---
title: 'Sonora: Substellar Atmosphere & Evolution Models'
url: https://www.emergentmind.com/topics/sonora
type: topic
---

# Sonora: Substellar Atmosphere & Evolution Models

Sonora is a family of self-consistent atmosphere and evolution models for substellar objects—brown dwarfs and giant exoplanets—developed to predict temperature–pressure structures, chemical compositions, spectra, photometry, and cooling tracks across the late-M, L, T, and Y regimes. Within the family, different branches target distinct physical regimes: cloudless atmospheres in rainout chemical equilibrium, cloudless atmospheres with disequilibrium chemistry, cloudy L/T-transition atmospheres, high-temperature extensions for early evolution, and very-cool cloud-free extensions for cold giant planets [2107.07434][2110.11824][2402.00756][2402.00758][2510.08694][2602.23423].

## 1. Scope and principal branches

Sonora is not a single grid but a suite of related grids sharing a common modeling philosophy while differing in cloud treatment, chemistry, and parameter coverage. The family structure is central to its use: cloud-free equilibrium models are commonly used as baselines, disequilibrium branches address vertical mixing and quenching, and cloudy branches address the refractory-cloud regime of warm brown dwarfs and young giant planets [2107.07434][2402.00756][2402.00758].

| Branch | Distinguishing treatment | Stated coverage |
|---|---|---|
| Bobcat | Cloudless, rainout chemical equilibrium | \(T_{\rm eff}=200\)–2400 K; \(2.5\le \log g \le 5.5\); \([{\rm M/H}] = -0.5\) to \(+0.5\); C/O \(=0.25\)–1.50 times solar [2107.07434] |
| Cholla | Cloud-free, solar-metallicity disequilibrium chemistry | \(T_{\rm eff}\sim 500\)–1300 K; \(\log g\sim 3.0\)–5.5; \(\log K_{zz}=2,4,7\) [2110.11824] |
| Elf Owl | Cloud-free disequilibrium chemistry with varying metallicity and C/O | \(T_{\rm eff}=275\)–2400 K; \(\log g=3.25\)–5.5; \([{\rm M/H}]=-1.0\) to \(+1.0\); C/O \(=0.22\)–1.14; \(\log K_{zz}=2\)–9 [2402.00756] |
| Diamondback | Cloudy equilibrium models with Ackerman–Marley clouds | \(T_{\rm eff}=900\)–2400 K; \(\log g=3.5\)–5.5; \([{\rm M/H}]=-0.5,0.0,+0.5\); \(f_{\rm sed}=1,2,3,4,8\) plus cloud-free [2402.00758] |
| Red Diamondback | Diamondback extended with SPHINX for early evolution | \(T_{\rm eff}=900\)–4000 K; \(\log g=3.0\)–5.5; \([{\rm M/H}]=-0.5,0.0,+0.5\); ages 1 Myr–15 Gyr [2510.08694] |
| Flame Skimmer | Colder cloud-free extension of Elf Owl | Described as extending cloud-free Elf Owl to colder effective temperatures, lower surface gravities, and a broader range of metallicities; in one application: \(T_{\rm eff}=150,175,200\) K, \(g=17,31\,{\rm m\,s^{-2}}\), \([{\rm M/H}]=-1.0\) to \(+1.0\), C/O \(=0.5,1.0,1.5,2.5\) times solar, \(\log K_{zz}=0,2,4,7,8,9\) [2602.23423] |

This branching structure reflects a practical division of labor. Bobcat and Cholla emphasize cloud-free atmospheres, Elf Owl emphasizes mixing-driven disequilibrium, Diamondback addresses the cloudy L/T regime, Red Diamondback extends Sonora into the young high-\(T_{\rm eff}\) regime, and Flame Skimmer targets very cold giant planets and Jupiter analogs [2107.07434][2110.11824][2402.00758][2510.08694][2602.23423].

## 2. Shared physical framework

Across the suite, Sonora models are 1D, plane-parallel, hydrostatic atmospheres in radiative–convective equilibrium. In the Bobcat description, hydrostatic equilibrium is written as
\[
\frac{dP}{dz} = -\rho g,
\]
and the flux balance condition as
\[
F_{\rm net}(P) = \sigma T_{\rm eff}^4.
\]
Convective zones are identified where \(\nabla_{\rm rad} > \nabla_{\rm ad}\), with convective adjustment imposing \(\nabla=\nabla_{\rm ad}\) [2107.07434].

Rainout chemistry is a defining assumption for the equilibrium branches. In Bobcat, condensates are removed from the overlying atmosphere once they form, so the observable gas phase is chemically distinct from a full local-equilibrium atmosphere with retained condensates. This treatment affects oxygen partitioning, alkali persistence, and the disappearance sequence of refractory species, and is one reason Sonora became a common baseline for cloud-free substellar modeling [2107.07434].

Radiative transfer is handled with correlated-\(k\) methods in the climate calculations, paired with higher-resolution post-processing for emergent spectra. The opacities incorporate updated molecular and atomic databases relative to earlier Saumon–Marley grids, including major absorbers such as H\(_2\)O, CH\(_4\), CO, CO\(_2\), and NH\(_3\), together with alkali line treatments and collision-induced absorption [2107.07434][2402.00758].

Disequilibrium branches introduce vertical mixing through an eddy diffusion coefficient \(K_{zz}\). In Elf Owl, the mixing timescale is
\[
\tau_{\rm mix} = \frac{H^2}{K_{zz}},
\]
and quench levels are defined by \(\tau_{\rm chem}=\tau_{\rm mix}\). Above a quench pressure, the abundances of species such as CH\(_4\), CO, CO\(_2\), H\(_2\)O, NH\(_3\), N\(_2\), HCN, and PH\(_3\) are frozen to their values at the quench level [2402.00756].

## 3. Clouds, disequilibrium chemistry, and branch-specific physics

The principal internal distinction within Sonora is between cloud-free and cloudy branches. Cholla and Elf Owl are cloud-free disequilibrium grids. Cholla focuses on solar metallicity and explicitly targets 500–1300 K atmospheres, while Elf Owl generalizes the disequilibrium treatment to non-solar metallicities and C/O ratios across nearly the full Bobcat \(T_{\rm eff}\) range [2110.11824][2402.00756].

Elf Owl showed that the impact of \(K_{zz}\) on the \(T(P)\) profile and spectra is a strong function of both \(T_{\rm eff}\) and metallicity. It also identified significant spectral degeneracies between varying \(K_{zz}\) and metallicity, particularly at \(3\)–\(5\,\mu{\rm m}\), where CH\(_4\), CO, and CO\(_2\) all respond strongly to both composition and mixing [2402.00756].

Diamondback introduces refractory clouds through the Ackerman & Marley framework. In that treatment, the vertical cloud structure is set by a balance between mixing and sedimentation,
\[
- K_{zz} \frac{\partial q_t}{\partial z} - f_{\rm sed} w_* q_c = 0,
\]
where \(f_{\rm sed}\) controls sedimentation efficiency. Low \(f_{\rm sed}\) corresponds to thick, vertically extended clouds; high \(f_{\rm sed}\) corresponds to thinner, more settled clouds. Diamondback includes Mg\(_2\)SiO\(_4\), MgSiO\(_3\), Fe, and Al\(_2\)O\(_3\) condensates, and its cloudy pressure–temperature structures can be hundreds of kelvin warmer at a given pressure than cloud-free structures [2402.00758].

Red Diamondback extends the cloudy Sonora framework to higher temperatures by stitching Diamondback to SPHINX M-dwarf atmospheres. The transition is handled through a blended \(T_{10}\) boundary condition between 2000 and 2400 K, allowing continuous evolution from cloudy brown-dwarf atmospheres into hotter, low-gravity M-type atmospheres. A major result is that young, massive brown dwarfs spend an early Hayashi-like phase at nearly constant \(T_{\rm eff}\), rather than following a simple backward extrapolation of older cooling tracks [2510.08694].

Flame Skimmer, as described in the \(\epsilon\) Eri b study, extends the cloud-free Elf Owl framework to colder effective temperatures, lower gravities, and a broader metallicity range. In that application it serves as a cloud-free baseline against which custom water-cloud models are compared, especially in the 150–200 K regime relevant to Jupiter analogs [2602.23423].

## 4. The CO\(_2\) revision and other recent corrections

A major recent revision in the Sonora family concerns Sonora Elf Owl v2. The original Elf Owl treatment quenched CO\(_2\) with respect to the full-atmosphere equilibrium, but CO\(_2\) should instead be quenched with respect to the disequilibrium abundance of CO. The corrected relation is
\[
f_{\rm CO_2} = K_{\rm eq}\,\frac{f_{\rm CO}\,f_{\rm H_2O}}{f_{\rm H_2}},
\]
with the crucial point that above the CO quench level, \(f_{\rm CO}\) must be the quenched abundance rather than the local equilibrium abundance [2505.03994].

The consequence was significant at low temperatures. The original Elf Owl grid underpredicted CO\(_2\) by more than an order of magnitude for objects with \(T_{\rm eff}\lesssim 600\) K, and in one 500 K example the mid-to-upper-atmosphere CO\(_2\) profile was too low by about two orders of magnitude. Version 2 corrected the CO\(_2\) treatment and removed PH\(_3\) as a spectral contributor because the spectra consistently contained too much PH\(_3\) absorption [2505.03994].

This correction matters observationally because JWST is sensitive to the affected CO\(_2\) bands. The same cold-planet context appears in Flame Skimmer, which is explicitly described as correcting the underestimation of CO\(_2\) found in earlier Sonora Elf Owl models and as incorporating both equilibrium and disequilibrium chemistry in a cloud-free framework with rainout chemistry for H\(_2\)O, CH\(_4\), and NH\(_3\) [2602.23423].

## 5. Coupling to evolution

Sonora was designed not only as an atmospheric grid suite but as an atmosphere–interior framework. In Bobcat, the evolution tables describe the cooling of substellar objects through time and provide \(T_{\rm eff}(t)\), \(L_{\rm bol}(t)\), \(R(t)\), \(\log g(t)\), and moment of inertia as functions of mass, age, and composition [2107.07434].

Diamondback generalizes this to cloudy evolution. Its evolution code uses atmospheric boundary conditions from the atmosphere grid and computes cloud-free, hybrid, and gravity-dependent hybrid clearing tracks. In the hybrid-grav prescription, the cloud-clearing temperature depends on surface gravity, reflecting the observational inference that low-gravity objects remain cloudy to cooler \(T_{\rm eff}\). The paper states that refractory clouds and metallicity both alter the evolution of substellar objects, changing the inferred temperature at a given age by up to 100–200 K [2402.00758].

Red Diamondback extends this coupling into the early-evolution regime. It computes evolution tracks from 1 Myr to 15 Gyr and from planetary masses to above the hydrogen-burning minimum mass, with atmospheric boundary conditions spanning \(T_{\rm eff}\approx 900\)–4000 K. Its early tracks show that low-gravity, high-\(T_{\rm eff}\) boundary conditions materially change young brown-dwarf cooling histories, especially during the deuterium-burning and Hayashi phases [2510.08694].

The evolutionary side of Sonora has also been used in Galactic population studies. A Solar-neighborhood synthesis employing Sonora Bobcat and Diamondback showed that luminosity functions and temperature distributions vary with height above and below the Galactic Plane, and that Diamondback’s gravity-dependent cloudy–cloudless transition produces a pronounced feature near \(T_{\rm eff}\approx 1300\) K that fades with increasing \(|z|\) as older, cooler populations dominate [2411.06330].

## 6. Observational applications and empirical performance

Sonora grids have been used widely as forward models for brown dwarfs and giant planets. In cold-brown-dwarf analyses with JWST, Sonora Elf Owl and ATMO2020++ delivered relatively consistent effective temperatures, but Sonora typically returned surface gravities about 1 dex lower than ATMO2020++, indicating strong model dependence in \(\log g\) even when \(T_{\rm eff}\) is stable [2409.19191]. In a large JWST/NIRSpec search for distant brown dwarfs, Sonora-Bobcat was used for initial candidate selection and Sonora Elf Owl v2 for atmospheric characterization, but the study found that cloud-free Sonora Elf Owl and LOWZ sometimes fail to reproduce L/T-transition spectra, whereas the cloudy SAND grid performs better in that regime [2510.02026].

Benchmark studies have repeatedly stressed the same point. In COCONUTS-1B, cloudless Sonora atmospheric models and Sonora evolutionary models agreed on temperature and radius but disagreed in gravity by \(\Delta\log g=-0.58\pm0.13\) dex, with the atmospheric models implying an unphysically young age. Assuming the evolutionary parameters, cloudless Sonora atmospheres produced Y- and J-band fluxes brighter than the data and a W2 flux fainter than expected, suggesting residual condensate clouds near 1300 K and non-equilibrium CO chemistry at \(4.6\,\mu{\rm m}\) [2002.05723]. In Ross 458c, Sonora Bobcat and Sonora Elf Owl both preferred super-stellar metallicities, but the paper argued that the poor fit quality and required error inflation made the retrieval-based metallicity inference more reliable [2509.22803].

Cloudy Sonora models have also been benchmarked directly. A uniform analysis of 142 late-M, L, and T benchmarks found that Sonora Diamondback and Saumon & Marley (2008) yield broadly consistent results, but also derived explicit empirical calibrations because raw atmospheric-fit \(T_{\rm eff}\), \(R\), and \(\log g\) show systematic offsets relative to evolution-based values. The same study found a statistically significant age dependence of \(f_{\rm sed}\) among L4–L9 dwarfs: young objects (\(<300\) Myr) exhibit systematically lower \(f_{\rm sed}\) than older counterparts, while no comparable trend appears across T0–T5 [2601.18866].

On the exoplanet side, Sonora Diamondback was one of the statistically preferred grids in a multi-modal analysis of \(\beta\) Pictoris b, alongside Exo-REM, but Sonora’s posteriors frequently lay at grid edges in \([\mathrm{M/H}]\) and \(f_{\rm sed}\), underscoring the difference between formal fit quality and physically robust inference [2509.25338]. Flame Skimmer was used as the cloud-free reference grid for \(\epsilon\) Eri b, where the JWST/NIRCam F444W non-detection was found to be consistent with a metal-enriched atmosphere and/or water ice clouds, making Sonora part of the interpretation of a cold Jupiter analog rather than only brown-dwarf spectra [2602.23423].

## 7. Limitations, recurrent tensions, and present interpretation

Several limitations recur across Sonora applications. First, cloud-free branches are inadequate for the L/T transition. This is explicit in the distant JWST brown-dwarf study, where Sonora Elf Owl “fails entirely” for at least one L/T-transition dwarf because it cannot reproduce cloud-driven near-infrared morphology, and more broadly in Ross 458c and COCONUTS-1B, where missing cloud opacity affects J, H, and K band structure [2510.02026][2509.22803][2002.05723].

Second, surface gravity remains the most model-dependent atmospheric parameter. An intercomparison of 14 brown-dwarf grids found that Sonora Bobcat and Sonora Cholla tend to predict \(\log g\sim 3\)–4 for many objects, even after excising data blueward of \(1.2\,\mu{\rm m}\), while effective temperature is comparatively robust across model families [2305.07719]. The same gravity sensitivity appears in the JWST cold-dwarf and benchmark analyses, suggesting that line-wing physics, cloud opacity, and disequilibrium chemistry remain entangled in present-day low-resolution gravity inference [2409.19191][2002.05723].

Third, model residuals point to missing opacity sources and imperfect rainout chemistry. In the large Diamondback benchmarking study, stacked residuals isolated persistent mismatches in TiO, FeH, CH\(_4\), and pressure-broadened K I bands, and an added ISM-like extinction term significantly improved fits across much of the sample. The paper interpreted this not as actual interstellar reddening but as evidence for missing high-altitude opacity in current cloudy models [2601.18866].

These tensions do not negate Sonora’s utility. They delimit where different branches are informative. Bobcat remains a baseline cloudless equilibrium grid; Cholla and Elf Owl expose the spectral role of disequilibrium chemistry; Diamondback provides the principal Sonora treatment of cloudy L/T atmospheres; Red Diamondback connects the suite to very young, high-\(T_{\rm eff}\) evolution; and Flame Skimmer extends the family to very cold giant planets. A plausible implication is that future Sonora development will continue along the directions already indicated by the literature: more flexible cloud physics, improved alkali and molecular opacities, and more self-consistent treatment of disequilibrium chemistry in the regimes where JWST is now most constraining [2402.00758][2505.03994][2601.18866].

Source: https://www.emergentmind.com/topics/sonora