Three-Dimensional Phase-Native Transition Mapping in Tokamak Data

We report a three-dimensional phase-native analysis of tokamak data in which transition structure becomes more coherent when the system is represented in intrinsic rather than purely laboratory-time coordinates. Using a reduced proxy space constructed from measured discharge observables, we analyze the public TCV L-H ensemble (179 events from 112 unique shots) using normalized accumulated heating, normalized stored energy, and a phase-distortion proxy. In this representation, the transition localizes to a compact fold-like region in 3D state space and exhibits sharper organization than in standard time-indexed plots or pairwise projections alone. The result is phenomenological and data-driven: we do not claim a first-principles plasma theory, but rather a practically useful re-representation that regularizes the observed transition geometry. The central claim is narrow by design. For the analyzed TCV dataset, a 3D phase-native mapping improves visibility of transition structure and yields a falsifiable prediction that comparable future L-H events should cluster in the same bounded region of the constructed state space.