Interpretation of adiabatic and diabatic populations from trajectories of branching corrected surface hopping

The branching corrected surface hopping (BCSH) has been demonstrated as a robust approach to improve the performance of the traditional fewest switches surface hopping (FSSH) for nonadiabatic dynamics simulations of standard scattering problems [J. Chem. Phys. 150, 164101 (2019)]. Here, we study how reliable populations of both adiabatic and diabatic states can be interpreted from BCSH trajectories. Using exact quantum solutions and FSSH results as references, we investigate a series of one-dimensional two-level scattering models and illustrate that excellent time-dependent populations can be obtained by BCSH. Especially, we show that different trajectory analysis strategies produce noticeable differences in different representations. Namely, the method based on active states performs better to get populations of adiabatic states, while the method based on wavefunctions produces more reliable results for populations of diabatic states.