Modeling wind-driven seed dispersal using a coupled Lagrangian particle tracking and 1-D [formula omitted]-[formula omitted] turbulence model

Abstract

We present a one-dimensional k-ɛ model and a Markov chain stochastic Lagrangian particle tracking model to study wind-driven dispersal of forest and grassland seeds. The turbulence model within and above the canopy is implemented for transient winds so that time series of profiles of mean velocity, turbulence kinetic energy, dissipation rate, and the root-mean-square of fluctuating velocity components can be modeled. We validate the model for steady state conditions using observations in a wide range of canopies, and apply the model for predicting time series of seed dispersal of Andropogon gerardii in a tall grass prairie in Missouri, United States. The stochastic seed dispersal model is validated against experimental data from a Quercus-Carya (oak-hickory) forest where seeds were released from different heights for a range of wind speeds. Using the model, we explore the detailed seed movement trajectories, instantaneous velocities, and investigate the relationship between dispersal distance and the seed falling duration and the mean fluctuating velocity during the falling. Moreover, examining the role of individual turbulent velocity components, we elucidate the different roles of horizontal and vertical velocity fluctuations, i.e., the horizontal component mainly contributes to the seed spreading in the wind direction, while the vertical component enhances the long tail of the distribution in the downstream direction, i.e., promoting long-distance-dispersal of seeds by wind.