The goal of this project is to improve models of the electrodynamic behavior of the magnetosphere, ionosphere, and thermosphere through statistical techniques for finding optimal values of unobserved parameters. Three such parameters will be estimated: 1) a coefficient of proportionality between electron energy and plasma temperature; 2) a parameter relating the flux of precipitating electrons to the electron energy and the plasma density; and 3) a parameter relating the energy gain of electrons traversing a potential drop to the plasma density, electron energy, and the strength of the electric current aligned with the magnetic field. While these parameters cannot be observed, optimal values for them can be estimated by comparing model output generated using specific parameter values against appropriate observations (magnetospheric storms as seen by the POLAR Ultraviolet Imager, for example). Models used to simulate the magnetosphere, ionosphere, and thermosphere are computationally expensive, so that only a limited number of parameter values can be tested. This research will develop strategies, based on hierarchical statistical modeling and the development of scalar error measures, to efficiently search the parameter space for optimal parameter values.
Better models of the ionosphere, magnetosphere, and thermosphere are of interest for practical as well as scientific reasons. Strong currents during magnetic disturbances heat the upper atmosphere, causing it to expand upward and greatly increase the air density at satellite altitudes, which can significantly affect the satellite orbits. In addition, the electron density of the ionosphere is strongly affected by the electric fields on both quiet and disturbed days, and variations in this density have impacts on satellite networks used for communication and geolocation. Thus, research conducted here could contribute to an upper-atmospheric forecasting capability of great practical benefit.
