Automated detection and characterision of CMEs in near real-time coronagraph data: lessons and challenges arising from the SWEEP project

The Space Weather Empirical Ensemble Package (SWEEP) is a 3-year project that, very recently, delivered a suite of software tools to the UK Met Office to improve their space weather forecasting capabilities. Part of this package was Automated CME Characterisation (ACMEC) software to automatically detect, and characterise, CMEs in near real-time (NRT) coronagraph data. ACMEC ingests STEREO COR2 beacon data, and SOHO LASCO C2/C3 NRT data, and given the availability of clean data, provides estimates of the CME’s 3D trajectory, angular extent, and speed. It also provides estimates of uncertainties in these values, enabling an ensemble forecast at Earth. We present example case studies to show the efficiacy of ACMEC, and reflect on the challenges faced, and lessons learned during the development stages. Future perspectives are given, including new missions, and the realisation that the development of machine learning methods are required for such complicated tasks in the next 10 years.