Improving CME modelling with data assimilation of Heliospheric Imager observations into the HUXt solar wind numerical model.

Coronal Mass Ejections that impact Earth drive the most severe space weather. To better enable effective space weather mitigation plans, there is much interest in improving the quality of CME arrival time predictions, particularly by quantifying and reducing the prediction uncertainty. A limited set of observatories, challenges in interpreting observation data, and limiting assumptions in CME parameterisations all play important roles in the uncertainty of the predicted CME evolution.

Data assimilation techniques provide a path for improving the predictive skill, by integrating observations into a modelling framework in a way that returns model states that better reflect the true state of a system. Furthermore, such techniques can self-consistently account for uncertainty in the observations, and uncertainty in the models structure and parameterisations.

We present some early results from our work to build a particle filter data assimilation scheme around the HUXt solar wind model. Assimilating the time-elongation profiles of CME flanks observed by the Heliospheric Imagers on NASAs STEREO mission, we demonstrate that such methods have good potential to improve modelled CME arrival time predictions. Using a simulation study, we present an estimate of the potential CME arrival time prediction improvements gained by using this particle-filter approach with an L5 Heliospheric Imager.