Physical Processes Affecting Prediction of Upper Atmosphere Space Weather

When the National Space Weather Program Strategic Plan was released in 1999, it was expected that investments in scientific research directed towards the Earth’s thermosphere and ionosphere would lead to the practical benefit of space weather prediction. If experience with prediction of tropospheric weather serves as a guide, then it must be admitted that the link between scientific knowledge and prediction of natural phenomena is not a straightforward one. While the tropospheric weather community discovered “chaos” as a fundamental limit to predictability for 5-10 day forecast lead times, the space weather community faces additional challenges. The global behavior of the thermosphere-ionosphere (T-I) domain is strongly dependent on driving from above and below, so that accurate characterizations of the driving forces are needed. As part of the NASA/NSF Partnership For Collaborative Space Weather Modeling, we are emphasizing methods of improving predictive skill in the T-I as scientific knowledge and observational capabilities improve. Our approach is meant to succeed despite varying degrees of scientific understanding, and despite environmental factors that are often poorly constrained. We will present results from a group of physics-based models of the coupled thermosphere-ionosphere that simulate what a forecast would be, emphasizing the physical processes that create storm-time changes. Energy flow through the system and its representation in the models is a major focus. We will describe how data-driven and statistical forecasts can provide insight into the physical processes acting during storms, and how such approaches naturally have a role to play within the space weather domain, even if the emphasis remains on physics-based forecast models.