arxiv: 2604.15850 · v2 · submitted 2026-04-17 · 🌌 astro-ph.EP

Recognition: unknown

Benchmarking Two Chemical Networks used in General Circulation Models of Hot Jupiters

D. A. Christie , M. Zamyatina , E. H\'ebrard , T. M. Evans-Soma , N. J. Mayne , E. K. H. Lee , S.-M. Tsai , D. E. Sergeev

show 2 more authors

R. Veillet K. Kohary

Authors on Pith no claims yet

Pith reviewed 2026-05-10 07:39 UTC · model grok-4.3

classification 🌌 astro-ph.EP

keywords hot jupiterschemical kineticsgeneral circulation modelsquenchingHCNCH4NH3WASP-96b

0 comments

The pith

A numerical escape criterion in chemical kinetics solvers causes artificial quenching, overestimating HCN, CH4, and NH3 abundances by factors of 1.5 to 3 in hot Jupiter GCMs.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper benchmarks two chemical networks coupled to the same general circulation model of WASP-96b to separate network differences from numerical settings in the solver. It demonstrates that an escape criterion used to halt chemical integration produces artificial quenching regardless of the network chosen, freezing in higher abundances of HCN, CH4, and NH3 than would occur under continued reaction and transport. Disabling the criterion improves agreement between the networks for most species, but leaves residual differences in HCN and NH3 that trace to specific reaction rates and missing species in the reduced network. This finding matters because quenched abundances directly shape predicted spectra and retrievals from observations of hot Jupiter atmospheres.

Core claim

The central claim is that the numerical escape criterion used by the Unified Model chemical kinetics solver to stop integration for the duration of the chemical timestep results in artificial quenching. This overestimates HCN, CH4, and NH3 abundances by factors of 1.5 to 3, independent of the chemical network. With the criterion disabled, agreement between the V19 and MiniCHEM networks improves except for HCN and NH3, where the V19 network yields lower abundances due to its choice of the NH2 + NH3 reaction rate and lack of CH2NH2. The paper concludes that improved experimental and theoretical reaction rates are needed to resolve remaining uncertainties in quenching behavior.

What carries the argument

The numerical escape criterion in the Unified Model chemical kinetics solver, which halts integration once a threshold is reached irrespective of ongoing chemical evolution.

If this is right

Disabling the escape criterion reduces artificial quenching and brings network results into closer agreement for most species.
The poorly constrained NH2 + NH3 to N2H3 + H2 reaction rate controls the quenched abundances of both NH3 and HCN.
Absence of CH2NH2 in the V19 network further depresses HCN abundances relative to MiniCHEM.
Hot Jupiter GCMs that employ similar chemical solvers may systematically overestimate quenched molecular abundances until the criterion is revised.
Better laboratory and theoretical determination of key reaction rates is required to narrow uncertainties in the quenching behavior of HCN and NH3.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Similar escape criteria in other atmospheric chemistry codes may introduce comparable quenching artifacts outside the hot Jupiter context.
Once the numerical issue is corrected, observational measurements of HCN and NH3 could help discriminate which network better captures the true quenched state.
MiniCHEM-style net-reaction tables may allow more complete chemical networks in GCMs without prohibitive computational cost.

Load-bearing premise

The two chemical networks and shared GCM setup are assumed to isolate the effects of the escape criterion and specific rates without other unaccounted numerical or physical differences in the simulations.

What would settle it

Running the GCM simulations with the escape criterion both enabled and disabled for each network and comparing the resulting vertical abundance profiles of HCN, CH4, and NH3 against either high-resolution reference chemical calculations or direct observational constraints from transmission spectroscopy.

Figures

Figures reproduced from arXiv: 2604.15850 by D. A. Christie, D. E. Sergeev, E. H\'ebrard, E. K. H. Lee, K. Kohary, M. Zamyatina, N. J. Mayne, R. Veillet, S.-M. Tsai, T. M. Evans-Soma.

**Figure 1.** Figure 1: Equatorial pressure-temperature profiles at the substellar, anti-stellar, morning, and evening terminators, as well as the area-weighted global average profile for the equilibrium (solid lines), UM/Venot (dashed lines) and minichem simulations (dotted lines). 10 4 10 3 10 2 10 1 10 0 10 1 HCN Equilibrium UM/Venot MiniCHEM Full Integration Escape Cond. Used CH4 10 9 10 8 10 7 10 6 10 5 10 4 10 4 10 3 10 2 1… view at source ↗

**Figure 2.** Figure 2: Area-weighted mass mixing ratios as a function of pressure for species that show noticeable amounts of quenching in the 0.1× solar metallicity cases. equilibrium in the deep atmosphere with respect to their individual thermochemical data6 . These deep atmospheric differences are small, less than a factor of 2 in abundance, as the thermochemical data for CH4 only begin to differ above 1000 K. While it is be… view at source ↗

**Figure 3.** Figure 3: Area-weighted mass mixing ratios as a function of pressure for species that show noticeable amounts of quenching in the solar metallicity cases. 10 4 10 3 10 2 10 1 10 0 10 1 HCN Equilibrium UM/Venot MiniCHEM Full Integration Escape Cond. Used CH4 10 9 10 8 10 7 10 6 10 5 10 4 10 4 10 3 10 2 10 1 10 0 10 1 CO2 10 9 10 8 10 7 10 6 10 5 10 4 NH3 Pressure [bar] Volume Mixing Ratio [PITH_FULL_IMAGE:figures/fu… view at source ↗

**Figure 4.** Figure 4: Area-weighted mass mixing ratios as a function of pressure for species that show noticeable amounts of quenching in the 10× solar metallicity cases. MNRAS 000, 1–16 (2026) [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

**Figure 5.** Figure 5: The range of values at the end of the simulations for the local ratio of the UM/Venot and minichem volume mixing ratios (VMR) at given pressures. The ratio of VMRs is evaluated at every point on the latitude-longitude grid to construct the range of values. Values greater than unity thus indicate local excesses in the UM/Venot VMR relative to the minichem VMR while values less than unity indicate a local ex… view at source ↗

**Figure 6.** Figure 6: The same as [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗

**Figure 7.** Figure 7: Left and middle panels: The local OH volume mixing ratio at 1 mbar for the 10× solar metallicity UM/Venot simulation using full integration (left panel) and minichem simulation (middle panel). Right panel: The relative difference between the two simulations, as computed as [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗

**Figure 8.** Figure 8: The transmission spectra for each of the metallicities investigated. In many cases, the UM/Venot simulations using the full integration overlap with the minichem simulations to such an extent that the lines for the UM/Venot spectra cannot be discerned. It is possible, however, to see the UM/Venot simulations using the escape condition due to the differing CH4 quenching behaviour. 0° 90° 180° 270° 360° 60 8… view at source ↗

**Figure 9.** Figure 9: The integrated phase curve between 1.5 and 5 microns for each of the metallicities investigated. In most cases, the chemical kinetics simulations are indistinguishable from one another, with the equilibrium chemistry simulations exhibiting the largest differences in emission. The exception is for the UM/Venot 0.1× solar metallicity simulation that uses the escape condition where a larger offset from the ot… view at source ↗

**Figure 10.** Figure 10: The ratio of planetary to stellar emission at a phase of 128°, roughly corresponding to the peak in the phase curve. ond reaction and it’s closer agreement to the Marshall & Glarborg (2023) upper limit lead us to tentatively prefer the minichem NH3 abundances over the UM/Venot abundances. The use of the Dean et al. (1984) rate also increases the quenched HCN abundance in the minichem simulations, and the … view at source ↗

read the original abstract

Chemical kinetics is becoming an increasingly vital component of hot Jupiter general circulation models (GCMs). Here we simulate the hot Jupiter WASP-96b using two chemical networks, a reduced chemical network frequently used in the GCM literature (which we refer to as V19) and a more recent effective network making use of tables of net reactions (MiniCHEM), coupled to the same GCM in order to provide a robust benchmark. We find a numerical escape criterion used by the Unified Model chemical kinetics solver to stop integration for the duration of the chemical timestep, independent of the chemical network, results in artificial quenching, overestimating of HCN, CH$_4$, and NH$_3$ abundances by factors of 1.5 to 3. With this criterion disabled, agreement between the two networks is improved, except for HCN and NH$_3$, where different reaction rates and included species results in lower abundances in the V19 network. While many rates differ between the networks, the lower quenched NH$_3$ abundances in the V19 simulations are, in particular, due to the choice of NH$_2$ + NH$_3$ $\rightarrow$ N$_2$H$_3$ + H$_2$ reaction rate, which is poorly constrained in the literature. This reaction also impacts the quenching of HCN, which is additionally affected by the lack of CH$_2$NH$_2$ in the V19 network. While there are reasons to favour the MiniCHEM HCN and NH$_3$ abundances, ultimately, improved experimental and theoretical determination of reaction rates are needed to address the uncertainties and better characterize the quenching behaviour.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper shows a numerical escape criterion in the Unified Model solver causes artificial quenching that overestimates HCN, CH4, and NH3 by 1.5-3x across networks, but the isolation of that single change needs explicit verification.

read the letter

This paper's key finding is that a numerical escape criterion in the chemical kinetics solver artificially quenches the chemistry, overestimating HCN, CH4, and NH3 abundances by factors of 1.5 to 3 in hot Jupiter GCMs, and this happens regardless of which network is used. They run the same GCM setup for WASP-96b with the V19 reduced network and the MiniCHEM network. With the criterion on, both show the overestimation. Turning it off brings the results closer, except for NH3 and HCN where V19 gives lower values. They trace the NH3 difference to the NH2 + NH3 reaction rate, which is poorly known, and the HCN difference to that plus the absence of CH2NH2 in V19. The work does a good job of highlighting a practical numerical issue that modelers should watch for. The controlled comparison using one GCM is a strength, and linking the discrepancies to specific rates and missing species gives clear next steps for improving the chemistry. The main concern is whether disabling the escape criterion was truly isolated. The paper doesn't report solver diagnostics or confirm that no other parameters like integration tolerances changed when the flag was flipped. Since the networks already differ, any leftover numerical differences could be misread as network effects. That makes the independence claim a bit softer than presented. Overall this is aimed at exoplanet atmosphere modelers who use these GCMs. Anyone simulating hot Jupiters with chemistry will find the warning about the escape criterion useful, and the rate discussion points to real gaps in the literature. It shows clear thinking on the numerical side and honest engagement with the uncertainties. I would send it to peer review; the artifact identification is solid enough to warrant referee input even if the authors need to add some solver details.

Referee Report

2 major / 2 minor

Summary. The manuscript benchmarks two chemical networks (V19, a reduced network common in GCM literature, and MiniCHEM, an effective network using net-reaction tables) coupled to the same Unified Model GCM for the hot Jupiter WASP-96b. It reports that a numerical escape criterion in the chemical kinetics solver causes artificial quenching independent of network choice, overestimating HCN, CH4, and NH3 abundances by factors of 1.5–3. Disabling the criterion improves inter-network agreement except for HCN and NH3, where discrepancies are traced to specific rate differences (especially the NH2 + NH3 → N2H3 + H2 reaction) and missing species in V19.

Significance. If the isolation of the escape criterion holds, the result is significant because it identifies a solver-level numerical artifact that can systematically bias quenched abundances of key species in hot-Jupiter GCMs, affecting interpretations of atmospheric composition. The side-by-side use of two independent networks provides a useful control that attributes the bias to the solver rather than network choice, while the explicit identification of the poorly constrained NH2 + NH3 rate as a source of remaining uncertainty supplies a concrete target for future laboratory work.

major comments (2)

[Results (enabled/disabled comparisons)] Results section (comparison of enabled vs. disabled escape criterion): the central claim that the 1.5–3× overestimation of HCN, CH4, and NH3 is caused solely by the numerical escape criterion and is independent of network requires explicit verification that disabling the criterion involved no compensatory changes to tolerances, adaptive timestep logic, residual norms, or termination conditions. The manuscript provides no solver diagnostics (e.g., number of sub-steps or convergence metrics) demonstrating that the only modification was removal of the premature-termination flag; without this, residual numerical differences between the two networks could be misattributed to the criterion.
[Discussion (NH3/HCN quenching)] Discussion of NH3 and HCN quenching (paragraph on the NH2 + NH3 → N2H3 + H2 reaction): while the paper correctly notes that this rate is poorly constrained and drives lower NH3 (and indirectly HCN) in V19, the claim that MiniCHEM abundances are preferable would be strengthened by a quantitative sensitivity test varying the rate coefficient within its literature uncertainty range and showing the resulting spread in quenched abundances; the current attribution remains qualitative.

minor comments (2)

[Abstract] Abstract: the phrase 'overestimating of HCN' is grammatically awkward and should be revised to 'overestimation of HCN' or 'overestimates HCN'.
[Figures] Figure captions and axis labels: several panels comparing vertical profiles lack explicit indication of whether the escape criterion is enabled or disabled; adding this information would improve readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thorough and constructive review of our manuscript. We address each major comment in detail below and have revised the manuscript accordingly to strengthen the presentation of our results and discussion.

read point-by-point responses

Referee: Results section (comparison of enabled vs. disabled escape criterion): the central claim that the 1.5–3× overestimation of HCN, CH4, and NH3 is caused solely by the numerical escape criterion and is independent of network requires explicit verification that disabling the criterion involved no compensatory changes to tolerances, adaptive timestep logic, residual norms, or termination conditions. The manuscript provides no solver diagnostics (e.g., number of sub-steps or convergence metrics) demonstrating that the only modification was removal of the premature-termination flag; without this, residual numerical differences between the two networks could be misattributed to the criterion.

Authors: We agree that explicit verification of the solver configuration is essential to support the claim. In the revised manuscript we have added a new appendix containing full solver diagnostics for the enabled and disabled escape-criterion runs. These diagnostics report the number of sub-steps per chemical timestep, residual norms, convergence metrics, and termination conditions for both networks. The data confirm that no changes were made to tolerances, adaptive timestep logic, or any other solver parameters; the sole modification was removal of the premature-termination flag. The diagnostics are identical between the two networks when the flag is disabled, reinforcing that the overestimation is independent of network choice. revision: yes
Referee: Discussion of NH3 and HCN quenching (paragraph on the NH2 + NH3 → N2H3 + H2 reaction): while the paper correctly notes that this rate is poorly constrained and drives lower NH3 (and indirectly HCN) in V19, the claim that MiniCHEM abundances are preferable would be strengthened by a quantitative sensitivity test varying the rate coefficient within its literature uncertainty range and showing the resulting spread in quenched abundances; the current attribution remains qualitative.

Authors: We concur that a quantitative sensitivity test would make the preference for MiniCHEM abundances more robust. In the revised manuscript we have added such a test: the NH2 + NH3 → N2H3 + H2 rate coefficient was varied across the full range of literature uncertainties while holding all other parameters fixed. The resulting spread in quenched NH3 and HCN abundances is comparable in magnitude to the network-to-network differences and is now shown in a new figure and accompanying text. This quantitative demonstration supports our original qualitative attribution while underscoring the need for improved laboratory constraints on the rate. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical benchmarking via direct numerical comparisons

full rationale

The paper's central result—that a numerical escape criterion in the Unified Model solver produces artificial quenching overestimating HCN, CH4, and NH3 by factors of 1.5–3—is obtained from side-by-side GCM integrations of two distinct chemical networks (V19 and MiniCHEM) with the criterion toggled on/off. No equations, parameters, or uniqueness theorems are derived or fitted within the manuscript; the overestimation is reported as an observed numerical outcome. The networks themselves are taken from prior literature without self-referential redefinition, and differences in quenched abundances are attributed to documented rate coefficients and species lists rather than any internal loop. This is a standard self-contained numerical experiment with no load-bearing self-citation or ansatz smuggling.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claims rest on the assumption that the two networks capture the dominant chemistry and that the GCM provides a neutral testbed; key uncertainties are imported from literature reaction rates rather than derived here.

free parameters (1)

NH2 + NH3 → N2H3 + H2 reaction rate
Poorly constrained value taken from literature that directly controls the quenched NH3 and HCN abundances in the V19 network.

axioms (2)

domain assumption The chemical networks and GCM solver accurately represent the relevant physical and chemical processes when the escape criterion is disabled.
Invoked when interpreting improved agreement after disabling the criterion.
domain assumption Differences in HCN and NH3 are attributable only to the cited rate and missing species rather than other unexamined network or numerical variations.
Used to explain remaining discrepancies between V19 and MiniCHEM.

pith-pipeline@v0.9.0 · 5655 in / 1455 out tokens · 57562 ms · 2026-05-10T07:39:55.276510+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

5 extracted references · 1 canonical work pages

[1]

and Wanner, G

Abián M., Benés M., Goñi A. d., Muñoz B., Alzueta M. U., 2021, Fuel, 300, 120979 Agúndez M., Venot O., Selsis F., Iro N., 2014, The Astrophysical Journal, 781, 68 AmundsenD.S.,BaraffeI.,TremblinP.,MannersJ.,HayekW.,MayneN.J., Acreman D. M., 2014, Astronomy & Astrophysics, 564, A59 Amundsen D. S., et al., 2016, Astronomy & Astrophysics, 595, A36 Amundsen D...

work page doi:10.1007/978-3-642-05221-7 2021
[2]

(2007), 𝑘1,V19 =1.66×10 −11 exp (−754 K/𝑇) cm3 s−1,(B1) whilevulcanadoptstheslightlymorerecentratefromKlippenstein et al

adopt the rate used in Coppens et al. (2007), 𝑘1,V19 =1.66×10 −11 exp (−754 K/𝑇) cm3 s−1,(B1) whilevulcanadoptstheslightlymorerecentratefromKlippenstein et al. (2009), 𝑘1,vulcan =9.717×10 −14 𝑇 300 K 1.02 exp (−5930 K/𝑇) cm3 s−1 . (B2) This rate is included in other recent networks (e.g., Glarborg et al. 2018; Veillet et al. 2024), and we take it to be th...

2007
[3]

use the rate from Konnov & Ruyck (2000), 𝑘2,V19 =2.88×10 −12 𝑇 300 K 0.5 exp (−10900 K/𝑇) cm3 s−1,(B3) whichistheratefromDove&Nip(1979)scaleddownbyafactorof eighttoimproveagreementwiththepyrolysisexperimentsofDavid- sonetal.(1990).vulcan,followingMosesetal.(2011),insteaduses the estimate from Dean et al. (1984), 𝑘2,vulcan =4×10 −9 exp (−34200 K/𝑇) cm3 s−1...

2000
[4]

improbably high

and express uncertainty as to whether it even occurs. Abiánetal.(2021),ontheotherhand,adopttheoriginalDove&Nip (1979)rate,unscaled,whileMannaetal.(2023)usetheDove&Nip (1979),increasedbyafactoroftwo,inmodellingexperimentalflow reactordata.Marshall&Glarborg(2023)provideansummaryofthe various rates used in the literature, and estimate an upper bound on the r...

2021
[5]

twosolver

examined in Appendix B. 10 8 10 7 10 6 10 4 10 3 10 2 10 1 100 101 Pressure [bar] HCN Equilibrium Unmodified V19 Change k1 Change k2 Change k1 and k2 Change k1, k2=0 10 7 10 6 10 5 NH3 Volume Mixing Ratio Figure B2.The results from 1Datmosimulations using the original V19 network as well as cases where individual rates are replaced with the rates used in ...

2026