| Abstract: Understanding CME kinematics in the middle corona is essential for linking their initiation physics with heliospheric evolution. However, conventional height–time tracking along discrete position angles cannot resolve the internal velocity dispersion between CME substructures such as the core, cavity, and front.
We present a two-dimensional optical-flow (OF) framework that derives continuous velocity fields directly from coronagraph image sequences, enabling systematic tracking of CME expansion and internal dynamics. The method was validated using LASCO C2 and STEREO/COR1 data, showing strong agreement with conventional techniques despite their lower cadence and spatial resolution. We then applied the technique to Solar Orbiter/METIS and PROBA-3/ASPIICS observations after extensive background removal and temporal-Fourier filtering to mitigate current brightness-flicker artifacts around the occulter in PROBA-3's Level-2/3 data.
The retrieved flow fields yield both centre-of-motion and expansion velocities for multiple limb CMEs. For example, in the 16 April 2024 METIS event, the average centre-of-motion and expansion speeds were 337 km/s and 196 km/s, respectively. Multi-wavelength analysis using METIS Ly-alpha and FSI 174/304 Å, together with ASPIICS wideband observations, reveals a critical height range of 2.2–2.6 solar radii where systematic velocity dispersion between the core and front first develops. Furthermore, Intensity-based segmentation enables the extraction of internal velocity distributions, revealing significant dispersion within both regions, but with velocity density concentrated near the median. The lower-velocity contributions arise primarily from the CME flanks.
Finally, the measured time lag between the core and front velocity profiles, together with their radial separation, implies an information transfer speed of ~3000–3500 km/s. This exceeds both sound and Alfvén speeds, indicating that a fast-mode MHD wave likely mediates momentum coupling between CME substructures.
This method also enables automated CME detection through evolving velocity distributions, paving the way for coordinated kinematic studies with current and upcoming missions such as Solar Orbiter, PROBA-3, and PUNCH. |
