Dependencies¶

GEE ACOLITE relies on two categories of dependencies: Python packages (declared in pyproject.toml) and the ACOLITE processor (bundled as a git submodule).

Dependency Graph¶

Python Packages¶

earthengine-api >= 0.1.350¶

The official Google Earth Engine Python client. All server-side operations (image filtering, band arithmetic, reducers, masking) are expressed as GEE API calls that execute lazily on GEE infrastructure.

Used for:

• ee.ImageCollection — filtering Sentinel-2 catalog by region, date, tile
• ee.Image.expression() — DN → TOA reflectance, atmospheric correction formula
• ee.Reducer.percentile() / ee.Reducer.toList() — dark spectrum extraction (the only getInfo() call per image)
• ee.Reducer.linearFit() — pSDB calibration regression
• ee.Image.updateMask() — water, cloud, cirrus masking
• ee.batch.Export — export to Drive / Asset / Cloud Storage

Install: pip install earthengine-api

numpy >= 1.20.0¶

Used exclusively on the client side for numerical operations during AOT estimation.

Used for:

• np.interp() — interpolating LUT arrays to match observed dark spectrum reflectance values
• np.argsort(), np.mean(), np.std() — sorting and statistical analysis of AOT estimates across LUT models
• Array manipulation for multi-band LUT lookups

scipy >= 1.7.0¶

Used on the client side for statistical analysis during the dark spectrum fitting.

Used for:

• scipy.stats.linregress() — linear regression to estimate the dark spectrum intercept (the intercept spectrum option)
• RMSD computation for model selection (min_drmsd criterion)

netcdf4 >= 1.7.0¶

Required by the ACOLITE package for reading its atmospheric Look-Up Tables stored in NetCDF format. Not used directly by gee_acolite but required at runtime.

ACOLITE (Submodule)¶

The ACOLITE processor is bundled as a git submodule in the ACOLITE/ directory. GEE ACOLITE acts as an adapter that extracts the following ACOLITE components for use with GEE imagery:

acolite.acolite.settings¶

acolite.acolite.settings.parse('S2A_MSI', settings=settings)

Parses and validates processing settings (from a file path or dict), merging user-provided values with sensor-specific defaults for Sentinel-2.

acolite.aerlut¶

acolite.aerlut.import_luts(sensor, settings)
acolite.aerlut.import_rsky_luts(sensor, settings)

Loads atmospheric LUT (Look-Up Table) files for the specified sensor. Each LUT is a multidimensional array indexed by:

• Pressure (surface atmospheric pressure, hPa)
• Path reflectance (romix) — reflectance of the atmosphere-surface system
• Relative azimuth angle (RAA)
• Viewing zenith angle (VZA)
• Solar zenith angle (SZA)
• AOT at 550 nm (τ₅₅₀)

For each combination, the LUT stores per-band values of:

acolite.shared¶

acolite.shared.rsr_dict(sensor)

Returns the spectral response function (SRF) for each Sentinel-2 band, used to compute band-integrated gas transmittances.

acolite.ac¶

acolite.ac.gas_transmittance(settings, rsrd)
acolite.ac.ancillary.get(date, lon, lat)

• gas_transmittance: Computes per-band gas transmittance (O₃, H₂O, O₂, CO₂) given pressure, ozone, and water vapor.
• ancillary.get: Fetches ancillary atmospheric data (pressure, ozone, water vapor, wind speed) from NASA Earthdata for a specific date and location.

Optional Dependencies¶