Reconstruction of Chorus Wave Amplitude Profiles by Incorporating Electron Fluxes Into a Machine Learning Framework

Abstract

Chorus waves, typically excited by anisotropic electrons with energies from hundreds of eV to tens of keV, play a critical role in the dynamics of Earth's outer radiation belt. Previous chorus wave models are commonly driven by geomagnetic indices and solar wind parameters, which capture large‐scale activity but may not fully reproduce localized and rapid wave variations. This limitation may partly arise from the omission of microphysical drivers, such as electron flux associated with wave excitation and evolution. Here we incorporate 1–100 keV electron flux information into a machine learning framework to improve reconstruction of lower band chorus wave amplitudes. Compared with an index‐only baseline model, the new approach increases the overall Pearson correlation coefficient from 0.38 to 0.53 and reduces the root‐mean‐square error from 20.3 to 18.6 pT. The inclusion of in situ electron flux information also helps capture intense localized chorus activity during geomagnetically quiet periods, which is often underestimated by models driven only by macro‐scale inputs. These results show that local electron flux provides substantial additional predictive information for chorus reconstruction. The framework provides a practical pathway toward generating proxy chorus wave data from particle measurements, thereby enhancing the space weather utility of missions that do not carry dedicated wave instruments.