The FORTRAN GLobal airglOW (GLOW) Model in Python ≥ 3.7.

A Fortran compiler is REQUIRED.

Note: This version uses meson and ninja as the build system, and does not rely on distutils, and is Python 3.12 compatible.

This library also allows parallelization of model evaluation using the multiprocessing module.

Additionally, the GlowModel class exposes complete control over model evaluation, which is divided into the following fundamental steps:

1. initialize: Set the altitude, energy grids and solar flux model (Hinteregger, EUVAC, custom).
2. setup: Set the model up for evaluation, specifying time, location, and optionally geomagnetic parameters. If geomagnetic parameters are not specified, they are retrieved using geomagdata.
3. precipitation (optional): Evaluate the energy and flux of precipitating electrons (Maxwellian or monoenergetic).
4. atmosphere: Evaluate the neutral atmosphere (MSISE-00) and ionosphere (IRI-90).
5. radtrans: Evaluate the GLOW radiative transfer model to calculate volume emission rates and ion composition.
6. result: Retrieve the model output as a xarray.Dataset.

Note: Between the atmosphere and radtrans steps, the intermediate dataset can be accessed using the result method by passing copy=False. This returns a reference to the internal dataset. Modifying the dataset in place allows for modifying the atmosphere and ionosphere presented to the radtrans step.

Note: For macOS Big Sur and above, you may need to add the following line to your environment script (~/.zshrc if using ZSH, or the relevant shell init script):

export LIBRARY_PATH="$LIBRARY_PATH:/Library/Developer/CommandLineTools/SDKs/MacOSX.sdk/usr/lib"

Then reopen the terminal. This fixes the issue where -lSystem fails for gfortran. Additionally, installation of gfortran using homebrew is required.

Direct installation using pip:

$ pip install glowpython

Install from source repository by:

$ git clone https://github.com/sunipkm/glowpython -b development
$ cd glowpython && pip install .

Requires (and installs) geomagdata for timezone aware geomagnetic parameter retrieval.

Then use examples such as:

• Examples/Maxwellian.py: Maxwellian precipitation, specify Q (flux) and E0 (characteristic energy).
• Examples/NoPrecipitation.py: No precipitating electrons.

These examples are also available on the command line: GlowMaxwellian and GlowNoPrecip.

Or, use GLOW in your own Python program by:

import glowpython as glow

iono = glow.maxwellian(time, glat, glon, Nbins, Q, Echar)

Read the module documentation for more information.

iono is a xarray.Dataset containing outputs from GLOW:

• Coordinates:
  • alt_km: Altitude grid (km)
  • energy: Energy grid (eV)
  • state: Neutral/Ionic excitation states (string)
  • wavelength: Wavelength of emission features in Angstrom (string). Represented as strings to accommodate the LBH band.
  • tecscale: TEC scale factor (float)
• Data:
  • Dimension (alt_km):
    • O: Neutral atomic oxygen density (cm^-3)
    • O2: Neutral molecular oxygen density (cm^-3)
    • N2: Neutral molecular nitrogen density (cm^-3)
    • NO: Neutral nitric oxide density (cm^-3)
    • NS: N(4S) density (cm^-3)
    • ND: N(2D) density (not used, set to zero) (cm^-3)
    • NeIn: Electron density (IRI-90), input to GLOW radiative transfer model. (cm^-3)
    • O+: Ion atomic oxygen (4S) density (cm^-3)
    • O+(2P): Ion atomic oxygen (2P) density (cm^-3)
    • O+(2D): Ion atomic oxygen (2D) density (cm^-3)
    • O2+: Ion molecular oxygen density (cm^-3)
    • N+: Ion molecular nitrogen density (cm^-3)
    • N2+: Ion molecular nitrogen density (cm^-3)
    • NO+: Ion nitric oxide density (cm^-3)
    • N2(A): Molecular nitrogen (GLOW) density (cm^-3)
    • N(2P): N(2P) density (cm^-3)
    • N(2D): N(2D) density (cm^-3)
    • O(1S): O(1S) density (cm^-3)
    • O(1D): O(1D) density (cm^-3)
    • NeOut: Electron density (calculated below 200 km for kchem=4 using GLOW model, cm^-3)
    • Te: Electron temperature (K)
    • Ti: Ion temperature (K)
    • Tn: Neutral temperature (K)
    • ionrate: Ionization rate (s^-1)
    • pederson: Pederson conductivity (S/m)
    • hall: Hall conductivity (S/m)
    • eHeat: Ambient electron heating rate (eV/cm^3/s)
    • Tez: Total energetic electron energy deposition (eV/cm^3/s)
  • Dimension (alt_km, wavelength):
    • ver: Volume emission rate of various features (1/cm^3/s)
  • Dimension (alt_km, state):
    • production: Production rate of various species (1/cm^3/s)
    • loss: Fractional loss rate of various species (1/s)
  • Dimension (energy):
    • precip: Precipitation flux (cm^-2/s/eV)
• Attributes:
  • time: Time of evaluation (ISO 8601 formatted string)
  • glatlon: Geographic latitude and longitude of evaluation (degrees)
  • Q: Flux of precipitating electrons (erg/cm^2/s)
  • Echar: Characteristic energy of precipitating electrons (eV)
  • iscale: Solar flux model. 0: Hinteregger, 1: EUVAC
  • xuvfac: XUV enhancement factor.
  • itail: Low energy tail enabled/disabled. 0: disabled, 1: enabled
  • fmono: Monoenergetic energy flux (erg/cm^2).
  • emono: Monoenergetic characteristic energy (keV).
  • jlocal: Local calculation only (disable electron transport).
  • kchem: GLOW chemistry level.

All available keys carry unit and description.

This repository started its life as a fork of the NCAR-GLOW model to upgrade from geomagindices that does not provide flux and Ap data post-SWPC. The NCAR-GLOW implementation runs the GLOW model as a separate process, invoked from a compiled GLOW binary, and runtime parameters are passed to this binary over STDIN and STDOUT is parsed to produce the Dataset containing the results of the simulation. This context switching, and text data parsing adds significant runtime overhead, and loss of precision. Starting v3.0.0, the core of []glowpython](https://github.com/sunipkm/glowpython) was changed to use F2PY, and pass inputs and get outputs from the FORTRAN library directly. Apart from the public-facing API (no_precipitation, maxwellian), the internal implementation of this repository is completely independent from NCAR-GLOW.