Uncertainty Quantification and Parameter Estimation Utilizing a Global Ionosphere Thermosphere Model

In any large-scale model of a complex system there will always be sub-grid-scale or parameterized phenomena. These types of phenomena may include things such as reconnection, resistivity, viscosity, conduction, diffusion, heating and cooling processes, reaction rates, etc. In published literature, there are often conflicting values or parameterizations for these processes. So, often modelers choose one and use that in their simulations. We show examples of the consequences of these types of choices in the global ionosphere thermosphere model (GITM). Specifically, we show how utilizing different published conductivity coefficients, reaction rates and cooling rate affect the thermospheric temperature structure during idealized times and during storm times. We compare these results to measurements to show how the uncertainties within the model parameterizations can alter the comparisons. Further, we show a new technique for utilizing measurements within the simulation to estimate the parameters while the model is running. We demonstrate that this technique provides correct results and is quite useful.