profit from the same direct-solve approach, which would speed up coupled evolution models for hydrogen-rich or carbon-rich planets as well.","The reported 30 ms per radiative transfer call suggests that AGNI could be paired with machine-learning emulators to explore parameter space even faster, though the paper does not discuss that combination.","If the real-gas equation of state is a major source of fidelity, earlier lava-world models that used simplified equations of state may have systematically misestimated magma-ocean cooling rates, which would be worth re-examining with the new code."],"forward_implications":["Lava-planet atmosphere studies can be run in minutes without proprietary software, enabling grids of models over the poorly constrained gas compositions and redox states of magma-ocean planets.","Coupling to interior evolution models becomes practical over Gyr timescales, since an atmosphere solve from a good initial guess takes under one minute.","The real-gas equation of state moves lava-world modeling beyond pure-steam or semi-arbitrary gas mixtures toward realistic rock-vapour compositions.","Open availability means the community can reproduce, extend, and compare the model, including the planned additions of hazes, multi-column dynamics, and Ledoux-based convection inhibition."],"supporting_citations":[{"why":"Supplies the flexible radiative transfer code that AGNI leverages to compute fluxes from a given temperature structure.","marker":"Edwards & Slingo, 1996"},{"why":"Provides the current form of the radiative transfer code used by AGNI within the correlated-k and two-stream approximations.","marker":"Sergeev et al., 2023"},{"why":"Establishes the treatment of overlapping gaseous absorption with the correlated-k method that AGNI's flux calculations depend on.","marker":"Amundsen et al., 2017"},{"why":"Previous work by the same authors that develops the Newton-Raphson energy-conservation method and convective shutdown behavior on which AGNI's solver rests.","marker":"Nicholls, Pierrehumbert, et al., 2024"},{"why":"A GPU-accelerated radiative-convective code that AGNI compares against, underpinning its portability and openness claims.","marker":"Malik et al., 2017"},{"why":"A closed-source lava-world atmosphere model applied to observable planets, which AGNI contrasts with for coupling and availability.","marker":"Piette et al., 2023"},{"why":"An open-source pure-steam model that identified convective stability, which AGNI extends to mixed gases and coupling.","marker":"Selsis et al., 2023"},{"why":"A coupled magma-ocean protoatmosphere model with simplifying assumptions that AGNI aims to improve upon.","marker":"Lichtenberg et al., 2021"}],"fun_headline_variants":["AGNI simulates lava-planet atmospheres in minutes","Open-source AGNI models lava worlds in minutes","First lava-atmosphere model with real-gas EOS","AGNI speeds up radiative-convective modeling of lava planets"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that the correlated-k and two-stream radiative transfer approximations, the Newton-Raphson solver, and the real-gas equation of state together yield accurate, converged, energy-conserving solutions for the poorly constrained gas mixtures of lava planet atmospheres, without demonstrating this through comparisons to observations, analytic solutions, or established codes in the paper itself.","fun_headline_variants_meta":{"raw":{"variants":["AGNI simulates lava-planet atmospheres in minutes","Open-source AGNI models lava worlds in minutes","First lava-atmosphere model with real-gas EOS","AGNI speeds up radiative-convective modeling of lava planets"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000214,"raw_usage":{"total_tokens":1433,"prompt_tokens":960,"completion_tokens":473,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":576,"completion_tokens_details":{"reasoning_tokens":408}},"tokens_in":576,"tokens_out":473,"duration_ms":5800,"temperature":1.0,"reasoning_tokens":408,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:24:27.841516+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a benchmark lava-planet case, such as a pure-steam or sulfur-rich atmosphere, run AGNI to convergence, and compare the resulting temperature profile and emergent spectrum to a line-by-line radiative-convective calculation or to a well-tested independent model; if the profiles disagree by more than the stated energy-conservation tolerance, or if the Newton-Raphson solver fails to converge from a poor initial guess for some gas mixtures, the paper's central utility claim is weakened. A direct internal check would be to verify that the net energy flux residual is below tolerance at every level in a published AGNI output.","supporting_citations":[],"review_version":1}