AGNI-integrated Rocky Surfaces and Atmospheres Analysis

A modular Python-based pipeline for modeling exoplanetary atmospheres and surfaces, generating synthetic eclipse depths, and comparing outputs with observational contrast data. The system integrates with a Julia-based radiative transfer model (AGNI) to simulate and analyze a variety of atmospheric and surface conditions.

---

Overview

This project allows the user to:

• Model generation: Automatically configure and run AGNI using planetary and stellar parameters, surface albedo profiles, and atmospheric compositions.

• Synthetic eclipse depth computation: Simulate secondary eclipse depths for blackbodies, greybodies, bare-rock surfaces, and full atmosphere-surface radiative transfer models.

• Temperature fitting: Fit the planet’s dayside brightness temperature from observed contrast (eclipse depth) data.

• Bayesian model comparison: Use nested sampling to compute the Bayesian evidence for each model and evaluate their relative support given the data.

• JWST simulation: Calculate band-integrated thermal emission in JWST/MIRI filters (e.g., F1280W, F1500W), compute signal-to-noise ratios, and assess detectability across surface and atmosphere types.

---

Structure

VAYU/
│
├── AGNI/                     # Julia-based AGNI radiative transfer engine
├── misc/                     # Miscellaneous files
├── res/    
│   ├── atmospheres/          # Basic atmosphere composition TOMLs  
│   ├── config/               # Generated AGNI configuration TOMLs
│   ├── planetary_data/       # Basic planet data, and observational contrasts
│   ├── stellar_spectra/      # Stellar spectra files (.txt)
│   └── surfaces/             # Surface albedo files (.dat)
│
├── out/                      # Model results and plots saved here (per planet)
├── plt/                      # Additional plotting scripts, primarily used for my thesis
├── src/
│   ├── atmosphere_labels.py  # Labels for atmosphere keys
│   ├── chi2_table.py         # Creates chi-2 summary table
│   ├── config_gen.py         # Generates AGNI config TOMLs
│   ├── constants.py          # Physical and astronomical constants
│   ├── dataloader.py         # Loads data 
│   ├── emission_miri.py      # Simulates observed emission with JWST Miri filters
│   ├── nested_sampling.py    # Uses a nested sampling method to compare models to the data (Zhang et al, 2024)
│   ├── pipeline.py           # Main orchestration script
│   ├── plots.py              # Plotting utilities
│   ├── stat.py               # Statistical functions 
│   ├── temperature_fit.py    # Fit temperature to observational data
│   ├── throughput.py         # Calculate JWST Miri filter throughput for simulated emission
│   └── utils.py              # Math + I/O helpers (Planck, contrast, etc.)
│
├── agni_config.toml          # Paths to directories
├── atmos_list.toml           # Optional list of atmospheres to loop over
└── README.md                 # This file 

---

Installation

1. Clone the repository

   git clone https://github.com/stuitje/VAYU.git
   cd VAYU

2. Set up Python environment

   Make sure you have the required dependencies installed:

   pip install -r requirements.txt

   Required packages include:

   • numpy, pandas, matplotlib, scipy, toml, tomlkit, netCDF4, os

3. Install AGNI

   Follow the AGNI installation instructions as explained in: https://nichollsh.github.io/AGNI/dev/setup/. AGNI must be placed inside the root directory (VAYU/). From the root directory, you can test AGNI using: .

   julia AGNI/test/runtests.jl

4. Configure paths

   If needed, update agni_config.toml to reflect your machine's paths. Default:

   # Base AGNI directory (used to access the model engine)
   agni_dir = "AGNI"
   
   # Input files and directories 
   planet_csv = "res/planetary_data/exoplanetarchive.csv"
   obs_data_dir = "res/planetary_data"
   atmosphere_dir = "res/atmospheres"
   surface_dir = "res/surfaces"
   stellar_spectra_dir = "res/stellar_spectra"
   
   # AGNI input files 
   spectral_file = "AGNI/res/spectral_files/Honeyside/256/Honeyside.sf"
   spectral_file_H2O = "AGNI/res/spectral_files/Frostflow/256/Frostflow.sf"
   spectral_file_O2 = "AGNI/res/spectral_files/Honeyside_with_O2/256/Honeyside.sf"
   spectral_file_Si = "AGNI/res/spectral_files/Rocks/256/Rocks.sf"
   
   # JWST observational modelling
   pandeia_dir = "/dataserver/users/formingworlds/stuitje/pandeia"
   
   # Outputs
   output_dir = "out"
   config_dir = "res/config"
   
   

---

Usage

Run a full pipeline

python -m src.pipeline <planet_name> -s <surface_name|all|list> -a <atmosphere_name|all|list> [--flux true|false]

• planet_name: Name of the planet as in the CSV.
• -s: 'all' to run over all surface files, 'list' (from surface_list.toml), or specific name.
• -a: 'all', 'list' (from atmos_list.toml), or specific name.
• -T: temperature model to use, which modifies the heat redistribution factor used. 'dayside' is the default, for no heat redistribution but an average dayside flux (f = 2/3); 'substellar' assumes no heat distribution and only the substellar point (f = 1); and 'full' assumes full heat redistribution (f = 1/4).
• --flux-only: to just generate flux-and-contrast plots (no combined contrast plots).
• --no-run: skips running AGNI, but directly uses the atm.nc file generated by AGNI earlier to generate output. Can be used when, for example, modifying the plot itself.

Example: running a pipeline for Trappist-1c, using a 1 bar CO2 atmosphere with full heat redistribution (f = 1/4).

python -m src.pipeline trappist-1c -s greybody -a 1bar_CO2 -T full

Example: running a pipeline for Trappist-1c, using a list (surface_list.toml) of bare surfaces with no heat redistribution (f = 2/3, default).

python -m src.pipeline trappist-1c -s list -a bare_rock

---