Ozone and Radiation Sensitivity Modeling

A Python package to model ozone column changes and radiative effects in response to various aircraft emissions at different altitudes. This package includes a region-specific response model based on empirical emission sensitivities and Taylor expansion, as well as a pre-trained multilayer perceptron trained on 60 fleet emission scenarios.

Features

• Modeling of ozone change and radiative forcing (response model, MLP)
• Supports multiple geographic regions (response model)
• Emission-specific sensitivity interpolation to altitude of emission (response model, MLP)
• Optional Taylor expansion w.r.t. cruise altitude to compare to reference aircraft (response model)
• Prediction of seasonal variation of ozone column changes and radiative forcing (MLP)

Use the software

Installation

Requires Python 3.9 or newer. Install the package:

pip install -e .

For development, install the test and lint tools as well, then run the tests and the linter:

pip install -e ".[dev]"
pytest
ruff check src/ tests/

Running the response model

Calculate ozone change:

An example on how to use the code is shown here

from response_model.taylor_model import calculate_delta_F_emissions

# Define inputs
mode = 'Radiative_Forcing'  # or 'Ozone'
region = "Transatlantic_Corridor"
altitude_km = 18.0
emissions = {
    'NO': (100, "GgNO2"),  # Gigagrams of NO2 per year
    'SO': (5, "GgS"),      # Gigagrams of sulfur per year
    'H2O': (50, "TgH2O"),  # Teragrams of water vapor per year
}

# Calculate ozone change or radiative forcing
emission_values = {key: val[0] for key, val in emissions.items()}
delta_F = calculate_delta_F_emissions(altitude_km, emission_values, region, mode=mode)

# Format emissions into a readable string with units
emission_str = ', '.join([f"{key} = {value} {unit}" for key, (value, unit) in emissions.items()])
unit_str = "DU" if mode == "Ozone" else "mW/m²"

print(
    f"The {mode} effect of a supersonic aircraft flying across the {region} "
    f"at {altitude_km:.1f} km, emitting {emission_str}, is estimated to be ΔF = {delta_F:.2f} {unit_str}"
)

Example output (formatted)

The radiative forcing effect of a supersonic aircraft flying across the Transatlantic_Corridor at 18.0 km, emitting:

• NO = 100 GgNO₂
• SO = 5 GgS
• H₂O = 50 TgH₂O

is estimated to be:

ΔF = 9.76 mW/m²

Other examples

More extensive examples including reference aircraft for comparison are shown in src/scripts/example.py You can execute this via

python3 src/scripts/example.py

Running the multilayer perceptron

For an example on how to load and use the trained multilayer perceptron see src/scripts/example_MLP.ipynb.

Project Structure

supersonic-effects/
├── src/
│   ├── response_model/
│   |   ├── __init__.py
│   |   ├── taylor_model.py
│   |   ├── read_data.py
│   |   └── resources/ 
│   |       ├── __init__.py 
│   |       ├── sensitivity_*.csv 
│   |       └── taylor_param_*.csv
│   ├── multilayer_perceptron/
│   |   ├── __init__.py
│   |   ├── load_mlp.py
│   |   └── resources/ 
│   |       ├── __init__.py 
│   |       └── trained_mlp_vanthoff_etal_2026.onnx 
│   └── scripts/
│   |   ├── example_MLP.ipynb
│   |   ├── example.py
│   |   ├── test_altitude.py
│   |   └── test_emission.py
├── resources/
|   ├── __init__.py
│   ├── sensitivity_*.csv
│   └── taylor_param_*.csv
├── tests/
│   ├── test_mlp.py
│   ├── test_model.py
│   └── test_validate_*.py
├── README.md
├── pyproject.toml
...

Resources

The data underlying the software originates from Van 't Hoff et al. 2024

• resources/sensitivity_ozone.csv: Empirical sensitivities (mDU / unit / year)
• resources/sensitivity_radiative_forcing.csv: Empirical sensitivities (mW/m2 / unit / year)
• resources/taylor_param_ozone.csv: 1st and 2nd order coefficients for altitude effect (DU / km, DU / km²)
• resources/taylor_param_radiative_forcing.csv: 1st and 2nd order coefficients for altitude effect (mW/m2 / km, mW/m2 / km²)