Ring Current Proton Artificial Neural Network (RCPANN) Model
How to use the RCPANN Model:
Install the rcpann python package via "pip install rcpann"
For user manual and examples, please visit https://pypi.org/project/rcpann/
Model Details:
Input: Geomagnetic indexes with a look back window of ~10 days; Satellite orbit
Output: Proton flux at a specific energy
Neural Networks: Fully-connected artificial neural network with 3 hidden layers, and ~15 neurons per hidden layer
RCPANN Model Description:
Terrestrial ring current dynamics are a critical part of the near-space environment, in that they directly drive geomagnetic field variations that control particle drifts, and define geomagnetic storms. The Ring Current Proton Artifical Neural Network (RCPANN) model aims to specify a global and time-varying distribution of ring current proton using geomagnetic indices and solar wind parameters with their history as input. We train an Artificial Neural Network (ANN) model to reproduce proton fluxes measured by the RBSPICE instrument onboard Van Allen Probes. By choosing optimal feature parameters and their history length, the model results show a high correlation and a small error between model specifications and satellite measurements. The modeled results well capture energy-dependent proton dynamics in association with geomagnetic storms, including inward radial diffusion, acceleration and decay. The RCPANN model produces proton fluxes with their corresponding 3D spatiotemporal variations, capturing the latitudinal distribution and local time asymmetry that are consistent with observations and that can further inform theory.
Global dynamic distribution of proton fluxes from 28 February to 5 March 2017 produced by the RCPANN model.
(a) The Sym-H index; (b) The SME index; (c) The observed 55 keV proton fluxes along Van Allen Probe A orbit. (d) The modeled 55 keV proton fluxes along the satellite orbit. (Bottom) The dynamic global distribution of 55 keV proton fluxes over a moderate geomagnetic storm cycle. The L value is the geocentric distance in units of Earth Radii at magnetic equator.
Comparison between Machine-Learned Model predictions and satellite measurements of proton fluxes over the entire year of 2017.
The ANN model This model produces a global distribution of proton fluxes with spatiotemporal variations, capturing the latitudinal distribution and local time variations that are consistent with observations and useful for informing theory. The X_sm and Y_sm are geocentric distances in the Solar Magnetic coordinates in units of Earth Radii.
Download the Model: https://github.com/jimmyli87/RCPANN
Reference:
Li, J., et al. (2022), Modeling Ring Current Proton Fluxes Using Artificial Neural Network and Van Allen Probe Measurements, Space Weather, https://doi.org/10.1029/2022SW003257.