Estimating impacts of space weather on the upper atmosphere

Description:

Space weather drives large changes in the upper atmosphere. Variations in thermosphere density result in significant uncertainties in orbit estimation for satellites in low earth orbit. Consequently, the ability to accurately predict satellite collision risks, lifetimes, and re-entry is limited even as use of these orbits increases dramatically.

Physics-based models of the ionosphere-thermosphere system are one way of addressing this challenge. These methods numerically solve equations representing physical and chemical processes in the upper atmosphere. Physics-based models can be combined with data assimilation techniques to allow for improved estimation based on observations. Reduced order models can be derived from physics-based models, retaining important dynamics while dramatically reducing computational cost. Potential PhD topics may focus on one of these areas or the combination thereof.

Notes:

The project would combine work in up to three areas of modelling the ionosphere-thermosphere system, depending on the student's interests and expertise. Namely these are: improving the fidelity of physics-based models of the upper atmosphere, data assimilation, and reduced order modelling. Working with BoM, we would seek to bring these methods closer to operational application.

Improving the fidelity of physics-based models of the ionosphere-thermosphere system is an ongoing effort. As of November 2023, we have one student (Atishnal Chand) working in this area using the Global Ionosphere Thermosphere Model (GITM). His work has focused on improvement of the high latitude chemistry and particle precipitation models in this code. Another PhD project in this area would expand our contribution to the development of GITM. Improving the grid representation of GITM in polar regions has been identified as important to improving model accuracy and computational performance. Alternatively, an improved numerical solver could be used to extend the maximum altitude in the model.

While physics-based models of the ionosphere-thermosphere system have not yet proven more accurate than statistical alternatives, they allow incorporation of observations via data assimilation. The ensemble Kalman filtering technique uses an array of model runs to represent the state of the system and its uncertainty. These are periodically adjusted based on observational data, considering uncertainties in both the prior state and these observations. The Data Assimilation Research Testbed (DART) provides a framework for implementing this technique with models such as the Thermosphere Ionosphere General Circulation Model (TIE-GCM). We have previously used these tools to estimate the state of the ionosphere-thermosphere system during geomagnetic storms. A PhD project in this area would include novel types of observation into the data assimilation method, such as GOLD UV spectrograph and SuperDARN radar measurements.

Physics-based models tend to be computationally expensive, often requiring high performance computing. Recently, there has been research into reduced order modelling of the ionosphere-thermosphere system. These models develop a simplified representation of the system state and its response to space weather drivers. This may be derived from physics-based models, like TIE-GCM or GITM, or empirical data, such as the SET HASDM database. Such reduced order models could be applied with data assimilation techniques. Very large ensembles, which are impractical with computational models, could be achieved with ensemble Kalman filtering or particle filtering techniques.