Geomagnetic forcing in T-I models: comparison of empirical, data assimilative, and physics-based forcing evaluated with radar measurements

Forecasting and mitigating space weather effects requires accurate modelling of the coupled Magnetosphere-Ionosphere-Thermosphere (MIT) system. There are multiple ways to couple the magnetospheric dynamics to a thermosphere-ionosphere model. Most commonly, empirical models such as Heelis and Weimer are applied. To improve upon empirical models, data assimilative techniques such as AMIE and AMGeO have been developed. These techniques assimilate various radar, magnetometer, and satellite-based measurements into an empirical background model. A comparably recent development is the MAGE geospace model, which couples multiple physics-based models of the entire MIT system. We compare these methods with each other and evaluate them with various measurements.

One of the most important geomagnetic impacts on the thermosphere-ionosphere is Joule heating due to Pedersen currents. We evaluate the different forcing approaches by comparing the resulting Joule heating in reference to local measurements with the EISCAT incoherent scatter radar. We show that data assimilative methods provide a significant improvement over empirical forcing.

Physics-based geomagnetic forcing promises a model representation of small-scale processes that cannot be achieved with empirical methods. However, an initial assessment showed significant discrepancies between the polar plasma convection pattern given by a physics-based geospace model and SuperDARN radar network measurements. Since Joule heating is affected by changes in electron density and plasma convection potential, we evaluate the model representation of these quantities separately with EISCAT, SSUSI, and SuperDARN measurements.