Last Updated: 07/09/2025
RANDM
Version: v.0.1A random forest model to predict atmospheric/thermospheric neutral densities at 400 km.
Caveats:
Only predicts at 400 km. Local time may be used in place of magnetic local time with a small additional error.
Inputs
The model inputs are : 4 FISM bands [1.3nm, 43.00nm, 85.55nm, 94.40nm] ; SYM-H index [and optional AE index]; Latitude, cos(MLT), sin(MLT) where MLT is magnetic local time. The model also needs datetime to download new model inputs when they are not part of the base input file.
Outputs
Thermospheric density at 400 km.
Change Log
version 0.0.1 is the first version of the model
Domains
- Thermosphere
Space Weather Impacts
- Atmosphere variability (satellite/debris drag)
Publications
- Murphy, K., Halford, A. J., Liu, V., Klenzing, J., Smith, J., Garcia-Sage, K., et al. (2025). Understanding and modeling the dynamics of storm-time atmospheric neutral density using random forests. Space Weather, 23, e2024SW003928.
- zenodo: mltdm- Initial Release including Model skops file and feature data
Code
Code Languages: Python
Contacts
- Alexa Halford, NASA GSFC (Model Developer)
- Kyle Murphy, null (Model Developer)
- Chris Bard, NASA GSFC (Model Developer)
- Jia Yue, NASA/GSFC (CCMC Model Host)
Publication Policy
In addition to any model-specific policy, please refer to the General Publication Policy.