Trajectory Curvature As A Selection Criterion For valid Lagrangian Stochastic Dispersion Models

Abstract

Among well-mixed multi-dimensional Lagrangian stochastic (LS) dispersion models, we observe that those in poorest agreement with observations produce ‘spiralling trajectories,’ with an associated reduction in dispersion. We therefore investigate statistics of increments dθ ' to the orientation θ'= arctan(W'/U') of the Lagrangian velocity-fluctuation vector – as a possible means to distinguish the better LS models within the well-mixed class. ‘Zero-spin’ models, having 〈 dθ'〉 = 0, are found to provide best agreement with observations. It is not clear however, whether imposition of the zero-spin property selects (in conjunction with the well-mixed condition) a unique model.