System identification based models for electron fluxes measured by GOES 16

System identification is employed to develop a set of models to forecast the electron fluxes at geostationary orbit. The energy of the electron fluxes are based on the measurements of (GOES 16) (MPS-HI). The GOES 16 generation of spacecraft measures a wide range of electron energies, 10 energy channels from 50 keV to 4 MeV, compared to its predecessors, which has allowed the development of data based electron flux forecast models covering a wide range of radiation belt electrons. The models are deduced using the system identification technique based on Nonlinear AutoRegressive Moving Average eXogenous (NARMAX) models, which deduces a model from input-output data and accounts for the spatial variation of the electron flux. The performance of these models were compared to the persistence of GOES 16 measurements and the develop NARMAX models were shown have higher performance metrics. These performance metrics were shown to have a dependence between energy and magnetic local time. The models have higher performance metrics at the dayside compared to the nightside and, generally, the models have higher performance metrics for higher energies compared to lower energies. The forecast horizon of these models also exhibit a relationship with energy, where the forecast horizon increases with energy, with the 69 keV electron flux model having a forecast horizon of only 2 hr, while the 2,907 keV electron flux model has a forecast horizon of 27 hr.