Virga Cloud Modeling Framework

• Virga Cloud Modeling Framework is a Python-based suite that simulates cloud vertical structures and particle size distributions in exoplanet and substellar atmospheres.
• The framework applies steady-state mass continuity to balance turbulent mixing and gravitational settling with variable sedimentation efficiency and fractal aggregate properties.
• It integrates microphysical parameters and radiative transfer coupling, enabling robust spectral interpretation from instruments like HST and JWST.

The Virga Cloud Modeling Framework is a Python-based, 1D equilibrium condensation suite that models cloud vertical structure, particle size distributions, aggregate morphology, and spectral properties in exoplanet and substellar atmospheres. It formalizes the steady-state balance of upward turbulent mixing and downward sedimentation for condensates, grounded in the Ackerman & Marley (2001) EddySed theory and extended to capture variable sedimentation efficiency, fractal aggregate aerosols, and multi-species cloud compositions. Virga is widely employed to interpret atmospheric spectra, benchmark microphysical models, and retrieve physically motivated cloud parameters from observations ranging from HST to JWST (Batalha et al., 20 Aug 2025, Kiefer et al., 13 Mar 2026, Moran et al., 8 Sep 2025, Lodge et al., 3 Dec 2025, Rooney et al., 2021, Mang et al., 2022).

1. Governing Physics and Core Equations

Virga operationalizes the vertical mass continuity for condensates as a balance between eddy diffusion and gravitational settling: $$-K_{zz} \frac{\partial q_t}{\partial z} - f_{\rm sed} w_* q_c = 0$$ where $K_{zz}$ is the eddy diffusivity, $q_t$ the total (vapor + condensate) mass mixing ratio, $q_c$ the condensate mixing ratio, $w_*$ the convective velocity scale ($w_*=K_{zz}/H$), and $f_{\rm sed}$ is the dimensionless sedimentation efficiency (Kiefer et al., 13 Mar 2026, Batalha et al., 20 Aug 2025). The first term quantifies upward turbulent transport and the second represents downward particle flux.

Steady-state closure integrates sources and sinks across vertical layers under local phase equilibrium, relating $q_v$ (vapor), $q_{\rm sat}$ (saturation, via Clausius-Clapeyron), and $q_c$ by

$K_{zz}$0

Virga adopts a log-normal particle size distribution: $K_{zz}$1 characterized by mean radius $K_{zz}$2 and width $K_{zz}$3 (Kiefer et al., 13 Mar 2026).

The particle fall speed $K_{zz}$4 is computed using physical drag laws across all Reynolds and Knudsen regimes (Batalha et al., 20 Aug 2025), with transition between Stokes, slip-corrected, and kinetic drag formulations for spheres; aggregates follow specific free-molecular and continuum scaling (Moran et al., 8 Sep 2025).

2. Sedimentation Efficiency and Altitude Dependence

The sedimentation efficiency parameter $K_{zz}$5 is central to Virga's predictive power. It encodes the mass-weighted mean particle settling velocity relative to the convective velocity scale: $K_{zz}$6 Physically, larger $K_{zz}$7 yields rapid settling, large particles, and geometrically thin clouds; small $K_{zz}$8 yields slow settling, small particles, and vertically extended clouds (Kiefer et al., 13 Mar 2026, Rooney et al., 2021).

Virga supports both constant and variable altitude-dependent $K_{zz}$9 parameterizations. The latter are necessary to reproduce cloud profiles from comprehensive microphysical models (e.g., CARMA). Example forms include exponential density dependence

$q_t$0

where $q_t$1, $q_t$2, and $q_t$3 are tunable parameters enforcing normalization and physical limits (Rooney et al., 2021, Mang et al., 2022). The analytic structure of the governing equation is retained for computational efficiency.

Comparison with microphysical models shows that no constant-$q_t$4 profile can reproduce the curvature and vertical extent of clouds under heterogeneous nucleation, necessitating flexible forms for retrievals and for jointly interpreting different spectral regions (Rooney et al., 2021, Mang et al., 2022, Kiefer et al., 13 Mar 2026).

3. Treatment of Particle Morphology: Fractal Aggregates

Virga v2.0 and later generalize beyond compact spheres to arbitrary fractal aggregates (Moran et al., 8 Sep 2025, Lodge et al., 3 Dec 2025). Particle morphology is controlled via the fractal dimension $q_t$5 (1 for chains, 3 for spheres) and monomer radius $q_t$6 or number $q_t$7: $q_t$8 with $q_t$9 a prefactor (Tazaki 2021). Aggregate growth is parametrized either with fixed $q_c$0 or fixed $q_c$1 (Moran et al., 8 Sep 2025).

Aggregate settling velocities are derived for both free-molecular and continuum drag regimes, explicitly accounting for the reduced density and cross-sectional scaling of fluffy structures. The closure ensures mass conservation and permits robust solution of the equilibrium fall-speed-radius condition.

Virga's optical property module for aggregates leverages the Modified Mean Field (MMF) theory as implemented in Optool to produce extinction, scattering, and asymmetry parameters across parameter grids of $q_c$2, $q_c$3, and $q_c$4 (Moran et al., 8 Sep 2025, Lodge et al., 3 Dec 2025).

4. Optical Properties and Radiative Transfer Coupling

Virga computes wavelength-dependent cloud optical properties using Mie theory for spheres and MMF theory for aggregates (Batalha et al., 20 Aug 2025, Moran et al., 8 Sep 2025, Lodge et al., 3 Dec 2025). Layer-by-layer extinction optical depth, single-scattering albedo, and asymmetry parameters are integrated over the local particle size distribution and passed to external radiative transfer codes (e.g., PICASO) for spectral synthesis (Batalha et al., 20 Aug 2025).

For a fixed cloud mass, the spectral impact of morphology is regime dependent:

• Rayleigh regime ($q_c$5): fluffier aggregates are more opaque.
• Geometric regime ($q_c$6): compact particles dominate extinction.

Morphology alters transmission slope, NIR/IR band muting, and the strengths of diagnostic features such as the $q_c$7m silicate band (Lodge et al., 3 Dec 2025, Moran et al., 8 Sep 2025). The MMF implementation yields efficiency within $q_c$81–2% of DDA models but with $q_c$9150× computational speedup (Moran et al., 8 Sep 2025).

5. Input Parameters, Model Workflow, and Species Support

Virga requires as input vertical profiles of pressure, temperature, and eddy diffusivity. Cloud species selection is flexible, with support for at least 18 condensates (volatile ices, sulfides, silicates, metals, oxides) along with their corresponding vapor pressure parameterizations, refractive indices, and densities (Batalha et al., 20 Aug 2025). Users can add new species by supplying optical and vapor-pressure data.

A typical workflow involves:

1. Preparing atmospheric $w_*$0–$w_*$1 and $w_*$2 profiles.
2. Selecting condensate species and cloud/morphology parameters ($w_*$3, $w_*$4, $w_*$5).
3. Solving the mass-balance ODE to obtain $w_*$6 and particle size distributions.
4. Computing opacity profiles and synthesizing spectra using RT solvers (Batalha et al., 20 Aug 2025).

Model output includes vertical distributions of $w_*$7, $w_*$8, number densities, optical depths $w_*$9, and diagnostic properties for comparison with observed spectra.

6. Validation, Applications, and Limitations

Virga has been benchmarked against both observed exoplanet atmospheres (e.g., WASP-17b SiO$w_*=K_{zz}/H$0 clouds) and microphysics models (CARMA, Sonora grid) (Batalha et al., 20 Aug 2025, Mang et al., 2022, Moran et al., 8 Sep 2025). For retrieval studies, $w_*=K_{zz}/H$1 and aggregate parameters are tuned to match spectral features, such as the strength and shape of $w_*=K_{zz}/H$2m silicate absorption seen in JWST data. Very small $w_*=K_{zz}/H$3 and low sticking coefficients $w_*=K_{zz}/H$4 are required to reproduce high-altitude, small-grain clouds as implied by observations of WASP-107b, VHS-1256b, and YSES-1c (Kiefer et al., 13 Mar 2026).

Morphology critically impacts spectra, especially for broadband and short-wavelength datasets. Virga's ability to flexibly explore $w_*=K_{zz}/H$5 and morphology parameter spaces makes it amenable to joint retrieval and forward-modeling efforts, but it should be cross-validated against full microphysics (e.g., growth and coagulation) where possible (Rooney et al., 2021, Kiefer et al., 13 Mar 2026).

The current model is parametric and does not resolve explicit kinetic microphysics (nucleation, coagulation cascades, break-up, or full radiative feedback), but it is extensible and computationally efficient for retrieval and sensitivity studies (Lodge et al., 3 Dec 2025, Moran et al., 8 Sep 2025). Guidance is provided for physical parameter ranges and consistency checks to prevent unphysical solutions (e.g., porous aggregate compressibility limits).

7. Future Development and Recommendations

Ongoing and planned Virga developments include support for aggregate/hollow-core particles, coupling to 2-D GCM outputs for spatially resolved clouds, and retrieval-mode interfaces (Batalha et al., 20 Aug 2025). Community contributions are encouraged via the open-source code repository. Constraints from laboratory sticking coefficients for silicates, together with panchromatic spectra, are refining microphysics-to-observable connections.

Modelers are encouraged to utilize shape and $w_*=K_{zz}/H$6 as free parameters, benchmark against laboratory and microphysical models, and leverage the rapid analytic solutions of Virga for comprehensive parameter space exploration and joint retrievals on multi-instrument datasets (Lodge et al., 3 Dec 2025, Kiefer et al., 13 Mar 2026, Rooney et al., 2021).